|
Citation |
- Permanent Link:
- http://ufdc.ufl.edu/AA00055802/00001
Material Information
- Title:
- The relative predictive value of selected variables in measuring teacher success in an urban school district
- Creator:
- Thompson, Jon Danforth, 1936-
- Publisher:
- University of Florida
- Publication Date:
- 1974
- Language:
- English
- Physical Description:
- vii, 109 leaves. : ; 28 cm.
Subjects
- Subjects / Keywords:
- Educational evaluation ( jstor )
Employment interviews ( jstor ) Graduates ( jstor ) Mathematical dependent variables ( jstor ) Personnel evaluation ( jstor ) Psychological assessment ( jstor ) Recommendations ( jstor ) School districts ( jstor ) Teacher evaluation ( jstor ) Teachers ( jstor ) Dissertations, Academic -- Educational Administration and Supervision -- UF Educational Administration and Supervision thesis Ed. D Teachers -- Selection and appointment ( lcsh ) Duval County ( local )
Notes
- Thesis:
- Thesis (Ed.D.)--University of Florida.
- Bibliography:
- Bibliography: leaves 106-108.
- General Note:
- Typescript.
- General Note:
- Vita.
- Statement of Responsibility:
- by Jon D. Thompson.
Record Information
- Source Institution:
- University of Florida
- Holding Location:
- University of Florida
- Rights Management:
- This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
- Resource Identifier:
- 029478165 ( ALEPH )
14276447 ( OCLC )
Aggregation Information
- UFIR:
- Institutional Repository at the University of Florida (IR@UF)
- UFETD:
- University of Florida Theses & Dissertations
- IUF:
- University of Florida
|
Downloads |
This item has the following downloads:
|
Full Text |
THE RELATIVE PREDICTIVE VALUE OF
SELECTED VARIABLES IN MEASURING TEACHER
SUCCESS IN AN URBAN SCHOOL DISTRICT
By
JON D. THOMPSON
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF EDUCATION
UNIVERSITY OF FLORIDA 1974
ACKNOWLEDGMENTS
To my doctoral committee, Dr. Michael Nunnery, Dr. William Alexander, Dr. Ruthellen Crews and Dr. Forbis Jordan my sincere gratitude is extended for their personal efforts and their influence on the final product. Special appreciation is extended to Dr. Nunnery for his patience and helpful suggestions.
To my wife Connie, and my children Bryan and Julie, we share in the joy of the conclusion of the study.
11
TABLE OF CONTENTS
PAGE
ACKNOWLEDGMENTS. . 1. 1
ABSTRACT. . . . . v
CHAPTER
I. INTRODUCTION. . 1
The Problem. . 2
Statement of the Problem . . . . 2 Delimitations 3
Limitations 4
Justification 5
Assumptions 6
Definition of Terms 7
Dependent Variable 7
Independent Variable . . . . . 7 Interview Assessment *. . 7 Interview Assessment Sheet . . . 7 Personal History Data . . . . 8 Recommendations Assessment . . . 8 Sample A 8
Sample B 8
Sample C 9
Procedures 9
The Samples 9
Sources and Collection of Data . . 10 Treatment of Data 13
Organization of the Remainder
of the Study 14
II. REVIEW OF LITERATURE. . . 15
Criteria Determination. . 16
Recruitment 19
Application or Personal History Data . 20 Recommendations 22
Interviews 24
Synopsis of the Review of Literature . 32
iii
CHAPTER PAGE
III. PRESENTATION OF DATA. . .33
Basic Descriptive Data Relative to the
Independent and Dependent Variables
for Samples A and B. . . .34
Personal History Data for Samples
A and B . . .34
Interview Assessment Data for
Samples A and B. . 40
Recommendations Assessment for
Samples A and B .44
Evaluation Assessment for Samples
A and B 44
Independent Variable Data in Relationship
to the Dependent Variable for Samples
A and B *. . 47
Personal History Data in Relationship
to Evaluation Assessment .48
Interview Assessment Related to
Evaluation Assessment. . . 61
Recommendations Assessment in
Relationship to Evaluation
Assessment. . . .64
Results of Regression Analysis for
Samples A and B .69
A Comparison of the Distribution of the
Independent Variables for Samples A,
B, and C .72
IV. DISCUSSION OF THE DATA. . .8.7
V. SUMMARY, CONCLUSIONS AND IMPLICATIONS . 97
Summary .97
Conclusions and Implications . . . 103
SELECTED BIBLIOGRAPHY. . . . 106
BIOGRAPHICAL SKETCH. . . . 109
iv
Abstract of Dissertation Presented to
the Graduate Council of the University of Florida
in Partial Fulfillment of the Requirements for
the Degree of Doctor of Education
THE RELATIVE PREDICTIVE VALUE OF
SELECTED VARIABLES IN MEASURING TEACHER
SUCCESS IN AN URBAN SCHOOL DISTRICT By
Jon D. Thompson
August, 1974
Chairman: Michael Y. Nunnery Major Department: Educational Administration
The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district.
Specifically, the focus was on the following:
1. For a sample of teachers who had performed
satisfactorily enough to be offered a third contract by the urban school district, to determine
the relative predictive value of personal history
data, interview assessment and recommendations
assessment and any combination thereof, when the
average of three principals' annual evaluations was
used as the criterion measure.
v
2. For a sample of teachers in their first year of teaching within the urban school district and about whom no decision had been made relative to their continued status, to determine the relative predictive value of personal history data, interview assessment and recommendations assessment and
any combination thereof, when the principal's annual
evaluation was used as the criterion measure.
3. To compare the two aforementioned samples of teachers in regard to their basic characteristics
with a third sample of teachers consisting of those
teachers who did not perform to a sufficiently
satisfactory level to be retained under contract
in the urban school district.
Data for each independent and dependent variable for
each member of the three teacher samples involved were collected from the personnel records of the single urban school district. Frequency distributions, joint frequency distributions and multiple regression equations were developed for the two samples of 100 teachers each. The characteristics of the two samples of teachers then presently employed (1974) were compared to the third sample of 21 teachers who were not reappointed by the urban school district. Based upon the preceding. analysis, the following major findings emerged:
1. The values of multiple R, the intercorrelation
of the independent variables as well as the dependent vi
variable for the two samples of 100 teachers,
ranged from .25153 to .31898.
2. R-square, the proportion of the variance in the dependent variable accounted for by the re-gression equation was .05834 for the sample of teachers in their third year of employment and
.10175 for the sample of teachers in their first
year of employment.
3. When the three samples were compared, significant
differences were found on the variables of race,
interview assessment and recommendations assessment. It was concluded that:
1. The relative usefulness of personal history data,
interview assessment and recommendations assessment
when utilized in the selection of teachers for an
urban school district was limited.
2. The findings of the present investigation were
consistent with findings of other studies under
similar circumstances.
3. Even though peripheral to the major intent of the investigation, based on the data obtained when Samples
A, B, and C, were compared, it seems justifiable to conclude that if persons are selected for teaching positions when these individuals have poor interview and recommendations ratings, then the likelihood of their success is
quite limited.
vii
CHAPTER I
INTRODUCTION
Personnel administrators in local school systems have continually faced the problem of selecting the best qualified applicants to fill teaching positions. Two events have occurred within the last few years that have resulted in greater emphasis being placed on insuring that the teacher selection process was effective. First, the number of applicants per vacancy has increased rapidly. This fact makes it indefensible for school systems to choose teachers at random. Secondly, society has placed greater pressure on schools to achieve their goals. Despite the fact that the goals have often been ill-defined, the task of achieving any educational goal is more readily accomplished with the best qualified teacher applicant being selected for employment. 2
Numerous methods have been used in attempting to ,elect the best qualified applicant. They indlude: college grades, previous experience, letters of recommendation, personal interviews, background data (e.g., age, sex, socioeconomic status) and statements of an applicant's educational philosophy.
Research conducted on the effectiveness of these methods
1Dale L. Bolton, Selection and Evaluation of Teachers. (Berkeley: McCutchan Publishing Corporation, 1973), p. 1.
2Ibid., p. 2.
1
2
has not been conclusive. It would appear that there have been no universal guidelines developed that apply to a cross section of school districts. Bolton suggested that it was necessary for each school district to validate its own selection method.3
In view of the wide range of techniques used and the obvious need to improve the selection process, the present study was undertaken. Utilizing data from a single urban school district the intent was to upgrade the selection process in the urban school district and to add to the present body of knowledge.
The Problem
Statement of the Problem
The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district.
Specifically, the focus was on the following:
1. For a sample of teachers who had performed
satisfactorily enough to be offered a third contract by the urban school district, to determine
the relative predictive value of personal history
data, interview assessment and recommendations
assessment and any combination thereof, when the
average of three principals' annual evaluations was
3Ibid., p. 2.
3
used as the criterion measure.
2. For a sample of teachers in their first
year of teaching within the urban school district and about whom no decision had been made relative
to their continued status, to determine the relative predictive value of personal history data, interview assessment and recommendations assessment and any combination thereof, when the principal's annual evaluation was used as the criterion
measure.
3. To compare the two aforementioned samples of
teachers in regard to their basic characteristics
with a third sample of teachers consisting of those
teachers who did not perform to a sufficiently
satisfactory level to be retained under contract
in the urban school district. Delimitations
The study was confined to three samples of teachers in
a single urban school district. Two samples, of 100 teachers each, were randomly chosen. Sample A consisted of teachers hired in 1971-72 and who were in continuous employment in the urban school district to January 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the district. Sample B included those teachers initially employed in 1973-74 by the urban school district and who were in continuous employment
4
to January 1974. This sample constituted the sample about whom no decision had been made regarding their continued employment in the urban school district. Sample C, con-sisting of 21 arbitrarily chosen teachers, composed the group of teachers who had not performed at a sufficiently satisfactory level to be retained by the urban school district and were no longer employed at the time of the investigation.
Data for each individual teacher in each of the three samples relative to the independent variables (personal history data, interview assessment and recommendations assessment) were collected exclusively from the personnel records of the urban school district. Data regarding the dependent variable, principal's annual evaluations, were also collected from the personnel records of the urban school district.
Limitations
Four major limitations were recognized in conducting
the present study. First, the data used in making judgments about the relative usefulness of the various independent variables did not include samples from the full array of applicants who were available to the urban school district. That.is, there was no attempt to select samples from among applicants who were not offered employment in the urban school district. Second, from the literature it was found that for any evaluation instrument to be effective, a position
5
analysis should be developed for each job. The evaluation instrument used by the urban school district was of a general type and was used for all instructional personnel and had no relationship to a specific job. Third, if an individual principal's annual evaluation is viewed as being an insufficient assessment of one's performance, then the study suffered from a weakness in the strength of the dependent variable. Fourth, since the study was confined to a single urban school district, its external validity beyond the district was open to question. Justification
The development and refinement of methods to accurately predict which applicants will be successful on the job and therefore should be employed have been studied extensively. The major portion of these studies however, has focused on the selection procedures for industry and government. Bolton, in Selection and Evaluation of Teachers,4 indicated that few systems utilize validated selection techniques for teacher employment. Due to teacher shortages in the past, this situation has been tolerated by educators and the public. The teacher supply in the early 1970's, and the increased concern over accountability in education should cause school personnel officials to develop more sophisticated selection methods.
4Ibid., p. 49.
6
Based upon personal experience in personnel work and a review of the literature on the selection process, it was the opinion of the writer that data collected in regard to certain variables aie superior to data collected in regard to other variables in predicting teacher success on the job and that a refined selection procedure can be designed for school systems to improve the decision-making process.
The conclusions of this study should provide a basis for designing a selection procedure that would assist the urban school district in making more effective teacher selections. In addition to being of value to the urban school district, the proposed study was perceived as important in that it adds to what Bolton has described as the meager body of knowledge available on teacher selection.5 Assumptions
The basic assumption for this study was that the teachers used as the sample would be similar enough to future teachEr applicants in the urban school district so that any conclusions reached would be reasonably valid for teacher selection decisions within the single urban school district.
5Ibid., p. 51.
7
Definition of Terms
Dependent Variable
For Sample A the dependent variable consisted of the mean of three annual evaluations, each completed by the principal of the school the teacher was assigned to for a given school year. The dependent variable for Sample B consisted of the mean of the annual evaluation for the 1973-74 school year completed by the principal of the school to which the teacher was assigned. Independent Variable
The values determined for personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status and race), interview assessment data and recommendations assessment data.
Interview Assessment
The rating assigned each applicant by the interviewer of the urban school district, based upon his evaluation of the applicant at the time of the interview. (The method for quantification of the rating is explained in the procedures section of this report.) Interview Assessment Sheet
The sheet used by the interviewer of the urban school district when recording his assessment of the applicant interviewed. The applicant was rated on a continuum in
8
six categories: appearance, voice, poise, manner, enthusiasm and interest.
Personal History Data
The term included: age, college location, previous experience, transcript assessment, marriage status and race. The items of information were supplied to the urban school district by the applicant on the application blank or college transcript, prior to employment. Recommendations Assessment
The numerical value assigned by the recommending party converted to a 1-10 scale or, in the case where only narrative data were provided by the recommending party, the writer determined the numerical value utilizing the procedures detailed in the procedures section of this report. Sample A
This sample consisted of 100 teachers, randomly
chosen, who were hired in 1971-72 and who were in continuous employment in the urban school district to January 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the urban school district.
Sample B
This sample consisted of 100 teachers, randomly chosen, who were initially employed in 1973-74 by the urban school district and who were in continuous employment to January 1974. This sample constituted the sample about whom no
9
decision had been made regarding continued employment in the urban school district.
Sample C
This sample consisted of 21 arbitrarily chosen
teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district and were no longer employed at the time of the investigation. The teachers in this sample were employed in the urban school district for varying lengths of time during the period 197173.
Procedures
In order to determine the relative usefulness of
personal history data, interview assessment and recommendations assessment in the selection of teachers in a single urban school district, the following procedures were utilized.
The Samples
Three samples, which included elementary and secondary teachers, were utilized in this study. Sample A consisted of 100 teachers, randomly chosen,who were hired in 1971-72 and who were in continuous employment in the urban school district to January 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the district. Sample B consisted of 100 teachers, randomly chosen, who were initially employed in
10
1973-74 by the urban school district and who were in continuous employment to January 1974. This sample constituted the sample about whom no decision had been made regarding continued employment in the urban school district. Sample C, consisting of 21 arbitrarily chosen teachers, composed the group of teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district and were no longer employed at the time of the investigation. The teachers in this sample were employed by the district for varying lengths of time during the period from 1971-73. Sources and Collection of Data
The files of the urban school district were utilized
for securing the data for each of the independent variables: personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status, and race), interview assessment and recommendations assessment for each teacher in each of the three samples. Data relative to the dependent variable, evaluations assessment, were also secured in the files of the urban school district for each teacher in Sample A and Sample B.
1. Personal history data. With the exception of grade point average, the data included in the category were contained on the application form that the urban school district required each applicant to complete. The application form became part of the employee's permanent file along with the transcript, after the applicant was employed. (The permanent files of all personnel were made available to the writer.)
11
2. Interview assessment. The interviewer for the urban school district evaluated the applicant on each of six personal characteristics at the conclusion of the interview. The results of all personal interviews were recorded on the interview evaluation sheet, designed for that purpose.
The six personal characteristics were: appearance, voice, poise, manner, enthusiasm and interest. The interviewer indicated his assessment of the applicant by placing a check on a continuum for each of the characteristics. The writer divided each continuum into ten equal parts for purposes of quantification. Thus each applicant attained an interview assessment score from 6-60. (Six represented a favorable rating, 60 an unfavorable rating.)
3. Recommendations assessment. Recommendations submitted about applicants were filed with the Interview Evaluation Sheet, separate from the permanent file. Most recommendations were submitted on a form with a scale to which a numerical value could be assigned. For those recommendations that were submitted in a narrative form, the writer determined the rating of the applicant based upon his experience in assessing recommendations for the single urban school district and a study by Peres and Garcia, reported in the review of literature.6 Specifically, the writer
6Sherwood H. Peres and J. Robert Garcia, "Validity
and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnel Psychology 16: 279-86 (Autumn 1962).
12
examined the letters of recommendation to locate the most discriminating categories of adjectives as determined by Peres and Garcia. Peres and Garcia found that adjectives dealing with mental ability were the most discriminating. Adjectives dealing with vigor, dependability-reliability, urbanity and cooperation-consideration were found to be correspondingly less discriminating.7 The writer determined whether the adjectives in the letters of recommendation were favorable or unfavorable to the applicant. A numerical rating from one to ten was then assigned by the writer to represent the recommendations assessment for a particular teacher.
4. Evaluation assessment. Data in regard to the
dependent variable, the teacher's annual evaluation as completed by the principal to whom the teacher was assigned for the given school year, were secured in the personnel files of the urban school district. The evaluation for the school year 1971-72 contained 31 items that were used to rate the teachers. The evaluation for the 1972-73 and 1973-74 school years contained 21 items that were used by the principal to rate the teachers. The mean was calculated for each teacher's evaluation for each year. For Sample A, the evaluation assessment represented the mean of the teacher's three annual evaluations. For Sample B, the evaluation assessment represented the mean of the teacher's 1973-74 annual evaluation.
7Ibid., p. 285.
13
Treatment of Data
To determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers for an urban school district, the following operations were performed.
1. Frequency distributions were developed for Samples A and B in regard to each of the nine variables to provide some idea of the basic characteristics of the data in terms of distribution and variability. Specifically, frequency distributions were developed that showed: absolute frequency, percent and cumulative percent.
2. Joint frequency distributions were developed to show how the dependent variable related to each of the independent variables. Separate joint frequency distributions were developed for Sample A and Sample B. The format consisted of:
a. Listing the numerical total of participants
for each cell of the joint frequency distribution.
b. Indicating row percent, the percent of each
dependent variable category in a particular
category of the independent variable.
c. Indicating the column percent, the portion of
each independent variable category in a particular
category of the dependent variable
d. Indicating total percent, the portion of
the total (100 percent) each cell represented.
3. Two regression equations were computed using: age, graduate of a university in Florida, years of previous experience, grade point average, race, interview assessment
14
and recommendations assessment as independent variables to predict the dependent variable (evaluations assessment). The multiple regression analysis produced the highest possible correlation between the independent variables and the dependent variable. The summary tables for the regression equations present the following information: multiple r, r-square, r-square change, B and Beta. Multiple r indicates the linear relationship between the independent variables and the dependent variable under investigation. R-square change is the change in r-square from the value of the previous step. The B and Beta values are the regular and normalized regression coefficients, respectfully.
4. Frequency distributions were developed to compare Samples A, B, and C in regard to each of the eight independent variables. Specifically, frequency distributions were developed that showed the percent of Samples A, B, and C in each category of the eight independent variables.
Organization of the Remainder of the Study
Chapter II contains a review of literature which
focuses primarily on previous studies in the selection of personnel. Chapter III is a presentation of the data. Chapter IV contains a discussion of the data and Chapter V contains the summary, conclusions and implications.
CHAPTER II
REVIEW OF LITERATURE
Approximately 240,000 teachers were selected for employment in the United States in 1970-71.1 For each teacher selected, other applicants were considered and rejected. Yet, the decisions for choosing these teachers and rejecting the others were based upon data that generally had not been validated and ignored research relating to the selection of employees.2 Apparently the limited research dealing directly with the selection of teachers or at least the teacher shortage did not provide an impetus for school systems to develop more sophisticated methods of teacher selection.
Considerable research on the selection process has been performed in industry and government and has been gradually applied to the selection process for teachers. The present review is arranged under subject headings based upon a review of the literature by Bolton in Selection and Evaluation of Teachers.3 These are: criteria determination, recruitment, application or personal history data, recommendations and interviews.
1Dale L. Bolton, Selection and Evaluation of Teachers, (Berkeley: McCutchan Publishing Corporation, 1973), p. 1.
2 Ibid., pp. 1-4.
Ibid., pp. 50-92.
15
16
Criteria Determination
To determine if selection methods are successful, a criterion for success on the job must be established. Traditionally in education this has taken the form of deriving a single standard of a "good" teacher. Robert Gui6n disputed this practice,
The fallacy of the single criterion lies in its assumption that everything that is to be
predicted is related to everything else to be
predicted--that there is a general factor in
all criteria accounting for virtually all of
the important variance in behavior at work and
its various consequences of value.4
Dunnette stated that personnel involved in the selection process must be willing to back off from the all encompassing criteria of job effectiveness. Dunnette advocated a more careful study of job behavior with emphasis upon differing styles within the same job. The selection process should avoid assuming that there is a single all-encompassing measure of occupational success.5 Dunnette pointed out that data processing equipment has given employers greater capabilities to include more selection data and evaluation criteria to make selection decisions, than was once felt possible.6
4Robert M. Guion, and Richard Bottier, "Validity
of Personality Measures in Personnel Selection," Personnel Psychology 18: 135-64 (Summer 1965).
5M. D. Dunnette, "A Modified Model for Test Validation and Selection Research," Journal of Applied Psychology 47: 317-23 (October 1963).
6Ibid., p. 320.
17
Supporting Dunnette's view that a single criterion
measure is a weakness of many selection methods is a modification of typical prediction models offered by Guetzkow and Forehand. Their model included the complex interactions which may occur between predictors and various predictor combinations, different groups of individuals, different behaviors of the job and the consequences of these behaviors relative to the goals of the organization.7
Similar job behaviors may be predicted by quite different patterns of interaction between various predictors and individuals, according to Guetzkow and Forehand. The modified model for test validation and selection research proposed by Guetzkow and Forehand did not rely upon the correlation coefficient as the sole criterion for validation of selection techniques. Their model provided a "richer schematization" for prediction research and offered important implications for the directions of future research.8
The fact that criteria change from time to time and
from one situation to another has been demonstrated. Bernard Bass in a study of food product salesmen measured evaluations over a period of 48 months. The correlation coefficient of two separate merit ratings by essentually the same supervisors fell from .62 to .29 over a 42 month period, indicating the criterion used by the supervisors varied from the
7Harold Guetzkow and Garlic A. Forehand, "A Research Strategy for Partial Knowledge Useful in the Selection of Executives," In Research Needs ip Executive Selection, edited by Ranato Tagiuri, Boston: Harvard University, Graduate School of Business Administration, Division of Research, 1961.
8Ibid., p. 48.
18
beginning of the period to the end. In the same study Bass found a higher correlation coefficient between esteem and merit (,42) than merit and ability (.18).
To avoid the situation where evaluators vary on the factors used to rate employees, researchers have recommended developing a position analysis. In making this determination Bolton suggested that the employer ask, "What must the individual do and what characteristics must he possess to be a success."10 Bolton listed the following guides.
1. The number of behaviors listed should be
limited.
2. Behavior should relate to the purpose of
the situation in which they are to be
exhibited.
3. Behavior should be measurable.
4. Behavior should include verbal interactions
as well as non-verbal communications, and
in-class as well as out-of-class action. 11
Garland DeNelsky and Michael G. McKee conducted a
study for the Central Intelligence Agency in which a modified Q-Sort was used that reliably expanded the variances of predictor and criterion variables.12 Seven staff psychologists evaluated 32 male employees. When the researchers
9Bernard M. Bass, "The Leaderless Group Discussion," Psychological Bulletin 51: 465-92 (September 1954).
10Bolton, p. 58.
11Ibid., pp. 59-60
12Garland Y. DeNelsky and Michael B. McKee, "Prediction
of Job Performance from Assessment Reports: Use of a Modified Q-Sort Technique to Expand Predictor and Criterion Variants," Journal of Applied Psychology 53: 439-45 (December 1969).
19
correlated the composite assessment-report predictions of the seven judges with the individual predictions of each of the seven judges, the results produced correlation coefficients of .63 for Group One and .66 for Group Two for individual Drediction, and .92 and .93 for the respective groups with composite predictions. Although the psychologists were more accurate in predicting weaknesses than strengths, it did suggest that pooled ratings for either selection or evaluation produced more accurate decisions.13
Recruitment
Fundamental to the effective selection of teachers is
the existence of a sufficient number of qualified candidates 14
from which to make selections. Since selection implies choice, recruitment is necessary to attract high quality applicants in a sufficient number so that validated selection techniques can be utilized. The intensity of a recruitment program is related to the availability of teachers in a given area. Some school districts are forced to a more aggressive recruitment program due to their remoteness to an adequate supply.
The basis for any selection program is to attain quality personnel who will remain for a period of time with their
Ibid., p. 441.
14Bolton, p. 60.
20
employer. Recruitment of personnel who fit this requirement has been shown to justify the additional cost of the recruitment. A study by Fitzgerald found the average cost per teacher hired in a midwestern state to be $146. Business and industry were found to pay up to $1,822 for recruitment of each professional employed.15 If the good teacher has a more positive effect on students than an average or poor teacher then it is difficult to rationalize the variance in expenditure.
Application or Personal History Data
Information on the application form,usually demographic data, has been widely used to predict employee success as measured by supervisory ratings. Dunnette commenting in 1963 about such data stated:
We cannot and should not try to avoid the fact
that statistics of selection (i.e. validity
coefficients) are far from gratifying and offer
little support to anyone claiming to do much
better than chance in the selection process.16
In contrast to studies compiled by Dunnette, that
reported relatively low correlation coefficients for validating biographical data, was a report entitled "The Prediction of Research Competence and Creativity," from Personal History,
15Paul Fitzgerald, "Recruitment of Teachers-- A Need
for Reevaluation," Personnel Journal 49: 312-14 (April 1970).
16Dunnette, p. 320.
17Ibid., p. 318.
21
by Smith, Albright, Glennen and Owens.18 The personal history questionnaire contained 484 multiple-choice items that were designed for the study. The results of the study produced a correlation coefficient of .613 when comparing the 484 items to supervisory ratings. This was considerably higher than found in previous studies.19
It was acknowledged by the authors that the personal
history technique can be more effective with highly skilled employees than might be found in studies involving lower ability groups. The authors also admitted that some of the personal history items were similar in nature to tests of personality, interests and values and as such were not purely demographic in nature.20
Charles Adair conducted a study in 1973 to determine the correlation between the undergraduate grade point average of education majors and their success on the job as rated by their supervisors in their first year of teaching. Adair found the only significant grade point category was
3.50 4.00 on a four-point scale. The 3.00 3.49 category had the next highest correlation coefficient, while the
2.50 2.99 category had the lowest correlation coefficient with success.21
18
W. I. Smith, L. E. Albright, J. R. Glennen and W. A. Owens, "The Prediction of Research Competence and Creativity from Personal History," Journal of Applied Psychology 45: 59-62 (February 1961).
19Ibid., P. 61 20Ibid., p. 62.
22
Adair concluded that there were few significant
correlation coefficients between undergraduate grade point average and success on the job as measured by the supervisor for first year teachers. There was some evidence to suggest that the most successful teachers had the highest undergraduate grade point average.22
In a study by Langdale, two groups of personnel interviewers were given completed application blanks to judge. One group was given a general job description and the other group a complete description of the job to be filled. The group that had the complete job description achieved a higher degree of interrater reliability (.87) than the group using the nonspecific description (.35).23
Recommendations
Most studies evaluating letters of recommendation indicate that this method, used independent of other selection methods,
21Charles D. Adair, "Relationship Between Undergraduate
Grades and First Year Teaching Success," School and Community 60: 22 (January 1974).
22Ibid., p. 22.
23John A. Langdale and Joseph Weitz, "Estimating the Influence of Job Information on Interviewer Agreement," Journal of Applied Psychology 57: 23-27 (February 1973).
23
has not produced useful results.24 However, certain kinds of information from previous eiiployers may be valuable. Stone and Kendall suggested that dates of employment, salary, whether the employer would seek an appraisal and whether the former employer would rehire the applicant should be useful in evaluating the applicant.25 The researchers found that letters of recommendation carried by the applicant are generally worthless. 6
The United States Civil Service Commission recommended
that the information an employer seeks from previous employers correspond to the unique aspects of the job to be filled. They also emphasize that a letter signed by an executive can be helpful.27
An extensive study was made by Peres and Garcia in
which they analysized 625 reference letters. They synthesized 170 adjectives from the letters and sorted them into six categories:
1. Extraversion
2. Interpersonal Relations
3. Dependability-Reliability
24Bolton, p. 65.
25Harold C. Stone and William E. Kendall, Effective
Personnel Selection Procedures, (Englewood Cliffs: PrenticeHall, 1956), p. 42.
26Ibid., p. 55.
27U.S. Civil Service Commission, A Guide for Executive Selection, Personnel Methods Series, No. 13, Washington, D. C.: Government Printing Office, 1961.
24
4. Forcefulness
5. Polish
6. Physical Vim
Two hundred supervisors were then asked to describe their "best" and "poorest" employee considering all 170 adjectives. The authors found that adjectives dealing with mental agility were most discriminating, and cooperationconsideration least discriminating. In order of discrimination:
1. Mental Agility
2. Vigor
3. Dependability-Reliability
4. Urbanity
5. Cooperation-Consideration
Peres and Garcia concluded it can be hypothesized that when a person is asked to submit a letter for an applicant he feels is not tryly qualified, the person is likely to say, "Mike is a pretty nice guy." This is a practice, it might be said, whereby former employers could "damn with faint praise." The authors suggested however, that the best alternative was a forced choice scale of adjectives that would be completed by the former employer.29 Interviews
No part of the selection process is more widely
used than the personal interview.30
28Sherwood H. Peres and J. Robert Garcia, "Validity
and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnel Psychology 16: 279-86 (Autumn 1962).
29Ibid., p. 285.
30Bolton, p. 67.
25
Most people feel that the interview contributes something to the process that cannot be gained
in any other way. 31
The above statements appeared to represent what has actually taken place in the selection process in most school systems. Interviews actually serve many purposes. The interviewer can insure that the applicant understands the job, and thereby he can reduce turnovers resulting from persons accepting jobs they might have rejected if they had fully understood the requirements. Interviews are also designed to gather information on the applicant that could not be otherwise acquired.32
One study, conducted by Chruden and Sherman,supported the use of nondirective or unstructured interviews in which the applicant was allowed to talk about whatever seem important.33 In another study, Crissy stressed that structure was the only reliable form.34 He stated that the interview continued to be the most widely used personal selection method because:
31
Richard A. Fear, The Evaluation Interview, (New York: McGraw-Hill Book Company, 1958), p. 102.
32Bolton, p. 67.
33Herbert J. Chruden and Arthur W. Sherman, Junior, Readings in Personnel Management, (Cincinnati: SouthWestern Publishing Company, 1961), p. 72.
34W. J. E. Crissy, "The Employment Interview-Research Areas, Methods and Results," Personnel Psychology 5: 73-85 (Summer 1952).
26
1. The supplementary, non-evaluative functions
served by the interview,
2. An almost universal conviction on the part
of supervisory and managerial personnel that
they are good "pickers of men,"
3. The expectation on the part of job applicants
of person treatment such as is accordedin the
interview.
In referring to three studies conducted by students under his direction, Crissy identified three fundamental aspects of interview reliability:
1. Intra-rater consistency, i.e., the agreement
of the interviewer with himself insofar as
his appraisals of interviewees are concerned;
2. Inter-rater consistency, i.e., the agreement
among interviewers insofar as their respective
judgements of interviewees are concerned;
3. The consistency of behavior elicited in the
interview, i.e., the extent to which the
interviewee behaves and responds in the same
way to similar stimuli in successive interviews.
One generalization stands out. The more structured
the interview, the more reliable it becomes in relation to the above three points. The more structured it becomes, however, the more it loses the supplementary, non-evaluative nature that some employers prefer.37
35Ibid., p. 73.
Ibid., p. 74.
Ibid., p. 74.
27
Crissy suggested methods of achieving consistency in each type of interview, but in addressing himself to the validity of selecting a person that will be successful on the job, he concurred with other writers by recommending that specific traits of the job to be filled be identified and that the interviewer perform a structured interview assessing those traits. The employer must however, determine if the selection method is effective. To accomplish this task Crissy suggested:
1. The person who rated the applicant should
not rate the same person as an employee.
2. Employees selected for employment represent
the upper portion of applicants and as such
are more homogeneous by traits. This fact
limits the distribution of any selection
method.
3. The experimenter should be aware of the
inherent relationship between selection
ratios and validity coefficients.
4. The employer should evaluate the interviewers
and determine what percent of employees
hired by each interviewer ultimately become
successful employees.38
K. A. Yonge, in assessing the success of the interview, felt that the only just criticism of the interview method was based upon the practice of conducting interviews with. no definite purpose or aim. Yonge does not agree that because interviews become structured that they will necessarily be valid. Some interviews by purpose depend upon the dynamic
38Ibid., p. 79.
28
situation, or interaction that occurs in an unstructured environment. Such items as poise, motivation, sincerity or purpose, emotional balance and responsiveness can best be judged under unstructured settings.39
Yonge cited studies by Moriwaki40 and Snedden 41in
which interview assessment was compared to success on the job and correlation coefficients of .63 and .82 were achieved when rigid specifications for the interview were used, as evidence that higher validity can be achieved by structured interviews. Yonge conducted a pilot study himself which included an outline for each interviewer. He obtained correlation coefficients ranging from .48 to .99 when comparing interview assessment to job success, with relatively few subjects accounting for most of the variance. lie concluded that, properly used, the interview structured or unstructured, can play a reliable part in the overall assessment of an individual's qualities for employment.42
Several studies have been conducted to determine how
interviewers react to favorable and unfavorable information
39K. A. Yonge, "The Value of the Interview: An
Orientation and a Pilot Study," Journal of Applied Psychology 40: 25-31 (February 1956).
40E. Moriwaki, "Note on the Comparative Validities of Judgements of Intelligence Based on-Photographs and on Interviews," Journal of Applied Psychology 13: 630-631 (November 1947).
41D. Snedden, "Measuring General Intelligence by
Interview," Psychological Clinic 19: 131-134 (July 1930).
42Yonge, p. 30.
29
received in an interview. Bolster and Springbett reported in 1961 that interviewers reacted more strongly to unfavorable information about the applicant than they did to favorable information.43 The study involved over 200 officer cadet selection reports and 16 personnel officers. The design involved giving the personnel officers a hypothetical rating on a man and introducing contrary information to determine how it affected his opinion. The shifts in opinion tended to follow the patterns listed below.
1. The amount of interview commitment at
the time a decision shifts and the weight of the challenging information both affect
the amount of change made in decisions.
2. It is easier to shift a rating in the
direction of rejection than in the direction
of acceptance.
3. People have different amounts of sensitivity
to negative evidence.
4. An item of information received toward the
end of the interview carries more weight than it would if received earlier in the
interview. Thus, when information is
received affects how it is perceived. 44
Hollmann conducted a study in 1972 designed to refute the findings of Bolster and Springbett. In this study, experienced interviewers were given hypothetical applicants to rate. The information was arranged so that negative information preceeded positive information on some applicants
43B.I.Bolster and B.M.Springbett, "The Reaction
of Interviewers to Favorable and Unfavorable Information," Journal of Applied Psychology 45: 98-103 (February 1961).
Ibid., p. 103.
30
and followed positive information on other applicants. Hollmann found that per unit of importance, the interviewers gave less weight to positive information rather than more weight to negative information as Bolster and Springbett had found.45 Hollmann suggested that the reason for the apparent emphasis on negative information was due to greater feedback from supervisors on negative information than on positive information.46,
In addition to the impact of the timing of information received during an interview, interviewers should be aware of the influence of several items. Mandell reported that such items as the applicant's resemblance to someone the interviewer knows and nervousness on the part of the interviewee are sources of error related to the interviewer's attitude.47
Springbett, in a separate study, concluded that no
generalizations about the value and precision of the interview method can be made, because the interviewer is part of the process. Springbett further stated that three separate operations are being performed by the interviewer: gathering, processing and evaluating. These operations are to some
45Thomas D. Hollmann, "Employment Interviewers' Error in Processing Positive and Negative Information," Journal of Applied Psychology 56: 130-134 (April 1972).
46Ibid., p. 134.
47Milton M. Mandell, The Employment Interview, Research Study number 47, (New York: American Management Association, 1961), p. 47.
31
degree independent. An interviewer can be adept at one
operation and ineffective in the others.48
Bolton found that the interviewer feels more confident about his decision in the selection process when
he has had the opportunity to both see and hear the applicant. Bolton offered the following summary of the
value of the interview in the selection process.
1. The validity of the interview depends
upon the skill of the interviewer, the situation and type of applicant. The
assumption that the mere process of
gathering, weighing and evaluating information attained in an interview will' be an accurate one has not been' proven.
2. Interviewers need to be trained and become familiar with the common errors that
research has identified.
3. More research is needed. The part the
interview plays, the type training the
interviewer should receive, the number of interviews necessary and who makes a good
interviewer need to be decided.
4. A better evaluation of the interview as a
selection instrument needs to be made. In
addition to the evaluation of the applicants selected, a follow up on the applicants rejected must be made to fully evaluate the
interview method. Correspondingly, the evaluation of the interviewer and the recognition
that all people don't make good interviewers
must be accomplished.49
B. M. Springbett, "Factors Affecting the Final
Decision in the Employment Interview," Canadian Journal of Psychology 12: 13-22 (March 1958).
49Bolton, p. 74.
32
Synopsis of the Review of Literature
To determine whether selection methods are successful, a criterion for success on the job must be established. There has been disenchantment with the single criterion measure of success and several alternatives have been presented. The selection methods most widely used and studied include: personal history data, interviews and recommendations. The validity of these various methods has ranged from unacceptable to useful levels. Examples of successful application of the methods seemed to occur when there was an understanding of how data should be collected on employees and where the selection methods have been validated against whatever measure of success on the job the organization uses. The extent of research available, while not conclusive, did not support the random choices many school districts made on teacher selection. The research did not point toward the identification of universal guidelines, but rather toward validation of the selection methods each district chose to utilize.
CHAPTER III
PRESENTATION OF DATA
As stated in Chapter I, the problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district.
Three samples of teachers from the single urban school district were chosen for the study. Sample A consisted of teachers who had performed satisfactorily enough to be offered a third contract. Sample B contained teachers in their first year of employment in the urban school district about which no decision had been made concerning their reemployment. A third sample, Sample C, consisted of teachers who had not performed to a sufficiently satisfactory level to be retained by the urban school district.
To determine the relative usefulness of the data
gathered on the three samples the following operations were performed.
1. Frequency distributions were developed for
Samples A and B.
2. Joint frequency distributions were developed
for Samples A and B to show how the dependent 33
34
variable (evaluation assessment) related
to each of the independent variables
(personal history data, interview
assessment and recommendations assessment).
3. Regression equations were computed for
Samples A and B to determine the relative
predictive value of the independent variables.
4. Frequency distributions were developed to
compare Samples A, B and C in regard to each
of the independent variables.
The presentation of data in this chapter is contained in four sections to correspond to the four operations described above.
Basic Descriptive Data Relative to the Independent
and Dependent Variables for Samples A and B
Frequency distributions were developed on each of the nine variables for Samples A and B to provide a synthesis of the basic characteristics of the data in terms of distribution and variability. The distributions show the numerical frequency in each variable category, the percent in each variable category and the cummulative percent. In Tables 4, 7, and 8, missing data resulted in the percents becoming decimals. Obviously, the numerical frequencies remained whole numbers.
Personal History Data for Samples A and B
The following variables are included under personal history data: age, graduate of a university in Florida,
35
years of previous experience, grade point average, marriage status and race. The distribution for age is shown in Table 1. In the age category 25 years and younger, Sample A had 59 percent and Sample B.-had 69 percent, a 10 percent difference. A 10 percent difference also existed in the category over 40 years. Sample A had 12 percent, Sample B has 2 percent in this category.
The mean age for Sample A was 28.52 years as compared to a mean of 25.38 years for Sample B. This represented a difference of 3.12 years.
Table 2 presents the distribution of teachers in
Samples A and B who graduated from a university in Florida and those who graduated from a university outside the state of Florida. Sixty-two percent of the teachers in Sample A graduated from a university in Florida while 51 percent of Sample B graduated from a university in Florida.
The distribution of years of previous experience is found in Table 3. Sixty-two percent of Sample A had no previous teaching experience while 55 percent of Sample B were first year teachers. Twenty-five percent of Sample A were teachers with one to five years of experience while 37 percent of Sample B fell in this category. The mean for Sample A was 5.16 years of previous experience as compared to 3.60 for Sample B.
Table 4 contains the distribution of grade point averages based on a four-point scale. The grade point average so computed was based on all college credit recorded in the personnel file of the individual teacher. The
36
Table 1
Age Distribution for Samples A and B
Sample A Sample B
Age Cum. Cum.
Category Freq. Percent Percent Freq. Percent Percent
25 years
and under 59 59 59 69 69 69
26-30 yrs. 15 15 74 17 17 86
31-35 yrs. 10 10 84 5 5 91
36-40 yrs. 4 4 88 7 7 98
Over 40
yrs. 12 12 100 2 2 100
Total 100 100 100 100 100 100
Mean for Sample A = 28.52 years Mean for Sample B = 25.38 years
37
Table 2
Distribution of Teachers in Samples A and B Who Graduated from a University in Florida and Who Graduated from a University Outside of Florida
Sample A Sample B
Cum. CuM.
Freq. Percent Percent Freq. Percent Percent
Grad. of
Univ in
Florida 62 66 66 51 51 51
Grad. of
Univ.
outside of
Florida 32 34 100 49 49 100
Missing
Data 6
Total 100 100 100 100 100 100
38
Table 3
Distribution of Years of Previous Experience
for Samples A and B
Sample A Sample B
Yrs. of
prev. Cum. Cum.
exper. Freq. Percent Percent Freq. Percent Percent
No. prev.
exper. 62 62 62 55 55 55
1 yr. of
exper. 10 10 72 16 16 71
2-5 yrs. of
exper. 15 15 87 21 21 92
6-10 yrs.
of exper. 10 10 97 5 5 97
Over 10
yrs. of
exper. 3 3 100 3 3 100
Total 100 100 100 100 100 100
Mean for Sample A = 5.16 years Mean for Sample B = 3.60 years
39
Table 4
Distribution of Grade Point Averages for Samples A and B
Sample A Sample B
Grade
point Cum. Cum
average Freq. Percent Percent Freq. Percent Percent
2.1 and
below 2 2.5 2.5 5 5.1 5.1
2.2-2.5 18 22.2 24.7 18 18.2 23.2
2.6-3.0 37 45.7 70.4 43 43.4 66.7
3.1-3.5 23 28.4 98.8 23 28.3 94.9
3.6-4.0 1 1.2 100.0 5 5.1 100.0
Missing
Data 19 1
Total 100 100.0 100.0 100 100.0 100.0
Mean for Sample A = 2.8 Mean for Sample B = 2.8
40
difference between Sample A and Sample B for the five grade point average categories was minimal. The greatest
difference existed in categories which involved fewer than five teachers. The mean grade point average for Sample A was 2.81, for Sample B it was 2.83.
Marriage status is displayed in Table 5. The percent of married teachers who began teaching in 1971-72 (Sample A) was 60 percent. The corresponding figure for Sample B was 57 percent. Thirty-four percent of the teachers in Sample A were single, 40 percent in Sample B. The group of divorced or separated teachers represented 6 percent of Sample A and 3 percent of Sample B.
Race is presented in Table 6. There were 22 percent black teachers for Sample A and 14 percent of Sample B were black. This resulted in 78 percent of Sample A and 86 percent of Sample B having been white. Interview Assessment Data for Samples A and B
Interview assessment was represented by the numerical
score attained by assigning values to personal characteristics listed on the interview rating sheet. The interviewer rated each applicant on a continuum for each characteristic with a potential range of scores from a most favorable rating of six to a most unfavorable rating of 60. The mean score for Sample A was 22.90, Sample B's mean was 19.23. This represented a difference of 3.61. Table 7 shows the interview assessment ratings by arbitrarily assigned categories. In
41
Table 5
Distribution by Marriage Status for Samples A and B
Sample A Sample B
Cum. Cum.
Status Freq. Percent Percent Freq. Percent Percent
Married 60 60 60 57 57 57
Separated
or
Divorced 6 6 66 3 3 60
Single 34 34 100 40 40 100
Total 100 100 100 100 100 100
42
Table 6
Distribution by Race for Samples A and B
Sample A Sample B
Cum. Cum.
Race Freq. Percent Percent Freq. Percent Percent
Black 22 22 22 14 14 14
White 78 78 78 86 86 86
Total 100 100 100 100 100 100
43
Table 7
Interview Assessment Distribution for Samples A and B
Sample A Sample B
Cate- Cum. Cum.
gories Freq. Percent Percent Freq. Percent Percent
6-15 9 9.5 9.5 21 21.6 21.6
16-20 25 26.3 35.8 38 39.2 60.8
21-25 37 38.9 74.7 30 30.9 91.8
26-30 14 14.7 89.5 7 7.2 99.0
Above 30 10 10.5 100.0 1 1.0 100.0
Missing
Data 5 3
Total 100 100.0 100.0 100 100.0 100.0
Mean for Sample A = 22.90 Mean for Sample B = 19.29
44
the most favorable category, a score of 6-15, Sample A had 9.5 percent while Sample B had 21.6 percent. In the second most favorable category, Sample B was also higher, 39.2 percent to 26.3 percent. Twenty-five percent of Sample A was rated above 30, while 8.2 percent of Sample B received ratings above 30. Recommendations Assessment for Samples A and B
Recommendations assessment was the numerical value assigned by the recommending party converted to a 1-10 scale as outlined in the procedures section. The distribution of ratings is shown in Table 8. A recommendation assessment of one represented the most positive rating while an assessment of ten represented the least positive rating. Ratings of one, four and five-ten contained minimal differences between Samples A and B. Sample A had 20.9 percent with a rating of two as compared to 26.1 percent for Sample B. In the category with a recommendation assessment of three, Sample A had 34.1 percent and Sample B had 26.1 percent. The means for Sample A and Sample B were 3.29 and 3.35 respectfully. Evaluation Assessment for Samples A and B
Evaluation assessment was used as the dependent
variable in the study. The distribution for Sample A is found in Table 10 and the distribution for Sample B appears
45
Table 8
Recommendations Assessment Distribution
for Samples A and B
Sample A Sample B
Cate- Cum. Cum.
gories Freq. Percent Percent Freq. Percent Percent
1 8 8.8 8.8 6 6.5 6.5
2 19 20.9 29.7 24 26.1 32.8
3 31 34.1 63.7 24 26.1 58.7
4 13 14.3 78.0 13 14.1 72.8
5-10 20 22.0 100.0 25 27.2 100.0
Missing
Data 9 8
Total 100 100.0 100.0 100 100.0 100.0
Mean for Sample A = 3.29 Mean for Sample B = 3.35
46
Table 9
Evaluation Assessment Distribution for Sample B
Sample B
Cum.
Categories Freq. Percent Percent
1.0 72 72 72
1.1 15 15 87
1.2 6 6 93
1.3 2 2 95
1.4-3.0 5 5 100
Total 100 100 100
Mean for Sample B = 1.05
Pages
are
misnumbered
following
this
insert
46
Table 10
Evaluation Assessment Distribution for Sample A
Sample A
Cum.
Categories Freq. Percent Percent
1.8 and below 20 20 20
1.9-2.0 21 21 41
2.1-2.2 29 29 70
2.3-2.5 20 20 90
2.6-5.0 10 10 100
Total 100 100 100
Mean for Sample A = 2.12
47
in Table 9. The evaluation assessment for Sample A represented the mean score of three annual evaluations for each teacher as completed by the principal. The evaluation assessment for Sample B represented the mean of one annual evaluation as completed by the principal. The annual evaluation for Sample B contained three choices for rating the teachers: satisfactory, needs improvement and unsatisfactory. This scale produced a smaller distribution of scores and necessitated using separate tables for Samples A and B. In both samples the lower the numerical value the more positive the evaluation.
The evaluation assessment for Sample A were distributed somewhat evenly in each category: 20 percent, 21 percent, 29 percent, 20 percent and 1-0 percent respectfully. In Sample B, 72 percent of the teachers were in the most positive (lowest) category.
Independent Variable Data in Relationship to the
Dependent Variable for Samples A and B
Joint frequency distributions were established to show
how the dependent variable related to the independent variables. Separate tables were constructed to show each independent variable for Sample A related to the dependent variable and each independent variable for Sample B related to the dependent variable. Tables 11 through 26 showing these data follow the same format:
1. The five categories of the dependent variable
(evaluation assessment) are shown on the left
of each table.
48
2. The categories of each independent variable
are shown across the top of each table.
3. Each cell within each tables shows:
a. The actual number of teachers,
(see b in:each table).:
b. The percent of teachers in an evaluation
category contained in each independent
variable category (see c in each table).
c. The percent of teachers in each independent
variable category who were in each evaluation category (see d in each table).
d. The percent of the total sample
(see e in each table).
4. The numerical total and percent of the total
sample in each evaluation category (see f in
each table).
5. The numerical total and percent of the total
sample in each independent variable category
(see g in each table).
Personal History Data in Relationship to Evaluation Assessment
The independent variables included in personal history data were: age, graduate of a university in Florida, years of previous experience, grade point average, marriage status and race. In Table 11, age is related to evaluation assessment for Sample A. Fifty-five percent of all teachers who received the most positive evaluation assessment (1.8 and below), were under 26 years old (see c, Table 11). Fifty-nine percent of all teachers in Sample A were under 26 years old (see g, Table 11). The over 40 category
49
Table 11
Age Related to Evaluation Assessment for Sample A
Eval. Age
cate
gory Under 26 26-30 31-35 36-40 Over 40 Row Total
1.8 11.0 b 2.0 2.0 1.0 4.0 20.0
and 55.0 c 10.0 10.0 5.0 20.0 20.0
below a 18.6 d 13.3 20.0 25.0 33.3
11.0 e 2.0 2.0 1.0 4.0
13.0 3.0 2.0 1.0 2.0 21.0
1.9- 61.9 14.3 9.5 4.8 9.5 21.0
2.0 22.0 20.0 20.0 25.0 16.7
13.0 3.0 2.0 1.0 2.0
19.0 3.0 3.0 1.0 3.0 29.0
2.1- 65.5 10.3 10.3 3.4 10.3 29.0
2.2 32.2 20.0 30.0 25.0 25,0
19.0 3.0 3.0 1.0 3.0
10.0 5.0 2.0 1.0 2.0 20.0
2.3- 50.0 25.0 10.0 5.0 10.0 20.0
2.5 16.9 33.3 20.0 25.0 16.7
10.0 5.0 2.0 1.0 2.0
6.0 2.0 1.0 0.0 1.0 10.0
2.6- 60.0 20.0 10.0 0.0 10.0 10.0
5.0 10.2 13.3 10.0 0.0 8.3
6.0 2.0 1.0 0.0 1.0
Column 59.0 15.0 10.0 4.0 12.0 100.0
Total 9 59.0 15.0 10.0 4.0 12.0 100.0
a. The lower the numerical value, the more positive
the rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
50
Table 12
Age Related to Evaluation Assessment for Sample B
Eval. Age
cate
gory Under 26 26-30 31-35 36-40 Over 40 Row Total
50.0 b 14.0 3.0 4.0 1.0 72.0
1.0 a 69.4 c 19.4 4.2 5.6 1.4 72.0
72.5 d 82.4 60.0 57.1 50.0 50.0 e 14.0 3.0 4.0 1.0
10.0 3.0 1.0 0.0 1.0 15.0
1.1 66.7 20.0 6.7 0.0 6.7 15.0
14.5 17.6 20.0 0.0 50.0
10.0 3.0 1.0 0.0 1.0
3.0 0.0 1.0 2.0 0.0 6.0
1.2 50.0 0.0 16.7 33.3 0.0 6.0
4.3 0.0 20.0 28.6 0.0
3.0 0.0 1.0 2.0 0.0
2.0 0.0 0.0 0.0 0.0 2.0
1.3 100.0 0.0 0.0 0.0 0.0 2.0
2.9 0.0 0.0 0.0 0.0
2.0 0.0 0.0 0.0 0.0
4.0 0.0 0.0 1.0 0.0 5.0
1.4- 75.0 0.0 0.0 25.0 0.0 5.0
1.5 5.7 0.0 0.0 14.3 0.0
4.0 0.0 0.0 1.0 0.0
Column 69.0 17.0 5.0 7.0 2.0 100.0
Total 9 69.0 17.0 5.0 7.0 2.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
51
represented 12 percent of the sample and received 20 percent of the ratings in the most positive (1.8 and below) category.
In Sample B found in Table 12, 69 percent of all
teachers were in the under 26 age category and 72 percent of all teachers under 26 received the most positive rating (1.0).
Tables 13 and 14 show graduates of universities in
Florida-,and outside Florida related to evaluation assessment for Samples A and B respectfully. As noted previously (see Table 2), Sample A contained 66 percent Florida graduates, while Sample B had 21 percent. In Sample A, Florida graduates (representing 66 percent of the teachers) accounted for 47.1 percent of the most positive ratings. In Sample B, Florida graduates (representing 51 percent of the teachers) accounted for 51.4 percent of the most positive ratings. In Sample A, 30.6 percent of all Florida graduates were rated 2.0 or lower. The percent of graduates of out of state schools who received a 2.0 or lower rating was 56.1 percent.
Previous teaching experience is shown related to
evaluation assessment in Tables 15 and 16. In examining Sample A (Table 15) it is found that the category none (no previous experience) contained 62 percent of the sample and 45 percent of the most positive ratings (1.8 and below). The category 6-10 years contained 10 percent of the sample and received 25 percent of the most positive ratings.
52
Table 13
Graduates of Universities in Florida and Graduates
of Universities Outside of Florida Related
to Evaluation Assessment for Sample A
Grad. Grad. of
Evaluation Grd. University Row
Category outside total f
in Fla. Fa
Fla.
8.0 b 9.0 17.0
1.8 and 47.1 c 52.9 18.1
below a 12.9 d 28.1
8.5 e 9.6
11.0 9.0 20.0
1.9-2.0 55.0 45.0 21.3
17.7 28.1
11.7 9.6
22.0 7.0 29.0
2.1-2.2 75.9 24.1 30.9
35.5 21.9
23.4 7.4
16.0 3.0 19.0
2.3-2.5 84.2 15.8 20.2
25.8 9.4
17.0 3.2
5.0 4.0 9.0
2.6-5.0 55.6 44.4 9.6
8.1 12.5
5.3 4.3
Column 62.0 32.0 94.0
Total 9 66.0 34.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
53
Table 14
Graduates of Universities in Florida and Graduates
of Universities Outside of Florida Related
to Evaluation Assessment for Sample B
Grad. Grad. of
Evaluation University University Row
Category in Fla outside total
Fla.
37.0 b 35.0 72.0
1.0 a 51.4 c 48.6 72.0
72.5 d 71.4
37.0 e 35.0
6.0 9.0 15.0
1.1 40.0 60.0 15.0
11.8 18.8
6.0 -9.0
3.0 3.0 6.0
1.2 50.0 50.0 6.0
5.9 6.1
3.0 3.0
2.0 0.0 2.0
1.3 100.0 0.0 2.0
3.9 0.0
2.0 0.0
3.0 2.0 5.0
1.4-1.5 60.0 40.0 5.0
5.9 4.0
3.0 2.0
Column 51.0 49.0 100.0
Total 9 51.0 49.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluE.tion category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
54
Table 15
Years of Previous Experience Related to
Evaluation Assessment for Sample A
Eval. Years of Previous Experience
cate- Over Row
gory None One 2-5 6-10 10 total
1.8 9.0 b 2.0 3.0 5.0 1.0 20.0
and 45.0 c 10.0 15.0 25.0 5.0 20.0
below a 14.5 d 20.0 20.0 50.0 33.3
9.0 e 2.0 3.0 5.0 1.0
13.0 3.0 2.0 3.0 0.0 21.0
1.9- 61.9 14.3 9.5 14.3 0.0 21.0
2.0 21.0 30.0 13.3 30.0 10.0
13.0 3.0 2.0 3.0 0.0
18.0 2.0 7.0 0.0 2.0 29.0
2.1- 62.1 6.9 24.1 0.0 6.9 29.0
2.2 29.0 20.0 46.7 0.0 66.7
18.0 2.0 7.0 0.0 2.0
15.0 2.0 2.0 1.0 0.0 20.0
2.3- 75.0 10.0 10.0 5.0 0.0 20.0
2.5 24.2 20.0 13.3 10.0 0.0
15.0 2.0 2.0 1.0 0.0
7.0 1.0 1.0 1.0 0.0 10.0
2.6- 70.0 10.0 10.0 10.0 0.0 10.0
5.0 11.3 10.0 6.7 10.0 0.0
7.0 1.0 1.0 1.0 0.0
Column 62.0 10.0 15.0 10.0 3.0 100.0 Total 9 62.0 10.0 15.0 10.0 3.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
55
Table 16
Years of Previous Experience Related to
Evaluation Assessment for Sample B
Years of Previous Experience Eval.Over Row
Category None One 2-5 6-10 10 Total
38.0 b 12.0 17.0 2.0 3.0 72.0
1.0 a 52.8 c 16.7 23.6 2.8 4.2 72.0
69.1 d 75.0 81.0 40.0 100.0 38.0 e 12.0 17.0 2.0 3.0
9.0 2.0 3.0 1.0 0.0 15.0
1.1 60.0 13.3 20.0 6.7 0.0 15.0
16.4 12.5 14.3 20.0 0.0
9.0 2.0 3.0 1.0 0.0
3.0 1.0 1.0 1.0 0.0 6.0
1.2 50.0 16.7 16.7 16.7 .0.0 6.0
5.5 6.3 4.8 20.0 0.0 3.0 1.0 1.0 1.0 0.0
2.0 0.0 0.0 0.0 0.0 2.0
1.3 100.0 0.0 0.0 0.0 0.0 2.0
3.6 0.0 0.0 0.0 0.0
2.0 0.0 0.0 0.0 0.0
3.0 1.0 0.0 1.0 0.0 5.0
1.4-1.5 40.0 20.0 0.0 20.0 0.0 5.0
5.4 6.3 0.0 20.0 0.0 3.0 1.0 0.0 1.0 0.0
Column 55.0 16.0 21.0 5.0 3.0 100.0
Total 9 55.0 16.0 21.0 5.0 3.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
56
The other categories contained minimal differences. In Sample B, shown in Table 16, 55 percent of all teachers had no previous experience and the category accounted for 51 percent of the most positive evaluations (1.0).
Tables 17 and 18 show grade point average related
to evaluation assessment. Grade point average was based upon a four-point scale and computed on all college credit recorded in the personnel file of the individual teacher. In examining the various categories of grade point average for Sample A, it is noted that 28.3 percent of all teachers achieved a grade point average of 3.1-3.5. This category accounted for 26.7 percent of the teachers who received the most favorable evaluations (1.8 and below). In the category 2.2-2.5, which represented 22.2 percent of all teachers in Sample A, 26.7 percent received ratings in the evaluation category 1.8 and below. The results for Sample B were very similar. The category of 3.1-3.5 represented 28.3 percent of the sample and received 26.8 percent of the most positive ratings (1.0). The 2.2-2.5 category represented 18.2 percent of the sample and received 14.1 percent of the most positive ratings.
Marriage status related to evaluation assessment for Samples A and B is displayed in Tables 19 and 20 respectfully. In Sample A married teachers represented 60 percent of the sample and 65 percent of the most positive ratings (1.8 and below). Single teachers accounted for 34 percent
57
Table 17
Grade Point Average Related to Evaluation
Assessment for Sample A
Eval. Grade Point Average
cate- Below 2.2- 2.6-- 3.1- 3.6- Row
gory 2.2 2.5 3.0 3.5 4.0 Total
1.0 b 4.0 6.0 4.0 0.0 15.0
1.8 and 6.7 c 26.7 40.0 26.7 0.0 18.5
below a 50.0 d 22.2 16.2 17.4 0.0
1.2 e 4.9 7.4 4.9 0.0
0.0 5.0 7.0 5.0 1.0 18.0
1.9- 0.0 27.8 38.9 27.8 5.6 22.2
2.0 0.0 27.8 18.9 21.7 100.0
0.0 6.2 8.6 6.2 1.2
1.0 6.0 10.0 7.0 0.0 24.0
2.1- 4.2 25.0 41.7 29,2 0.0 29.6
2.2 50.0 33.3 27.0 30.4 0.0
1.2 7.4 12.3 8.6 0.0
0.0 3.0 6.0 6.0 0.0 15.0
2.3- 0.0 20.0 40.0 40.0 0.0 18.5
2.5 0.0 16.7 16.2 26.1 0.0
0.0 3.7 7.4 7.4 0.0
0.0 0.0 8.0 1.0 0.0 9.0
2.6- 0.0 0.0 88.9 11.1 0.0 11.1
5.0 0.0 0.0 21.6 4.3 0.0
0.0 0.0 9,9 1.2 0.0
Column 2.0 18.0 37.0 23.0 1.0 81.0
total 9 2.5 22.2 45.7 28.3 1.2 100.0
Missing Data 19
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
58
Table 18
Grade Point Average Related to Evaluation
Assessment for Sample B
Eval. Grade Point Average
cate- Below 2.2- 2.6- 3.1- 3.6- Row
gory 2.2 2.5 3.0 3.5 4.0 Total
5.0 b 10.0 33.0 19.0 4.0 71.0
1.0 a 7.0 c 14.1 46.5 26.8 5.6 71.7
100.0 d 55.6 76.8 67.9 80.0
5.1 e 10.1 33.3 19.2 4.0
0.0 3.0 6.0 6.0 0.0 15.0
1.1 0.0 20.0 40.0 40.0 0.0 15.2
0.0 16.7 14.0 21.4 0.0
0.0 3.0 6.1 6.1 0.0
0.0 3.0 2.0 1.0 0.0 6.0
1.2 0.0 50.0 33.3 16.7 0.0 6.1
0.0 16.7 4.7 3.6 0.0
0.0 3.0 2.0 1.0 0.0
0.0 1.0 0.0 1.0 0.0 2.0
1.3 0.0 50.0 0.0 50.0 0.0 2.0
0.0 5.6 0.0 3.6 0.0
0.0 1.0 0.0 1.0 0.0
0.0 1.0 2.0 1.0 1.0 5.0
1.4- 0.0 20.0 40.0 20.0 20.0 5.0
1.5 0.0 5.6 4.7 3.6 20.0
0.0 1.0 2.0 1.0 1.0
Column 5.0 18.0 43.0 28.0 5.0 99.0
Total 9 5.1 18.2 43.4 28.3 5.1 100.0
Missing Data 1
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
59
Table 19
Marriage Status Related to Evaluation Assessment for Sample A
Eval. Marriage Status
Cate- Divorced or Row
gory Married Separated Single Total
13.0 b 2.0 5.0 20.0
1.8 and 65.0 c 10.0 25.0 20.0
below a 21.7 d 33.3 14.7
13.0 e 2.0 5.0
13.0 1.0 7.0 21.0
1.9- 61.9 4.8 33.3 21.0
2.0 21.7 16.7 20.6
13.0 1.0 7.0
16.0 2.0 .11.0 29.0
2.1- 55.2 6.9 37.9 29.0
2.2 26.7 33.3 32.4
16.0 2.0 11.0
13.0 1.0 6.0 20.0
2.3- 65.0 5.0 30.0 20.0
2.5 21.7 16.7 17.6
13.0 1.0 6.0
5.0 0.0 5.0 10.0
2.6- 50.0 0.0 50.0 10.0
5.0 8.3 0.0 14.7
5.0 0.0 5.0
Column 60.0 6.0 34.0 100.0
Total g 60.0 6.0 34.0 100.0
A. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
60
Table 20
Marriage Status Related to Evaluation
Assessment for Sample B
Eval. Marriage Status
Cate- Divorced or Row
gory Married Separated Single Total
45.0 b 3.0 24.0 72.0
1.0 a 62.5 c 4.2 33.3 72.0
78.9 d 100.0 60.0
45.0 e 3.0 24.0
7.0 0.0 8.0 15.0
1.1 46.7 0.0 53.3 15.0
12.3 0.0 20.0
7.0 0.0 8.0
4.0 0.0 2.0 6.0
1.2 66.7 0.0 '33.3 6.0
7.0 0.0 5.0
4.0 0.0 2.0
0.0 0.0 2.0 2.0
1.3 0.0 0.0 100.0 2.0
0.0 0.0 5.0
0.0 0.0 2.0
1.0 0.0 4.0 5.0
1.4- 16.7 0.0 83.4 5.0
1.5 .1.8 0.0 19.0 5.0
1.0 0.0 4.0
Column 57.0 3.0 40.0 100.0
Total g 57.0 3.0 40.0 100.0
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. C. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
61
of the sample and received 25 percent of the most positive ratings. The same categories for Sample B revealed that 57 percent of the sample were married and that married teachers received 62 percent of the most positive ratings. Forty percent of Sample B consisted of single teachers and this group received 33.3 percent of the most positive ratings. In examining the least positive evaluations as related to marriage status, married teachers were given 30 percent of the ratings in the two least positive evaluation categories. Thirty-two percent of the single teachers were placed in the same two categories.
The last variable in personal history data, race,
is related to evaluation assessment in Tables 21 and 22. In Sample A, 78 percent of the teachers were white and this group received 75 percent of the most positive evaluations (1.8 and below). The 22 percent black teachers thus received 25 percent of the most positive ratings. Similarily, in Sample B, the white teachers represented 86 percent of the sample and received 86 percent of the most positive ratings while the 14 percent black teachers received 14 percent of the most positive ratings.
Interview Assessment Related to Evaluation Assessment
Interview assessment represented the numerical score
attained by assigning values to the interview rating sheet.
62
Table 21
Race Related to Evaluation Assessment for Sample A
Eval. Race
Cate- Row
gory Black White Total
5.0 b 15.0 20.0
1.8 and 25.5 c 75.0 20.0
below a 22.7 d 19.2
5.0 e 15.0
3.0 18.0 21.0
1.9- 14.3 85.7 21.0
2.0 13.6 23.1
3.0 18.0
5.0 24.0 29.0
2.1- 17.2 82.8 29.0
2.2 22.7 30.8
5.0 24.0
8.0 12.0 20.0
2.3- 40.0 60.0 20.0
2.5 36.4 15.4
8.0 12.0
1.0 9.0 10.0
2.6- 10.0 90.0 1C.0
5.0 4.5 11.5
1.0 9.0
Column 22.0 78.0 100.0
Total g 22.0 78.0 100.0
a. The lower the numerical value, the more positive
the rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
63
Table 22
Race Related to Evaluation Assessment for Sample B
Eval. Race
cate- Row
gory Black White Total
10.0 b 62.0 72.0
1.0 a 13.9 c 86.1 72.0
71.4 d 72.1
10.0 e 62.0
1.0 14.0 15.0
1.1 6.7 93.3 15.0
7.1 16.3
1.0 14.0
1.0 5.0 6.0
1.2 16.7 83.3 6.0
7.1 5.8
1.0 5.0
0.0 2.0 2.0
1.3 0.0 100.0 2.0
0.0 2.3
0.0 2.0
2.0 3.0 5.0
1.4- 40.0 60.0 5.0
1.5 14.3 3.5
2.0 3.0
Column 14.0 86.0 100.0
Total g 14.0 86.0 100.0
a. The lower the numerical value, the more positive
the rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
64
These scores ranged from a possible low of six to a possible high of 60, with the low score representing the more positive assessment. In Table 23, the relationship between interview assessment and evaluation assessment for Sample A is displayed. The interview category 0-15 represented 9.5 percent of the sample and received 15.8 percent of the most positive ratings. The least positive interview assessment category 31-60, represented 10.5 percent of the sample and this group received 10.5 percent of the most positive ratings. A comparison of the same data for Sample B in Table 24, reveals the most positive interview category (0-15) represented 21.6 percent of the sample and this group received 23.2 percent of the most positive evaluation ratings (1.0). The least positive interview category (31-60) represented 1.0 percent of the sample and this group received 1.4 percent of the most positive ratings.
Recommendations Assessment in Relationship to Evaluation Assessment
Recommendations assessment related to evaluation
assessment for Sample A is shown in Table 25. The same variables for Sample B appear in Table 26. In Sample A, of those teachers who received the most positive recommendations assessment (1), 37.5 percent received the most positive evaluation rating (1.8 and below). Of those teachers who received the second most positive recommendations
65
Table 23
Interview Assessment Related to Evaluation
Assessment for Sample A
Eval. Interview Assessment
Cate- Row
gory 0-15 16-20 21-25 26-30 31-60 Total
3.0 b 5.0 6.0 3.0 2.0 19.0
1.8 and 15.8 c 26.3 31.6 15.8 10.5 20.0
below a 33.3 20.0 16.2 21.4 20.0
3.2 5.3 6.3 3.2 2.1
1.0 9.0 5.0 2.0 2.0 19.0
1.9- 5.3 47.4 26.3 10.5 10.5 20.0
2.0 11.1 36.0 13.5 14.3 20.0
1.1 9.5 5.3 2.1 2.1
4.0 5.0 11.0 4.0 3.0 27.0
2.1- 14.8 18.5 40.7 14.8 11.1 28.4
2.2 44.4 20.0 29.7 28.6 30.0
4.2 5.3 11.6 4.2 3.2
0.0 4.0 10.0 3.0 3.0 20.0
2.3- 0.0 20.0 50.0 15.0 15.0 21.1
2.5 0.0 16.0 27.0 21.4 30.0
0.0 4.2 10.5 3.2 3.2
1.0 2.0 5.0 2.0 0.0 10.0
2.6- 10.0 20.0 50.0 20.0 0.0 10.5
5.0 11.1 8.0 13.5 14.3 0.0
1.1 2.1 5.3 2.1 0.0
Column 9.0 25.0 37.0 14.0 10.0 95.0
Total 9 9.5 26.3 38.9 14.7 10.5 100.0
Missing Data 5
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
66
Table 24
Interview Assessment Related to Evaluation
Assessment for Sample B
Eval. Interview Assessment Cate- Row
gory 0-15 16-20 21-25 26-30 31-60 Total
16.0 b 27.0 20.0 5.0 1.0 69.0
1.0 a 23.2 c 39.1 29.0 7.2 1.4 71.1
76.2 d 71.1 66.7 71.4 100.0
16.5 e 27.8 20.6 5.2 1.0
4.0 7.0 3.0 1.0 0.0 15.0
1.1 26.7 46.7 20.0 6.7 0.0 15.5
19.0 18.4 10.0 14.3 0.0
4.1 7.2 3.1 1.0 0.0
0.0 2.0 4.0 0.0 0.0 6.0
1.2 0.0 33.3 66.7 0.0 0.0 6.2
0.0 5.3 13.3 0.0 0.0
0.0 2.1 4.1 0.0 0.0
1.0 1.0 0.0 0.0 0.0 2.0
1.3 50.0 50.0 0.0 0.0 0.0 2.1
4.8 2.6 0.0 0.0 0.0
1.0 1.0 0.0 0.0 0.0
0.0 1.0 3.0 1.0 0.0 5.0
1.4- 0.0 20.0 60.0 20.0 0.0 5.1
1.5 0.0 2.6 10.0 14.3 0.0
0.0 1.0 3.1 1.0 0.0
Column 21.0 38.0 30.0 7.0 1.0 97.0
Total 9 21.6 39.2 30.9 7.2 1.0 100.0
Missing Data 3
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
67
Table 25
Recommendations Assessment Related to Evaluation
Assessment for Sample A
Eval. Recommendations Assessment
Cate- Powf
gory 1 2 3 4 5-10 Total
3.0 b 6.0 5.0 0.0 4.0 18.0
1.8 and 16.7 c 33.3 27.8 0.0 22.2 19.8
below a 37.5 d 31.6 16.1 0.0 20.0
3.3 e 6.6 5.5 0.0 4.4
0.0 5.0 7.0 5.0 3.0 20.0
1.9- 0.0 25.0 35.0 25.0 15.0 22.0
2.0 0.0 26.3 22.6 38.5 15.0
0.0 5.5 7.7 5.5 3.3
2.0 4.0 11.0 5.0 5.0 27.0
2.1- 7.4 14.8 40.7 18.5 18.5 29.7
2.2 25.0 21.1 35.5 38.5 25.0
2.2 4.4 12.1 5.5 5.5
3.0 3.0 4.0 1.0 7.0 18.0
2.3- 16.7 16.7 22.2 5.6 38.9 19.8
2.5 37.5 15.8 12.9 7.7 35.0
3.3 3.3 4.4 1.1 7.7
0.0 1.0 4.0 2.0 1.0 8.0
2.6- 0.0 12.5 50.0 25.0 12.5 8.8
5.0 0.0 5.3 12.9 15.4 5.0
0.0 1.1 4.4 2.2 1.1
Column 8.0 19.0 31.0 13.0 20.0 91.0
Total 9 8.8 20.9 34.1 14.3 22.0 100.0
Missing Data 9
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
68
Table 26
Recommendations Assessment Related to Evaluation Assessment for Sample B
Eval. Recommendations Assessment
Cate- Row
gory 1 2 3 4 5-10 Total
4.0 b 18.0 17.0 9.0 16.0 64.0
1.0 a 6.3 c 28.1 26.6 14.1 25.0 69.6
66.7 d 75.0 70.8 69.2 64.0
4.3 e 19.6 18.5 9,8 17.4
1.0 5.0 2.0 1.0 6.0 15.0
1.1 6.7 33.3 13.3 6.7 40.0 16.3
16.7 20.8 8.3 7.7 24.0
1.1 5.4 2.2 1.1 6.5
1.0 1.0 1.0 2.0 1.0 6.0
1.2 16.7 16.7 16.7 '33.3 16.7 6.5
16.7 4.2 4.2 15.4 4.0
1.1 1.1 1.1 2.2 1.1
0.0 0.0 2.0 0.0 0.0 2.0
1.3 0.0 0.0 100.0 0.0 0.0 2.2
0.0 0.0 8.3 0.0 0.0
0.0 0.0 2.2 0.0 0.0
0.0 0.0 2.0 1.0 2.0 5.0
1.4- 0.0 0.0 40.0 20.0 40.0 5.4
1.5 0.0 0.0 8.3 7.7 8.0
0.0 0.0 2.2 1.1 2.2
Column 6.0 24.0 24.0 13.0 25.0 92.0
Total 9 6.5 26.1 26.1 14.1 27.2 100.0
Missing Data 8
a. The lower the numerical value, the more positive the
rating.
b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained
in this independent variable category.
d. Percent of those in this independent variable category
who were in this evaluation category.
e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this
evaluation category.
g. Numerical total and percent of total sample in this
independent variable category.
69
(2), 31.6 percent received the most positive evaluation rating. In the entire sample, 19.8 percent received ratings in the most positive category. Teachers in the two most positive recommendations categories (1,2) represented 29.7 percent of the sample. These teachers received 50 percent of the most favorable evaluation ratings.
Teachers in Sample A who received the least positive (5-10) recommendations represented 22 percent of the sample. Teachers in that category (5-10) received 22 percent of the most positive evaluations (1.8 and below). In Sample B, as found in Table 26, 64 percent of the teachers with the least positive recommendations assessment received evaluation ratings in the most positive category (1.0). The most positive recommendations assessment category (1.0), contained only six teachers, but four (66.7 percent) achieved the most positive evaluation rating (1.8 and below). Inthe second most positive recommendations assessment category (2.0), 75.0 percent of the teachers fell in the most positive evaluation category. The remaining recommendations categories (3,4) had 70.8 percent and 69.2 percent, respectfully, in the most positive evaluation categories.
Results of Regression Analysis for Samples A and B
Multiple regression is an extension of the use of the bivariate correlation coefficient to multivariate analysis.
70
Multiple regression permits the study of the linear relationships between a set of independent variables and a dependent variable while taking into account the interrelationships among the independent variables. The basic purpose of multiple regression is to produce a linear combination of independent variables which will correlate as highly as possible with the dependent variable. The linear combination can then be used to predict the values of the dependent variable. Regression equations were established for Sample A and Sample B incorporating the following independent variables: age, graduate of a university in Florida, years of previous experience, grade point average, marriage status, race, interview assessment and recommendation assessment.
The summary tables for the regression equations
present the following information: multiple R, R-square, R-square change, B and Beta. Multiple R indicates the linear relationship between the independent variables and the dependent variable under investigation. R-square can be interpreted as the proportion of the variance in the dependent variable accounted for by the regression equation. R-square change is the change in R-square from the value of the previous step. The B and Beta values are the
1William C. Mitchell, "Multiple-Regression Analysis: Subprogram Regression," in Statistical Package for the Social Sciences, edited by Norman H. Nie, Dale H. Bent and C Hadlai Hull, (New York: McGraw-Hill Book Company, 1970), p. 175-95.
71
regular and normalized regression coefficients, respect2
fully2. The summary table and prediction equation for Sample A is shown in Table 27. The correlation coefficients listed under multiple R represent the intercorrelation of the independent variables included to that point as well as the correlation between the independent variable and the dependent variable. The last correlation coefficient, .25153, represents the strength between the dependent variable and the eight independent variables taken together. Rsquare, the proportion of the variance in the dependent variable accounted for by the regression equation is found to be .05834 percent. Multiple R for Sample B appears in Table 28. The total multiple R for Sample B is .31898 The proportion of variance accounted for in the dependent variable for Sample B is .10175.
The prediction equation for Sample A appears below Table 27 while the prediction equation for Sample B is below Table 28. The prediction equation is designed to predict the value of the dependent variable, evaluation assessment. A practical example is included to help in understanding the regression equation. A hypothetical applicant from Sample B is used, with the following characteristics: 24 years old, a graduate of the University of Florida, no previous experience, a grade point average of 3.21, single, white, an interview assessment of 19 and a
2Ibid., p. 176.
72
recommendations assessment of two. Evaluation Assessment X. = (.01968) (24) + (-.06769) (1) + (-.001099) (0) + (0.0Y099) (3.21) + (.46941) (2) + (-.37482) (2) + (.02097) (19) + (.09234) (2) + 9.2091.3
The actual value of the variable is used in the
equation except for: graduate of a university in Florida where a graduate equals one and a non-graduate equals two, marriage status where married equals one and single, divorced or separated equals two, and race where black equals one and white equals two. The 9.20913 is a constant as shown in the table. The predicted evaluation assessment for the hypothetical teacher would 1.3913 whereas the mean for Sample B is 1.05.
A Comparison of the Distribution of the Independent
Variables for Samples A, B, and C
This section presents the comparison of the distribution of the independent variables: personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status and race), interview assessment and recommendations assessment, for Samples A, B, and C. Sample A consisted of 100 teachers, randomly chosen, who were hired in 1971-72 and who were in continuous employment in the urban school district to January, 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the urban school district. Sample B consisted of 100 teachers, randomly chosen, who were initially
Table 27
Multiple Regression Analysis for Evaluation Assessment for Sample A
RZ-Proportion B Normalized
of variance 2 Regression Regression
Variable Multiple R accounted for R Change Coefficient Coefficient
X1 Age 0.08423 0.00709 0.00709 0.00125 0.00363
X2 Graduate of Univ. 0.13970 0.01952 0.01242 -0.69066 -0.11617
in Florida
X3 Yrs. of Prev. Ex. 0.16047 0.02575 0.00624 -0.08973 -0.11589
X4 Grade Pt. Ave. 0.16610 0.02759 0.00184 0.01238 0.04282
X5 Marriage Status 0.17308 0.02996 0.00237 0.53363 0.078677
X6 Race 0.17359 0.03013 0.00018 0.34061 0.04246
X7 Interview Assess. 0.22557 0.05088 0.02075 0.06705 0.15803
X Recommendations 0.24153 0.05834 0.00745 -0.17113 -0.08830
8 Assessment
Constant 19.70982
Prediction Equation A
X= (0.00125)X1 + (-0.69066)X2 + (-0.08973)X3 + (0.01238)X4 + (0.53363)X5 + (0.34061)X6 + (0.06705)X7 + (-0.'7113)X8 + 19.70982
Table 28
Multiple Regression Analysis for Evaluation Assessment for Sample B
R2-proportion B Normalized
of variance 2 Regression Regression
Variable Multiple R accounted for R Change Coefficient Coefficient
Xl Age 0.02029 0.00041 0.00041 0.01968 0.10364
X Graduate of Univ. 0.05175 0.00268 0.00227 -0.06769 -0.03131
in Florida
X3 Yrs. of Prev. Ex. 0.10544 0.01112 0.00844 -0.01754 -0.04703
X4 Grade Pt. Ave. 0.10632 0.01130 0.00019 0.01099 0.05366
X5 Marriage Status 0.24694 0.06098 0.04967 0.46941 0.21500
X6 Race 0.26961 0.07269 0.01171 -0.37482 -0.12032
X7 Interview Assess. 0.28889 0.08346 0.01077 0.02097 0.11082
X Recommendations 0.31898 0.10175 0.01830 0.09234 0.13886
Assessment
Constant 9.20913
Prediction Equation
X 9= (0.01968)X1 + (-0.06769)X2 + (-0.01754)X3 + (0.01099)X4 + (0.4694')X5 + (-0.37482)X6 + (0.02097)X7 + (0.09234)X8 + 9.20913
75
employed in 1973-74 by the urban school district and who were in continuous employment to January, 1974. This sample constituted the sample about whom no decision had been made regarding their continued employment in the urban school district. Sample C consisted of 21 arbitrarily chosen teachers, who had not performed at a sufficiently satisfactory level to be retained in the urban school district and at the time of the investigation were no longer employed. The teachers in this sample were employed in the urban school district for varying lengths of time during the period 1971-73.
In Tables 29-36, the comparisons of the three samples on each of the independent variables by frequency percents are shown. Missing data in Sample C has rendered the comparison on grade point average relatively useless.
The chi-square test of independence was performed on all independent variables except grade point average. The purpose of the chi-square test is to test the null hypothesis: Samples A, B, and C are from the same population. The chi-square obtained for each sample and the chi-square from a table in Guilford, at the probability level of .01 for the appropriate degrees of freedom, is shown in each table.3 A chi-square that is greater than that found in the table at the .01 level, provides th' basis for rejecting the null hypothesis of no difference in the three samples. A chisquare less than that found in the table in Guilford permits
3J. P. Guilford, Fundamental Statistics in Psychology
and Education, (New York: McGraw-Hill Book Co., 1956) p. 145.
76
acceptance of the null hypothesis.4
The variables age, graduate of a university in
Florida, years of previous experience and marriage status contain chi-squares sufficiently low to accept the null hypothesis. The variables race, interview assessment and recommendations assessment have chi-squares large enough to reject the null hypothesis.
The comparison of age distribution is found in Table 29. In the age category 25 and under, Sample A had 59.0 percent, Sample B had 69.0 percent and Sample C had 38.1 percent. In the category 40 and over, Sample A had 12.0 percent while Sample B and Sample C had 2.0 percent and 19.0 percent respectfully. The value of chi-square computed on these distributions was 15.70, which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted.
Table 30 shows the comparison of graduates of
universities in Florida and graduates of universities outside of Florida for Samples A, B, and C. Graduates of universities in Florida accounted for 66.0 percent, 51.0 percent and 61.9 percent for Samples A, B, and C respectfully. The value of chi-square computed on these distributions was 4.69 which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted.
4Ibid., p. 145.
77
Table 29
Comparison of Age Distribution
for Samples A, B, and C
Age Sample A Sample B Sample C
Category Percent Percent Percent
25 and
under 59.0 69.0 38.1
26-30 15.0 17.0 23.8
31-35 10.0 5.0 9.5
36-40 4.0 7.0 9.5
40 and 12.0 2.0 19.0
Over
Total 100.0 100.0 100.0
Computed chi-square = 15.70 Critical value of chi-square at .01 level of significants = 20.09
78
Table 30
Comparison of Graduate of Universities in Florida
and Graduates of Universities Outside of Florida
for Samples A, B, and C
Sample A Sample B Sample C
Percent Percent Percent
Grad. of
univ. in 66.0 51.0 61.9
Florida
Grad. of
univ. out- 34.0 49.0 38.1
side of
Florida
Total 100.0 100.0 100.0
Computed chi-square = 4.69
Critical value of chi-square at .01 level of significance 9.21
79
Years of previous experience is compared in Table 31 for Samples A, B, and C. The percent of teachers with no previous experience was 62.0 percent for Sample A, 55.0 percent for Sample B and 47.6 percent for Sample C. Teachers with 2-5 years of experience were found to constitute 15.0 percent for Sample A and 21.0 percent and 33.3 percent for Samples B and C. The value of chi-square computed on these distributions was 7.22 which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted.
Table 33 shows marriage status for Samples A, B and C. Sixty percent of Sample A teachers were married while 57 percent of Sample B and 57 percent of Sample C fell in this category. The largest difference occurred in the category divorced or separated where Sample A showed 6.0 percent, Sample B-3.0 percent and Sample C-19.0 percent. The value of chi-square computed on these distributions was .19 which was not significant at beyond the .01 level of significance. Thus, at this significance level the null hypothesis is accepted.
The last personal history data variable, race, is
shown in Table 34. As can be seen from the table, Sample A had 22.0 percent black teachers, Sample B had 14.0 percent black teachers and Sample C had 57.1 percent black teachers. The value of chi-square computed on these distributions was 16.1 which was significant at the .01 level of significance. At this significance level the null hypothesis is rejected.
80
Table 31
Comparison of the Distribution of Years of Previous
Experience for Samples A, B, and C
Yrs. of
Yrs.iofs Sample A Sample B Sample C
pri Percent Percent Percent
experience
No previous 62.0 55.0 47.6
experience
1 year 10.0 16.0 14.3
2-5 years 15.0 21.0 33.3
6-10 years 10.0 5.0 4.8
Over 10 years 3.0 3.0 0.0
Total 100.0 100.0 100.0
To compute the chi-square value, cells in the over 10 years category were combined with cells in the 6-10 years category. Using these combined cells the computed value of chi-square was 7.22. The critical value of chi-square for six degrees of freedom at the .01 level was 16.81.
81
Table 32
Comparison of the Distribution of Grade Point
Average for Samples A, B, and C
Grade
point Sample A Sample B Sample C
point
average
2.1 and 2.5 5.1 9.1
below
2.2-2.5 22.2 18.2 72.7
2.6-3.0 45.7 43.4
3.1-3.5 28.4 28.3 18.2
3.6-4.0 1.2 5.1
Total 100.0 100.0 100.0
Missing data: Sample A 19
Sample B 1 Sample C 10
t2
Table 33
Comparison of Marriage Status
for Samples A, B, and C
Sample A Sample B Sample C
Percent Percent Percent
Married 60.0 57.0 57.1
Separated
or 6.0 3.0 19.0
Divorced
Single 34.0 40.0 23.8
Total 100.0 100.0 100.0
To compute the chi-square value, cells in the divorced or separated category were combined with the cells in the single category. Using these combined cells the computed value of chi-square was .19. The critical value of chi-square for two degrees of freedom at the .01 level was 9.21.
83
Table 34
Comparison of Race for Samples A, B, and C
Sample A Sample B Sample C
Percent Percent Percent
Black 22.0 14.0 57.1
White 78.0 86.0 42.9
Total 100.0 100.0 100.0
Computed chi-square = 16.04 Critical calue of chi-square at .01 level of significance = 9.21
84
The comparison of interview assessment is found in
Table 35. The lower the numerical score the more positive the interview rating. In the most positive category 6-15, Sample A had 9.5 percent, Sample B had 21.6 percent and Sample C had 11.8 percent. The greatest difference in interview assessment occurred in the least positive interview category 30-60. In this category Sample A had 10.5 percent, Sample B had 1.0 percent and Sample C had 29.4 percent. The value of chi-square computed on these distributions was 29.12 which was significant at the .01 level of significance. At this significance level the null hypothesis is rejected.
Table 36 contains the distribution of recommendations
assessment for Samples A, B, and C. The lower the numerical score the more positive the recommendations assessment. The total percent for the two most positive recommendations categories are as follows: Sample A 29.7 percent, Sample B 32.6 percent and Sample C 6.3 percent. In the least positive category 5 10, Sample A had 22.0 percent, Sample B had 27.2 percent and Sample C 68.8 percent. The value of chi-square computed on these distributions was 22.20 which was significant at the .01 level of significance. At this significance level the null hypothesis is rejected.
85
Table 35
Comparison of the Distribution of Interview
Assessment for Samples A, B, and C
Sample A Sample B Sample C
Percent Percent Percent
6-15 9.5 21.6 11.8
16-20 26.3 39.2 11.8
21-25 38.9 30.9 29.4
26-30 14.7 7.2 17.6
Above 30 10.5 1.0 29.4
Total 100.0 100.0 100.0
Missing Data: Sample A 5
Sample B 3 Sample C 4
To compute the chi-square value, cells in the above 30 category were combined with the cells in the 26-30 category. Using these combined cells the computed value of chi-square was 29.12. The critical value of chisquare for six degrees of freedom at the .01 level was 16.81.
86
Table 36
Comparison of the Distribution of Recommendations
Assessment for Samples A, B, and C
Sample A Sample B Sample C
Percent Percent Percent
1 8.8 6.5 0.0
2 20.9 26.1 6.3
3 34.1 26.1 6.3
4 14.3 14.1 18.8
5-10 22.0 27.2 68.8
Total 100.0 100.0 100.0
Missing data: Sample A 9
Sample B 8 Sample C 5
To compute the chi-square value, cells in the recommendations category (1) were combined with the cells in the category (2). Using these combined cells the computed value of chi-square was 22.20. The critical value of chi-square for six degrees of freedom at the .01 significance level was 16.81.
CHAPTER IV
DISCUSSION OF THE DATA
The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district. Data were gathered on three samples of teachers who performed in the urban school district. Sample A consisted of teachers who had performed satisfactorily enough to be offered a third contract. Sample B consisted of teachers in their first year of teaching and about whom no decision had been made concerning their continued employment. Sample C represented teachers who did not perform to a sufficiently satisfactory level to be retained under contract in the urban school district.
Regression equations were established for Sample A and Sample B to determine the relative predictive value of personal history data, interview assessment and recommendations assessment when principal's annual evaluation was used as the criterion measure. The regression equation for Sample A is shown in Table 27. Multiple R represents the intercorrelation of the independent variables as well as the correlation between the independent variable and the 87
88
dependent variable, evaluation assessment. The value of multiple R f-or Sample A was .25153. A correlation of this size was described by Guilford as a definite but small relationship.2 R-square represents Lhe proportion of the variance in the dependent vaiiable accounted for by the regression equation. For Sample A, R-square was .05834, meaning that the variables included in this equation accounted for less than 6 percent of the variance in the dependent variable, evaluation assessment.
The results for Sample B were only slightly better. Multiple R for Sample B, found in Table 28, was .31898. A multiple R of this size is also described by Guilford as a definite but small relationship.3 R-square, the proportion of the variance in the dependent variable accounted for by the regression equation, was slightly higher than in Sample A, but at .10175 it accounted for just over 10 percent of the variance.
Due to the manner in which the correlation coefficients between the independent variables and the dependent variable were presented in this study (multiple R's), the exact correlation coefficient for any one variable can not be observed. The multiple R listed for each independent variable
IJ. P. Guilford, Fundamental Statistics in Psychology and Education, (New York: McGraw-Hill Book Company, 1956), p. 145.
2Ibid., p. 145.
3Ibid., p. 145.
89
represents the relationship between not only the dependent variable but the interrelationships of the various independent variables to that point. The differences between the coefficients does give some idea of the relatively low coefficients obtained for each variable. Correlation coefficients below .40, as found in both Samples A and B, were cited in studies reported by Dunnette in 1963, as a basis for disenchantment with prediction models.4
A study by Smith, Albright, Glennen and Owens reported achieving correlation coefficients as high as .613 when comparing 484 personal history items to supervisory ratings.5 The findings in this study are in contradiction to the present study where no practical usefulness was found for personal history data. However, it is important to note that Smith, Albright, Glennen and Owens used 484 personal history items where as in the present study only six were included.6
Recommendations assessment added little to the intercorrelation of independent variables or the correlation between it and the dependent variable, evaluation assessment. This fact supports findings by Bolton that letters of recommendation have not produced usable results in the selection of recruits.' There are at least two possible
4M. D. Dunnette, "A Modified Model for Test Validation and Selection Research," Journal of Applied Psychology 47: 317-23 (October 1963).
5W. I. Smith, L. E. Albright, J. R. Glennen, and W. A. Owens, "The Prediction of Research Competence and Creativity from Personnel History," Journal of Applied Psychology 45:
59-62 (February 1961).
90
reasons that the recommendations produced such low correlations. First, the urban school district did not have a policy that governed who signed letters of recommendation and they accepted letters in all sorts of forms. On this point, the United States Civil Service Commission as the result of some of its work, recommended for improvement of the usefulness of letters of recommendation that it ought to be signed by a line executive in charge of the employee. Secondly, even though there was a districtdeveloped form available, its use was not required; as such, the district had not carefully defined discriminating terms to be commonly used. In this regard, as a result of a study by Peres and Garcia in which they synthesized adjectives used in recommendations, it was suggested that the identification of phrases would be helpful in establishing a more refined assessment procedure for handling recommendations.
Interview assessment produced correlation coefficients of disappointingly small magnitude just as recommendations
6Ibid., p. 61.
7
Dale L. Bolton, Selection and Evaluation of Teachers, (Berkeley: McCutchan Publishing Corporation, 1973), p. 65.
U. S. Civil Service Commission, A Guide for Executive
C.i.cction, Personnel Methods Series, No. 13, Washington, D. C.: Government Printing Office, 1961.
9Sherwood H. Peres and J. Robert Garcia, "Validity
and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnel Psychology 16: 279-86 (Autumn 1962).
91
assessment and personal history data. One problem with interviews, as was pointed out by Bolton, is that interviews can serve many purposes but a singular purpose must be 10
established for any one interview to be effective. No evidence was found on the urban school district interview assessment sheet that indicated that the interview was structured or unstructured. Studies reported by Crissy11 stressed that structured interviews were the only reliable form, while studies by Yongel2 supported structured or unstructured interviews as long as the purpose of the interview was clear. Bolton suggested that the validity of the interview depended upon the skill of the interviewer and that interviewers needed to be trained.13
The measure of success and thus the basis for determining if an independent variable was an effective predictor of success was based upon the principal's annual evaluation of the teacher. There was a considerable difference in the mean of the evaluation assessments for Sample A and Sample B due to the urban school district changing the evaluation form for the 1973-74 school year.
10Bolton, p. 67.
1W. J. E. Crissy, "The Employment Interview-Research Areas, Methods and Results," Personnel Psychology 5: 73-85 (Summer 1952).
K. A. Yonge, "The Value of the Interview: An Orientation and a Pilot Study," Journal of Applied Psychology 40: 25-31 (February 1956).
13Bolton, p. 74.
|
Full Text |
PAGE 1
THE RELATIVE PREDICTIVE VALUE OF SELECTED VARIABLES IN MEASURING TEACHER SUCCESS IN AN URBAN SCHOOL DISTRICT By JON D. 'l'HOMPSON A DISSERTA'I'ION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF EDUCATION UNIVERSITY OF FLORIDA 1974
PAGE 2
ACKNOWLEDGMENTS To my doctoral committee, Dr. Michael Nunnery, Dr. William Alexander, Dr. Ruthellen Crews and Dr. Forbis Jordan my sincere gratitude is extended for their personal efforts and their influence on the final product. Special appreciation is extended to Dr. Nunnery for his patience and helpful suggestions. To my wif~ Connie, and my children Bryan and Julie, we share in the joy of the conclusion of the study. ii
PAGE 3
ACKNOWLEDGMENTS ABSTRAC'I' CHAPTE R TABLE OF CON T ENTS I. INTRODUCTION The Proble m PAGE ii V Statement of the Problem ..... 1 2 2 3 4 5 II. Delimitations . Limitations ...... Justification .... Assumptions .... Definition of T e r m s Dependent Variable. .... Independent Variabl e .. Interview Assessment Interview Assessment Shee t ..... Personal History Data Recomm endations Asses s m e n t Sample A ... Sample B ..... 6 7 7 7 7 7 8 8 8 8 9 Sample C . Procedures .. The Sample s ... Sources and Collection of Data. Treatment of Data ..... Organization of the R emainder of the Study ... REVIEW OF LITERATURE Criteria Determination Recruitment Application or Personal History Data Reconunendations .. Interviews Synopsis of t h e Review of Literature iii 9 9 10 13 14 15 16 19 20 22 24 32
PAGE 4
CHAPTER III. PRESENTATION OF DATA. Basic Descriptive Data Relative to the Independent and Dependent V ariables PAGE .33 for Samples A and B . .34 Personal History Data for S ~mples A and B . .. 34 Interview Assessment Data for Sample s A and B . . .40 Recommendations Assessment for Samples A and B .44 Evaluation Assessment for Samples A and B . .44 Independent Variable Data in Relationship to the Dependent Variable for Sample s A and B 4 7 Personal History Data in Relationship to Evaluation Assessment .. 48 Interview Assessment Relatej to Evaluation Assessment Recommendations Assessment in Relationship to Evaluation Assessment Results of Regression Analysis for Samples A and B. .. A Comparison of the Distribution bf the Independent Variables for Samples A, .61 6 4 .69 B, and C . 72 IV. DISCUSSION OF THE DATA. 87 V. SUMMARY, CONCLUSIONS AND IMPLICATIONS .97 Summary Conclusions and Implications SELECTED BIBLIOGRAPHY. BIOGRAPHICAL SKETCH. iv 97 103 106 109
PAGE 5
Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Education THE RELATIVE PREDICTIVE VALUE OF SELECTED VARIABLES IN MEASURING TEACHER SUCCESS IN AN URBAN SCHOOL DISTRICT By Jon D. Thompson August, 1974 Chairman: Michael Y. Nunnery Major Department: Educational Administration The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district. Specifically, the focus was on the following: 1. For a sample of teachers who had performed satisfactorily enough to be offered a third contract by the urban school district, to determine the relative predictive value of personal history data, interview assessment and recommendations assessment and any combination thereof, when the average of three principals' annua l evaluations was used as the criterion measure. V
PAGE 6
2. For a sample of teachers in their first year of teaching within the urban school district and about whom no decision had been made relative to their continued status, to determine the relative predictive value of personal history data, interview assessment and recommendations assessment and any combination thereof, when the principal's annual evaluation was used as the criterion measure. 3. To compare the two aforementioned samples of teachers in regard to their basic characteristics with a third sample of teachers con~isting of those teachers who did not perform to a sufficiently satisfactory level to be retained under contract in the urban school district. Data for each independent and dependent variable for each member of the three teacher samples involved were collected from the personnel records of the single urban school district. Frequency distributions, joint frequency distributions and multiple regression equations were developed for the two samples of 100 teachers each. The characteristics of the two samples of teachers then presently employed (1974) were compared to the third sample of 21 teachers who were not reappointed by the urban school district. Based upon the preceding, analysis, the following major findings emerged: 1. The values of multiple R, the intercorrelation of the independent variables as well as the dependent vi
PAGE 7
variable for the two samples of 100 teachers, ranged from .25153 to .31898. 2. R-square, the prop o r tion of the variance in the dependent variable accounted for by the re gression equation was .05834 for the s ample of teachers in their third year of employment and .10175 for the s a m ple of teachers in their first year of employment. 3. When the three samples were com pared, significant differences were found on the variables of race, interview assessment and recommendations a~sessment. It was concluded that: 1. The relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers for an urban school district was limited. 2. The findings of the present investigation were consistent wit h findings of other studies under similar circumstances. 3. Even though peripheral to the major intent of the investigation, based on the data obtained when Samples A, B, and C, were compared, it seems justifiable to conclude that if persons are selected for teaching positions when these individuals have poor interview and recommendations ratings, then the likelihood of their success is quite limited. vii
PAGE 8
CHAPTER I INTRODUCTION P ersonnel administrator s in local school systems have continually faced the problem of selecting the best qualified applicants to fill teaching positions.1 Two events have occurred within the last few years that have resulted in greate r emphasis being placed on insuring that the teacher selection process w a s effective. First,.the numb e r of applicants per vacancy has increased rapidly. This fact makes it indefensible for school systems to choose teachers at random. Secondly, society has placed greater pressure on schools to achieve their goals. Despite the fact that the goals have often been ill-defined, the task of achieving any educationa l goal is more readily accomplished with the best qualified teacher applicant being selected for employ-2 ment. Numerous methods 11ave been used in attempting to :_;elect the best qualified applicant. They indlude: college grades, previous experience, letters of recommendation, personal interviews, background data (e.g., age, sex, socioeconomic status) and statements of an applicant's educational philosophy. Research conducted on the effectiveness of these methods 1 Dale L. Bolton, Selection and Evaluation of Teachers. (Berkeley: Mccutchan Publishing Corporation, 1973), p. 1. 2Ibid., p. 2. 1
PAGE 9
2 has not been conclusive. It would appear that there have been no uriiversal guidelines developed that apply to a cross section of school districts. Bolton suggested that it was necessary for each school district to ~alidate its own selection method.3 In view of the wide range of t echnique' s used and the obvious need to improve the selection process, the present study was undertaken. Utilizing data from a single urban school district the intent was to upgrade the selection process in the urban school district and to add to the present body of knowledge. The Problem Statement of the Problem The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessme n t when utilized in the selection of teachers in a single urban school district. Specifically, the focus was on the following: 1. For a sample of teachers who had performed satisfactorily enough to be offered a third contract by the urban school district, to determine the relative predictive value of personal history data, interview asses!=,ment and recommendations assessment and any corn~ination thereof, when the average of three principals' annual evaluations was 3Ibid., p. 2
PAGE 10
used as the criterion measure. 2. For a sample of teachers in their first year of teaching within the urban school district and about whom no decision had been made relative to their continued status, to determine the relative predictive value of personal history data, interview assessment and recornmendati,ons ass,ess ment and any combination thereof, when the principal's annual evaluation was used as the criterion measure. 3. To compare the two aforementioned samples of teachers in regard to their basic characteristics with a third sample of teachers consisting of those teachers who did not perform to a sufficiently satisfactory level to be retained under contract in the urban school district. Delimitations 3 The study was confined to three samples of teachers in a single urban school district. Two samples, of 100 teachers each, were randomly chosen. Sample A consisted of teachers hired in 1971-72 and who were in continuous employment in the urban school district to January 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the district. Sample B included those teachers initially employed in 1973-74 by the urban school district and who were in continuous employment
PAGE 11
4 to January 1974. This sample constituted the sample about whom no decision had been made regarding their continued employment in the urban school district. Sample C, con-sis ting of 21 arbitrarily chosen teachers, composed the group of teachers who had not performe d at a sufficiently satisfactory level to be retained by the urban school district and were no longer employed at the time of the investigation. Data for each individual teacher in each of the three samples relative to the independent variables (personal history data, interview assessment and recommendations assessment) were collected exclusively from the personnel records of the urban school district. Data regarding the dependent variable, principal's annual evaluations, were also collected from the personnel records of the urban school district. Limitations Four major limitations were recognized in conducting the present study. First, the data used in making judgments about the relative usefulness of the various independent variables did not include samples from the full array of applicants who were available to the urban school district. That i'.s, there was no attempt to select samples from among applicants who were not offered employment in the urban school district. Second, from the literature it was found that for any evaluation instrument to be effective, a position
PAGE 12
analysis should be developed for each job. The evaluation instrument used by the urban school district was of 5 a general type and was used for all instructional personnel and had no relatioriship to a specific job. Third, if an individual principal's annual evaluation is viewed as being an insufficient assessment of one's performance, then the study suffered from a weakness in the strength of the dependent variable. Fourth, since the study was confined to a single urban school district, its external validity beyond the district was open to question. Justification The development and refinement of methods to accurately predict which applicants will be successful on the job and therefore should be employed have been studied extensively. The major portion of these studies however, has focused on the selection procedures for industry and government. Bolton, in Selection and Evaluation of Teachers,4 indicated that few systems utilize validated selection techniques for teacher employment. Due to teacher shortages in the past, this situation has been tolerated by educators and the public. The teacher supply in the early 1970's, and the increased concern over accountability in education should cause school personnel officials to develop more sophisticated selection methods. 4rbid., p. 49.
PAGE 13
Based upon personal experience in personnel work and a review of the literature on the selection process, it 6 was the opinion of the writer that data collected in regard to certain variables are superior to data collected in regard to other v ariables in predicting t eacher success on the job and that a refined s election procedure can be designed for school systems to improve the decision-makin g process. The 'Conclusions of this study should provide a basis for designing a selection procedure. that would assist the urban school district in making more effective teacher selections. In addition to being of value to the urban school district, the proposed study was perceived as important in that it adds to what Bolton has described as the meager body of knowledge available on teacher selection.5 Assumptions The basic assumption for this study was that the teachers used as the sample would be similar enough to future teacher applic&nts in the urban school district so that any conclusions reached would be reasonably valid for teacher selection decisions within the single urban school district. 5Ibid., p. 51.
PAGE 14
7 Definition of Terms Dependent Variable For Sample A the dependent variable consisted of the mean of three annual evaluations, each completed by the principal of the school the teacher was assigned to for a given school year. The dependent variable for Sample B consisted of the mean of the annual evaluation for the 1973-74 school year completed by the p rincipal of the school to which the teacher was assigned. Independent Variable The values determined for personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status and race), interview assessment data and recomme.ndations assessment data. Interview Assessment The rating assigned each applicant by the interviewer of the urban school district, based upon his evaluation of the applicant at the time of the interview. (The method for quantification of the rating is explained in the procedures section of this report.) Interview Assessment Sheet The sheet used by the interviewer of the urban school district when recording his assessment of the applicant interviewed. The applicant was rated on a continuum in
PAGE 15
six categories: appearance, voice, poise, manner, enthusiasm and interest. 8 Personal History Data The term included: age college location, previous experience, transcript assessment, marriage status and race. The items of information were supplied to the urban school district by the ap~licant on the application blank or college transcript, prior to employment. Recommendations Assessment The numerical value assigned by the.recommending party converted to a 1-10 scale or, in the case where only narrative data were provided by the reconunending party, the writer determined the 1. numerical value utilizing the procedures detailed in the procedures section of this report. Sample A This sample consisted of 100 teachers, randomly chosen, who were hired in 1971-72 and who were in continuous employment in the urban school district to January 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the urban school district. Sample B This sample consisted of 100 teachers, randomly chosen, who were initially employed in 1973-74 by the urban school district and who were in continuous employment to January 1974. This sample constituted the sample about whom no
PAGE 16
decision had been made r e garding continued employment in the urban school district. Sample C This sample consiste d of 21 arbitrarily chosen 9 teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district and were no longer employed at the time of the investigation. The teachers in this sample were employed in the urban school district for varying lengths of time during the period 1971-73. Procedures In order to determine the relative usefulness of personal history data, interview assessment and recommend ations assessment in the selection of teachers in a single urban school district, the following procedure s were utilized. The Samples Three samples, which included elementary and secondary teachers, were utilized in this study. Sample A consisted of 100 teachers, randomly chosen,who were hired in 1971-72 and who were in continuous employment in the urban school district to J anuary 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the district. Sample B consisted of 100 teachers, randomly chosen, who were initially employed in
PAGE 17
10 1973-74 by the urban school district and who were in continuous employment to January 1974. This sample constituted the sample about whom no decision had been made regarding continued employment in the urban school district. Sample C, consisting of 21 arbitrarily chosen teachers, composed the group of teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district and were no longer employed at the time of the investigation. The teachers in this sample were employed by the district for varying lengths of time during the period from 1971-73. Sources and Collection of Data The files of the urban school district were utilized for securing the data for each of the independent variables: personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status, and race), interview assessment and recommendations assessment for each teacher in each of the three samples. Data relative to the dependent variable, evaluations assessment, were also secured in the files of the urban school district for each teacher in Sample A and Sample B. 1. Personal history data. With the exception of grade point average, the data included in the category were contained on the application form that the urban school district required each applicant to complete. The application form became part of the employee's permanent file along with the transcript, after the applicant was employed. (The permanent files of all personnel were made available to the writer.)
PAGE 18
11 2. Interview assessment. The interviewer for the urban school district evaluated the applicant on each of six personal characteristics at the conclusion of the inter-view. The results of all personal interviews were recorded on the interview evaluation sheet, designed for that purpose. The six personal characteristics were: appearance, voice, poise, manner, enthusiasm and interest. The interviewer indicated his assessment of the applicant by placing a check on a continuum for each of the characteristics. The writer divided each continuum into ten equal parts for purposes of quantification. Thus each applicant attained an interview assessment score from 6-60. (Six represer.ted a favorable rating, 60 an unfavorable rating.) 3. Recommendations assessment. Recommendations sub-mitted about applicants were filed with the Interview Evaluation Sheet, separate from the permanent file. Most recommendations were submitted on a form with a scale to which a numerical value could be assigned. For those recommendations that were submitted in a narrative form, the writer determined the rating of the applicant based upon his experience in assessing recommendat.ions for the single urban school district and a study by Peres and Garcia, reported in the review of literature.6 Specifically, the writer 6sherwood H. Peres and J. Robert Garcia, "Validity and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnei Psychology 16: 279-86 (Autumn 1962).
PAGE 19
12 examined the letters of recommendation to locate the most discriminating categories of adjectives a~ determined by Peres and Garcia. Peres and Garcia found that adjectives dealing with mental ability were the most discriminating. Adjectives dealing with vigor, dependability-reliability, urbanity and cooperation-consideration were found to be correspondingly less discriminating.7 The writer determined whether the adjectives in the letters of recommendation were favorable or unfavorable to the applicant. A numerical rating from one to ten was then assigned by the writer to represent the recommendations assessment for a particular teacher. 4. Evaluation assessment. Data in regard to the dependent variable, the teacher's annual evaluation as completed by the principal to whom the teacher was assigned for the given school year, were secured in the personnel files of the urban school district. The evaluation for the school year 1971-72 contained 31 items that were used to rate the teachers. The evaluation for the 1972-73 and 1973-74 school years contained 21 items that were used by the principal to rate the teachers. The mean was calculated for each teacher'3 evaluation for each year. For Sample A, the evaluation assessment represented the mean of the teacher's three annual evaluations. For Sample B, the evaluation assessment represented the mean of the teacher's 1973-74 annual evaluation. 7Ibid., p. 285.
PAGE 20
13 Treatment of Data To determine the relative usefulnes s of personal history data, intervie w a .ssessment and recommendations assessment when utilized in the selection of teacher s for an urban school district, the following operations were performed. 1. Frequency distributions were developed for Samples A and Bin r egard to each of the nine variables to provide some idea of the basic characteristics o f the data in terms of distribution and variability. Specifically, frequency distributions were developed that showe~: absolute frequency, percent and cumulative percent. 2. Joint frequency distributions were developed to show. how the dependent variable related to each of the independent variables. Separate joint frequency distributions were developed for Sample A and Sample B. The format consiste d of: a. Listing the numerical total of participants for each cell of the joint frequency distribution. b. Indicating row percent, the percent of each dependent variable category in a particular category of the independent variable. c. Indicating the column percent, the portion of each inde p e n dent variable category in a particular category of the dependent variable d. Indicating total percent, the portion of the total (100 percent) each cell represented. 3. Two regression equations were computed using: age, graduate of a university in Florida years of previous experience, grade point average, race, interview assessment
PAGE 21
14 and recommendations assessment as independent variables to predict the dependent variable (evaluations assessment). The multiple regression analysis produced the highest possible correlation between the independent variables and the dependent variable. The summary tables for the regression equations present the following information: multiple r, r-square, r-square change, Band Beta. Multiple r indicates the linear relationship between the independent variables and the dependent variable under investigation. R-square change is the change in r-square from the value of the previous step. The Band Beta values are the regular and normalized regression coefficients, respectfully. 4. Frequency distributions were developed to compare Samples A, B, and C in regard to each of the eight independent variables. Specifically, frequency distributions were developed that showed the percent of Samples A, B, and C in each category of the eight independent variables. Organization of the Remainder of the Study Chapter II contains a review of literature which focuses primarily on previous studies in the selection of personnel. Chapter III is a preseutation of the data. Chapter IV contains a discussion of the data and Chapter V contains the summary, conclusions and implications.
PAGE 22
CHAPTER II REVIEW OF LITERATURE Approximately 240,000 teachers were selected for employment in the United States in 1970-71.1 For each teacher selected, other applicants were considered and rejected. Yet, the decisions for choosing these teachers and rejecting the others were based upon data that generally had not been validated and ignored research relating to the 2 selection of employees. Apparently the limited research dealing directly with the selection of teachers or at least the teacher shortage did not provide an impetus for school systems to develop more sophisticated methods of teacher selection. Considerable research on the selection process has been performed in industry and government and has been gradually applied to the selection process for teachers. The present review is arranged under subject headings based upon a review of the literature by Bolton in Selection and Evaluation 3 of Teachers. These are: criteria determination, recruit-ment, application or personal history data, recorrmtendations and interviews. 1 Dale L. Bolton, Selection and Evaluation of Teachers, (Berkeley: Mccutchan Publishing Corporation, 1973), p. 1. 2 Ibid., pp. 1-4. 3Ibid., pp. 50-92. 15
PAGE 23
16 Criteria Determination To determine if selection methods are successful, a criterion for success on the job must be established. Traditionally in education this has taken the form of deriving a single standard of a "good" teacher. Robert Gui6n disputed this practice, The fallacy of the single criterion lies in its assumption that everything that is to be predicte d is rela t e d to everything else to be predicted--that there is a general factor in all criteria accounting for virtually all of the important variance in behavior at work and its various consequences of value.4 Dunnette stated that personnel involved in the selection process must be willing to back off from the all encompassing criteria of job effectiveness. Dunnette advocated a more careful study of job behavior with emphasis upon differing styles within the same job. The selection process should avoid assuming that there is a single all-encompassing mea-sure of occupational success.5 Dunnette pointed out that data processing equipment has given employers greater capabilities to include more selection data and evaluation criteria to make selection decisions, than was once felt possible.6 4Robert M. Guion, and Richard Battier, "Validity of Personality Measures in Personnel Selection," Personnel Psychology 18: 135-64 (Sununer 1965). 5M. D. Dunnette, "A and Selection Research," 317-23 (October 1963). 6 Ibid. p. 3 2 0. Modified Model for Test Validation Journal of Applied Psychology 47:
PAGE 24
17 Supporting Dunnette's view that a single criterion measure is a weaknes s of many r:;election methods is a modification of typical prediction models offered by Guetzkow and Forehand. Their model included the complex interactions which m a y occur between predictors and various predictor combinations, different groups of individuals, different behaviors of the job and the consequences of these behaviors relative to the goals of the organization.7 Similar job behaviors m a y be predicted by quite different patterns of interaction between various predictors and individuals, according to Gu etzkow and Forehand. The modified model for test validation and selection research proposed by Guetzkow and Forehand did not rely upon the correlation coefficient as the sole criterion for valida-tion of selection techr.iques. Their model provided a "richer schematization'' for prediction research and offered important implications for the directions of future research.8 The fact that criteria change from time to time and from one situation to another has been demonstrated. Bernard Bass in a study of food product salesmen measured evaluations over a period of 48 months. The correlation coefficient of two separate merit ratings by essentually the same supervisors fell from .62 to .29 over a 42 month period, indicating the criterion used by the supervisors varied from the 7Harold Guetzkow and Garlie A. Forehand, "A Research Strategy for Partial Knowledge Useful in the Selection of Executives," In Research Needs i o Exec~tive Selection, edited by Ranato Tagiuri, Boston: H arvard University, Grad u ate School of Business Administration, Division of Research, 1961. 8rbid., p. 48.
PAGE 25
18 beginning of the perioc1 to the end. In the same study Bass found a higher correlation coefficient between esteem and merit (.4;) than merit and ability (.18).9 To avoid the situation where evaluators vary on the factors used to rate employees, researchers have recommende d developing a position analysis. In making this determination Bolton suggested that the employer ask, "What must the individual do and what characteristics must he possess to be a success."10 Bolton listed the following guides. 1. The number of behaviors listed .should be l imited. 2. Behavior should relate to the purpose of the situation in which they are to be exhibited. 3. Behavior should be measurable. 4. Behavior should include verbal interactions as well as non-verbal communications, and in-class as well as out-of-class action. 11 Garland DeNelsky and Michael G. McKee conducted a study for the Central Intelligence Agency in which a modi- fied Q-Sort was used that reliably expanded the variances of predictor and criterion variables.12 Seven staff psychologists evaluated 32 male employees. When the researchers 9Bernard M. Bass, "The Leaderless Group Discussion," Psychological Bulletin 51: 465-92 (September 1954). 10 Bolton, p. 58. 11rbid., pp. 59-60 12Garland Y. DeNelsky and Michael B. McKee, "Prediction of Job Performance from Assessment Reports: Use of a Modified Q-Sort Technique to Expand Predictor and Criterion Variants," Journal of Applied Psychology 53: 439-45 (December 1969).
PAGE 26
19 correlated the composite assessment-report predictions of the seven judges with the individual predictions of each of the seven judges, the results produced correlation coefficients of .63 for Group One and .66 for Group Two for individual prediction, and .92 and .93 for the respective groups with composite predictions. Although the psychologists were more accurate in predicting weaknesses than strengths, it did suggest that pooled ratings for either selection or evaluation produced more accurate d 13 ec1s1.ons. Recruitment Fundamental to the effective selection of teachers is the existence of a sufficient number of qualified candidates from which to make selections.14 Since selection implies choice, recruitment is necessary to attract high quality applicants in a sufficient number so that validated selection techniques can be utilized. The intensity of a recruitment program is related to the availability of teachers in a given area. Some school districts are forced to a more aggressive recruitment program due to their remoteness to an adequate supply. The basis for any selection program is to attain quality personnel who will remain for a period of time with their 13rbid., p. 441. 14 Bolton, p. 60.
PAGE 27
20 employer. Recruitment of personnel who fit this requiremen~ has been shown to justify the additional cost of the recruit-ment. A study by Fitzgerald found the average cost per teacher hired in a midwestern state to be $146. Business and industry were found to pay up to $1,822 for recruitment 15 of each professional employed. If the good teacher has a more positive effect on students than an average or poor teacher then it is difficult to rationalize the variance in expenditure. Application or Personal History Data Information on the application form,usually demographic data, has been widely used to predict employee success as measured by supervisory ratings. Dunnette commenting in 1963 about such data stated: We cannot and should not try to avoid the fact that statistics of selection (i.e. validity coefficients) are far from gratifying and offer little support to anyone claiming to do much better than chan9e in the selection process.16 17 In contrast to studies compiled by Dunnette that reported relatively low correlation coefficients for validating biographical data, was a report entitled "The Prediction of Research Competence and Creativity," from Personal History, 15Paul Fitzgerald, "Recruitment of Teachers--A Need for Reevaluation," Personnel Journal 49: 312-14 (April 1970). 16 Dunnette, p. 320. 17 b'd 318 I 1 ., p.
PAGE 28
21 18 by Smith, Albright, Glennen and Owens. The personal history questionnaire contained 484 multiple-choice items that were designed for the study. The results of the study produced a correlation coefficient of .613 when comparing the 484 items to supervisory ratings. This was considerably h l f d t d. 19 h1g er t1an oun 1n previous s u 1es. It was acknowledged by the authors that the personal history technique can be more effective with highly skilled employees than might be found in studies involving lowe r ability groups. The authors also admitted that some of the personal history items were similar in n ature to tests of personality, interests and values and as such were not purely demographic in nature.20 Charles Adair conducted a study 1n 1973 to determine the correlation between the undergraduate grade point average of education majors and their success on the job as rated by their supervisors in their first year of teaching. Adair found the only significant grade point category was 3.50 -4.00 on a four-point scale. The 3.00 -3.49 category had the next highest correlation coefficient, while the 2.50 -2.99 category had the lowest correlation coefficient 21 with success. 18w. I. Smith, L. E. Albright, J. R. Glennen and w. A. Owens, "The Prediction of Research Competence and Creativity from Personal History," Journal of Applied Psychology 45: 59-62 (February 1961). 19 Ibid. p. 61 20ibid., p. 62.
PAGE 29
22 Adair concluded that there were few significant correlation coefficients between undergraduate grade point average and success on the job as measured by the supervisor for first year teachers. There was some evidence to suggest that the most successful teachers had the highest d 22 undergraduate gra e point average In a study by Langdale, two groups of personnel interviewers were given completed application blanks to judge. One group was given a general job description and the other group a complete description of the job to be filled. The group that had the complete job description achieved a higher degree of interrater reliability (.87) than the group using the nonspecific description (.35).23 Recommendations Most studies evaluating letters of recommendation indicate that this method, used independent of other selection methods, 21charles D. Adair, "Relationship Between Undergraduate Grades and First Year Teaching Success," School and Community 60: 22 (January 1974). 22 b'd I 1 ., p. 22. 23John A. Langdale and Joseph Weitz, "Estimating the Influence of Job Information on Interviewer Agreement," Journal of Applied Psychology 57: 23-27 (February 1973).
PAGE 30
23 24 has not produced useful results. However, certain kinds of information from previous employers may be valuable. Stone and Kendall suggeste d that dates of employment, salary, whether the employer would seek an appraisal and whether the former employer would rehire the applicant should be useful in evaluating the applicant.25 The researchers found that letters of recommendation carried by the applicant are 26 generally worthless. The United States Civil Service Commission recommended that the information an employer seeks from previous employers correspond to the unique aspects of the job to be filled. They also emphasize that a letter signed by an executive 27 can be helpful. An extensive study was made by Peres and Garcia in which they analysized 625 reference letters. They synthesized 170 adjectives from the letters and sorted them into six categorie s : 1. Extraversion 2. Interpersonal Relations 3. Dependability-Reliability 24 Bolton, p. 65. 25 Harold C. Stone and William E. Kendall, Effective Personnel Selection Procedures, (Englewood Cliffs: PrenticeHal 1, 19 5 6) p. 4 2. 26rbid. p. 55. 27 . 'd f U.S. Civi Service Commission, A Gui e or Executive Selection, Personnel Methods Series, No. 13, Washington, D. C.: Government Printing Office, 1961.
PAGE 31
24 4. Forcefulness 5. Polish 6. Physical Vim Two hundred supervisors were then asked to describe their "best" and "poorest" employee considering all 170 adjectives. The authors found that adjectives dealing with mental agility were most discriminating, and cooperationconsideration least discriminating. In order of discrimination: 1. Mental Agility 2. Vigor 3. Dependability-Reliability 4. Urbanity 28 5. Cooperation-Consideration Peres and Garcia concluded it can be hypothesized that when a person is asked to submit a letter for an applicant he feels is not tryly qualified, the person is likely to say, "Mike is a pretty nice guy." This is a. practice, it might be said, whereby former employers could "damn with faint praise." The authors suggested however, that the best alternative was a forced choice scale of adjectives that would be completed 29 by the former employer. Interviews No part of the selection process is more widely used than the personal interview.30 28sherwood H. Peres and J. Robert Garcia, "Validity and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnel Psychology 16: 279-86 (Autumn 1962). 29 b"d 285 I 1 ., p. 30 Bolton, p. 67.
PAGE 32
Most people feel that the interview contributes something to the process that c annot be gaine d in any other way. 31 The above statements appeared to represent what has actually taken place in the selection process in most 25 school systems. Interviews actually serve many purposes. The interviewer can insure that the applicant understands the job, and thereby he can reduce turnovers resulting from persons accepting jobs they might have rejected if they had fully understood the requirements. Interviews are also designed to gather information on the applicant th t ld b h 3 2 a cou not e ot erwise acquireu. One study, conducted by Chruden and ~herma~ supported the use of nondirective or unstructured interviews in which the applicant was allowed to talk about whatever seem im-33 portant. In another study, Crissy stressed that structure was the only reliable form.34 He stated that the interview continued to be the most widely used personal selection method because: 31 l d h 1 R1c1ar A. Fear, Te Eva uation Interview, (New York: McGraw-Hill Book Company, 1958), p. 102. 32 Bolton, p. 67. 33Herbert J. Chruden and Arthur w. Sherman, Junior, Readings in Personnel Management, (Cincinnati: Southwestern Publishing Company, 1961), p. 72. 34w. J. E. Crissy, "The Employment Interview-Research Areas, Methods and Results," Personnel Psychology 5: 73-85 (Summer 1952).
PAGE 33
1. The supplementary, non-evaluative functions served by the interview, 2. An almost universal conviction on the part of supervisory and managerial personnel that they are good "pickers of men," 26 3. The expectation on the part of job applicants of person~! treatment such as is accordedih the interview ) In referring to three studies conducted by students under his direction, Crissy identified three fundamental aspects of interview reliability: 1. Intra-rater consistency, i.e., the agreement of the interviewer with himself insofar as his appraisals of interviewees are concerned; 2. Inter-rater consistency, i.e., the agreement among interviewers insofar as their respective judgements of interviewees are concerned; 3. The consistency of behavior elicited in the interview, i.e., .the extent to which the interviewee behaves and responds in the same 3 ~ way to similar stimuli in successive interviews. 0 One generalization stands out. The more structured the interview, the more reliable it becomes in relation to the above three points. The more structured it becomes, however, the more it loses the supplementary, non-evaluative 37 nature that some employers prefer. 35Ibid., p. 73. 36rbid., p. 74. 3 7 Ibid. p. 74.
PAGE 34
Crissy suggested methods of achieving consistency in each type of interview, but in addressing himself to the validity of selecting a person that will be successful on the job, he concurred with other writers by recommending that specific traits of the job to be filled be identified and that the interviewer perform a structured interview assessing those traits. The employer must however, determine if the selection method is effective. To accomplish this task Crissy suggested: 1. The person who rated the applicant should not rate the same person as an employee. 2. Employees selected for employment represent the upper portion of applicants and as such are more homogeneous by traits. This fact limits the distribution of any selection method. 3. The experimenter should be aware of the inherent relationship between selection ratios and validity coefficients. 4. The employer should evaluate the interviewers and determine what percent of employees ~ired by each interviewer ultimately become successful employees.38 27 K. A. Yonge, in assessing the success of the interview, felt that the only just criticism of the interview method was based upon the practice of conducting interviews with no definite purpose or aim. Yonge does not agree that because interviews become structured that they will necessarily be valid. Some interviews by purpose depend upon the dynamic 38rbid., p. 79.
PAGE 35
28 situation, or interaction that occurs in an unstructured environment. Such items as poise, motivation, sincerity or purpose, emotional balance and responsiveness can best b d d d d 39 e JU ge un er unstructure settings. Yonge cited studies by Moriwaki40 and Snedden41 in which interview assessment was compared to success on the job and correlation coefficients of .63 and .82 were achieved when rigid specifications for the interview were used, as evidence that higher validity can be achieved by structured interviews. Yonge conducted a pilot study himself which included an outline for each interviewer. He obtained correlation coefficients ranging from .48 to .99 when comparing interview assessment to job success, with relatively few subjects accounting for most of the variance. He concluded that, properly used, the interview structured or unstructured, can play a reliable part in the overall assessment of an individual's qualities for 42 employment. Several studies have been conducteo to determine how interviewers react to favorable and unfavorable inform, .tion 39K. A. Yonge, "The Value of ~he Interview: An Orientation and a Pilot Study," Journal of Applied Psychology 40: 25-31 (February 1956). 40E. Moriwaki, "Note on the Comparative Validities of Judgements of Intelligence Based onPhotographs and on Interviews," Journal of Applied Psychology 13: 630-631 (November 1947). 410. Snedden, "Measuring General Intelligence by Interview," Psychological Clinic 19: 131-134 (July 1930). 42 Yonge, p. 30.
PAGE 36
29 received in an interview. Bolster and Springbett reported in 1961 that interviewers reacted more strongly to unfavorable information about the applicant tha n they did to favorable information.43 The study involved over 200 officer cadet selection reports arid 16 personne l officers. The design involved giving the personnel officers a hypothetical rating on a man and introducing contrary information to determine how it affected his opinion. The shifts in opinion tended to follow the patterns listed below. 1. The amount of interview com..mitm ent at the time a decision shifts and the w eight of the challenging information both affect the amount of change made in decisions. 2. It is easier to shift a rating in the direction of rejection than in the direction of acceptance. 3. People have different amounts of sensitivity to negative evidence. 4. An item of information received toward the end of the interview carries more weight than it would if received earlier in the interview. Thus, when information is received affects how it is perceived. 4 4 Hollmann conducted a study in 1972 designed to refute the findings of Bolster and Springbett. In this study, experienced interviewers were given hypothetical applicants to rate. The information was arranged so that negative information preceeded positive information on some applicants 43B.I.Bolster and B.M.Springbett, "The Reaction of Interviewers to Favorable and Unfavorable Information," Journal of Applied Psychology 45: 98-103 (February 1961). 44Ibid., p. 103.
PAGE 37
30 and followed positive information on other applicants. Hollmann found that per unit of importance, the interviewers gave less weight to positive information rather than more weight to negative information as Bolster and Springbett had found.45 Hollmann suggested that the reason for the apparent emphasis on negative information was due to greater feedback from supervisors on negative information than on positive information.461 In addition to the impact of the timing of information received du~ing an interview, interviewers should be aware of the influence of several items. Mandell reported that such items as the applicant's resemblance to someone the interviewer knows and nervousness on the part of the interviewee are sources of e~ror related to the interviewer's attitude.47 Springbett, in a separate study, concluded that no generalizations about the value and precision of the interview method can be made, because the interviewer is part of the process. Springbett further stated that three separate operations are being performed by the interviewer: gathering, processing and evaluating. These operations are to some 45 Thomas D. Hollmann, in Processing Positive and of Applied Psychology 56: 46Ib'" 134 l.Q. p. "Employment Interviewers' Error Negative Information," Journal 130-134 (April 1972). 47Milton M. Mandell, The Employment Interview, Research Study number 47, (New York: American Management Association, 1961), p. 47.
PAGE 38
degre~ independent. An interviewer can be adept at one operation and ineffective in the others.48 Bolton found that the interviewer feels more confident about his decision in the selection process when he has had the opportunity to both see and hear the applicant. Bolton offered the following summary of the value of th~ interview in the selection process. 1. The validity of the interview depends upon the skill of the interviewer, the situation and type of applicant. The assumption that the mere process of gathering, weighing and evaluating information attained in an interview will:, be an accurate one has not been proven. 2. Interviewers need to be trained and become familiar with the common errors that research has identified. 3. More research is needed. The part the interview plays, the type training the interviewer should receive, the number of interviews necessary and who makes a good interviewer need to be decided. 4. A better evaluation of the interview as a selection instrument needs to be made. In addition to the evaluation of the applicants selected, a follow up on the applicants rejected must be made to fully evaluate the interview method. Correspondingly, the evaluation of the interviewer and the recognition that all people don't make good interviewers must be accomplished.49 48B. M. Springbett, "Factors Affecting the Final Decision in the Employment Interview," Canadian Journal of Psychology 12: 13-22 (March 1958). 49 Bolton, p. 7 4 31
PAGE 39
32 Synopsis of the Review of Literature To determine whether selection methods are successful, a criterion for success on the job must be established. There has been disenchantment with the single criterion measure of success and several alternatives have been presented. The selection methods most widely used and studied include: personal history data, interviews and recommendations. The validity of these various methods has ranged from unacceptable to useful levels. Examples of successful application of the methods seemed to occur when there was an understanding of how data should be collected on employees and where the selection methods have been v~lidated against whatever measure of success on the job the organization uses. The extent of research available, while not conclusive, did not support the random choices many school districts made on teacher selection. The research did not point toward the identification of universal guidelines, but rather toward validation of the selection methods each district chose to utilize.
PAGE 40
CHAPTER III PRESENTATION OF DATA As stated in Chapter I, the problem in this study was to determine the relative usefulness of personal history data, interview assessmen t and recommendations assessment when utilized in the selection of teachers in a single urban school distric t Three s amples of teachers from the single urban school district were chosen for the study. Sample A consisted of teachers who had performed satisfactorily enough to be offered a third contract. S ample B contained teachers in their first year of employment in the urban school district about which no decision had b een made concerning their reemployment. A third sample, Sample C, consisted of teachers who had not p erformed to a sufficiently satisfactory level to be retained by the urban school district. To determine the relative usefulness of the data gathered on the three samples the following operations were performed. 1. Frequency distributions were developed for Samples A and B. 2. Joint frequency distributions were developed for Samples A and B to show how the dependent 33
PAGE 41
variable (evaluation assessment) related to each of the independent variables (personal history data, interview assessment and recommendations assessment). 3. Regression equations were computed for Samples A and B to determine the relative predictive value of the independent variables. 4. Frequency distributions were developed to compare Samples A, Band C in regard to each of the independent variables. 34 The presentation of data in this chapter is contained in four sections to correspond to the four operations des cribed above. Basic Descriptive Data Relative to the Independent and Dependent Variables for Samples A a~d B Frequency distributions were developed on each of the nine variables for Samples A and B to provide a synthesis of the basic characteristics of the data in terms of distribution and variability. The distributions show the numerical frequency in each variable category, the percent in each variable category and the cummulative percent. In Tables 4, 7, and 8, missing data resulted in the percents becoming decimals. Obviously, the numerical frequencies remained whole numbers. Personal History Data for Samples A and B The following variables are included under personal history data: age, graduate of a university in Florida,
PAGE 42
35 years of previous experience, grade point average, marriage status and race. The distribution for age is shown in Table 1. In the age category 25 years and younger, Sample A had 59 percent and Sample B ~'.had 69 percent, a 10 percent difference. A 10 percent difference also existed in the category over 40 years. Sample A had 12 percent, Sample B has 2 percent in this category. The mean age for Sample A was 28.52 years as compared to a mean of 25.38 years for Sample B. This represented a difference of 3.12 years. Table 2 presents the distribution of teachers in Samples A and B who graduated from a university in Florida and those who graduated from a university outside the state of Florida. Sixty-two percent of the teachers in Sample A graduated from a university in Florida while 51 percent of Sample B graduated from a university in Florida. The distribution of years of previous experience is found in 'l'able 3. Sixty-two percent of Sample A had no previous teaching experience while 55 percent of Sample B were first year teachers. Twenty-five percent of Sample A were teachers with one to five years of experience while 37 percent of Sample B fell in this category. The mean for Sample A was 5.16 years of previous experience as compared to 3.60 for Sample B. Table 4 contains the distribution of grade point averages based on a four-point scale. The grade point average so computed was based on all college credit recorded in the personnel file of the individual teacher. The
PAGE 43
Table 1 Age Distribution for Samples A and B Sample A Age Cum. Category Freq. Percent Percent 25 years and under 59 59 59 26-30 yrs. 15 15 74 31-35 yrs. 10 10 84 36-40 yrs. 4 4 88 Over 40 yrs. 12 12 100 Total 100 100 100 Mean for Sample A= 28.52 years Mean for Sample B = 25.38 years Freq 69 17 5 7 2 100 36 Sample B Cum. Percent Percent 69 69 17 86 5 91 7 98 2 100 100 100
PAGE 44
Table 2 .. Distribution of Teacher s in Samples A and B W h o Graduate d from a University in Florida and Who Graduated from a UniveYsity Outside of Florida Sample A Sam ple B Cu m. Fre q Percent Percent Freq. P ercent Grad. of Univ in Florida 62 66 66 51 51 Grad. of Univ. outside of Florida 32 34 100 49 49 Missing Data 6 Total 100 100 100 100 100 37 Cum. Percent 51 100 100
PAGE 45
'able 3 Distribution of Years of Previous Experience for Samples A and B Sample A Yrs. of prev. Cum. exper. Freq. Percent Percent No. prev. exper. 62 62 62 1 yr. of exper. 10 10 72 2-5 yrs. of exper. 15 15 87 6-10 yrs. of exper. 10 10 97 Over 10 yrs. of exper. 3 3 100 Total 100 100 100 Mean for Sample A= 5.16 years Mean for Sample B = 3.60 years Sample B Freq. Percent 55 55 16 16 21 21 5 5 3 3 100 100 38 Cum. Percent 55 71 92 97 100 100
PAGE 46
Table 4 Distribution of Grade Point Averages for Samples A and B Sample .A.. Sample B Grade point average Freq. Percent 2.1 and below 2 2.5 2.2-2.5 18 22.2 2.6-3.0 37 45.7 3.1-3.5 23 28.4 3.6-4.0 1 1.2 Missing Data 19 Total 100 100.0 Mean for Sample A= 2.8 Mean for Sample B = 2.8 Cum. Percent Freq. Percent 2.5 5 5.1 24.7 18 18.2 7 0. 4 43 43.4 98.8 23 28.3 100.0 5 5.1 1 100.0 100 100.0 39 Cum Percent 5.1 2::L 2 66.7 94.9 100.0 100.0
PAGE 47
40 difference between Sample A and Sample B for the five grade point average categories was minimal. The greatest difference existed in categories which involved fewer than five teachers. The mean grade point average for Sample A was 2.81, for Sample Bit was 2.83. Marriage status is displayed in Table 5. The percent of married teachers who began teaching in 1971-72 (Sample A) was 60 percent. The corresponding figure for Sample B was 57 percent. Thirty-four p ercent of the teachers in Sample A were single, 40 percent in Sample B. The group of divorced or separated teachers represented 6 percent of Sample A and 3 percent of Sample B. Race is presented in Table 6. There were 22 percent black teachers for Sample A and 14 percent of Sample B were black. This resulted in 78 percent of Sample A and 86 percent of Sample B having been white. Interview Assessment Data for Samples A and B Interview assessment was represented by the numerical score attained by assigning values to personal characteristics listed on the interview rating sheet. The interviewer rated each applicant on a continuum for each characteristic with a potential range of scores from a most favorable rating of six to a most unfavorable rating of _.60:. The mean score for Sample A was 22.90, Sample B's mean was 19.23. This represented a difference of 3.61. Table 7 shows the interview assessment ratings by arbitrarily assigned categories. In
PAGE 48
Status Married Separated or Divorced Single Total Table 5 Distribution by Marriage Status for Sru~ples A and B Sample A Cum. Freq. Percent Percent Freq. 60 60 60 57 6 6 66 3 34 34 100 40 100 100 100 100 41 Sample B Cum. Percent Percent 57 57 3 60 40 100 100 100
PAGE 49
42 Table 6 Distribution by Race for Samples A and B Sample A Sample B Cum. Cum. Race Freq. Percent Percent Freq. Percent Percent Black 22 22 22 14 14 14 White 78 78 78 86 86 86 Total 100 100 100 100 100 100
PAGE 50
Cate-gories 6-15 16-20 21-25 26-30 Above 30 Missing Data Total Table 7 Interview Assessment Distribution for Samples A and B Sample A Cum. Freq. Percent Percent Freq. 9 9.5 9.5 21 25 26.3 35.8 38 37 38.9 74.7 30 14 14.7 89.5 7 10 10.5 100.0 1 5 3 100 100.0 100.0 100 Mean for Sample A= 22.90 Mean for Sample B = 19.29 43 Sample B Cum. Percent Percent 21.6 21.6 39.2 60.8 30.9 91. 8 7.2 99.0 1. 0 100.0 100.0 100.0
PAGE 51
the most favorable category, a score of 6-15, Sample A had 9.5 percent while Sample B had 21.6 percent. In the second most favorable category, Sample B was also higher, 39.2 percent to 26.3 percent. Twenty-five percent of Sample A was rated above 30, while 8.2 percent of Sample B received ratings above 30. Recommendations Assessment for Samples A and B Recommendations assessment was the numerical value assigned by the recommending party converted to a 1-10 44 scale as outlined in the procedures section. The distribution of ratings is shown in Table 8. A recommendation assessment of one represented the most positive rating while an assessment of ten represented the least positive rating. R atings of one, four and five-ten contained minimal differences between Samples A and B. Sample A had 20.9 percent with a rating of ~wo as compared to 26.1 percent for Sample B. In the category with a recommendation assessment of three, Sample A had 34.1 percent and Sample B had 26.1 percent. The means for Sample A and Sample B were 3.29 and 3.35 respectfully. Evaluation Assessment for Samples A and B Evaluation assessment was used as the dependent variable in the study. The distribution for Sample A is found in Table 10 and the distribution for Sample B appears
PAGE 52
Table 8 Recommendations Assessment Distribution for Samples A and B Sample Cate-gories Freq. Percent 1 8 8.3 2 19 20.9 3 1 1 34.1 4 13 14.3 5-10 20 22.0 Missing Data 9 Total 100 100.0 Mean for Sample A= 3.29 Mean for Sample B = 3.35 A Sample B Cum. Percent Freq. Percent 8.8 6 6.5 29.7 24 26.1 63.7 24 26.1 78.0 13 14.1 100.0 25 27.2 8 100.0 100 100.0 45 Cum. Percent 6.5 32.8 58.7 72.8 100.0 100.0
PAGE 53
Categories 1.0 1.1 1.2 1. 3 1.4-3.0 Total Table 9 Evaluation Assessment Distribution for Sample B Sample B Freq. Percent 72 72 15 15 6 6 2 2 5 5 100 100 Mean for Sample B = 1.05 46 Cum. Percent 72 87 93 95 100 100
PAGE 54
Pages are misnumbered following this insert
PAGE 55
Table 10 Evaluation Assessment Distribution for Sample A Sample A Categories Freq. Percent 1. 8 and below 20 20 1. 9-2. 0 21 21 2.1-2.2 29 29 2.3-2.5 20 20 2.6-5.0 10 10 Total 100 100 Mean for Sample A= 2.12 46 Cum Percent 20 41 70 90 100 100
PAGE 56
in Table 9. The evaluation assessment for Sample A represented the mean score of three annual evaluations for each teacher as completed by the principal. The evaluation assessment for Sample B represented the mean 47 of one annual evaluation as completed by the principal. The annual evaluation for Sample B contained thre e choices for rating the teachers: satisfactory, n e e d s improvement and unsatisfactory. This scale produced a smaller distribution of scores and necessitated using separate tables for Samples A and B. In both samples the lower the numerical value the more positive the evaluation. The evaluation assessment for Sample A were distributed somewhat evenly in each category: 20 percent, 21 percent, 29 percent, 20 percent and 10 percent respectfully. In Sample B, 72 percent of the teachers w ere in the most positive (lowest) category. Independent Variable Data in Relationship to the Dependent Variable for Samples A and B Joint frequency distributions were established to show how the dependent variable related to the independent variables. Separate tables were constructed to show each independent variable for Sample A related to the dependent variable and each independent variable for Sample B related to the dependent variable. Tables 11 through 26 showing these data follow the same format: 1. The five categories of the dependent variable (evaluation assessment) are shown on the left of each table.
PAGE 57
2. The categories of each independent variable are shown across the top of each table. 3. Each cell within each tables shows: a. The actual number of teachers. (s~e b ini .each t able)~ b. The p ercent of teachers in an evaluation category containe d in each independent variable c ategory (see c in each table). c. The percent of teachers in each independent variable category who were in each evaluation category (see d in each table). d. The percent of the total sample (see e in each table). 4. The numerical total and percent of the total sample in each evaluation category (see fin each table). 5. The numerical total and percent of the total sample in each independent variable category (see gin each table). 48 Personal History Data in Relationship to Evaluation Assessment Theindependent variables included in personal history data were: age, grac.uate of a university in Florida, years of previous experience, grade point average, marriage status and race. In Table 11, age is related to evaluation assessment for Sample A. Fifty-five percent of all teachers who received the most positive evaluation assessment (1.8 and below), were under 26 years old. (see c, Table 11). Fifty-nine percent of all teachers in Sample A were under 26 years old (see g, Table 11). The over 40 category
PAGE 58
49 Table 11 Age Related to Evaluation Assessment for Sample A Eval. Age cate-gory Under 26 26-30 31-35 36-40 Over 40 Row Total 1.8 11. 0 b 2.0 2.0 1.0 4.0 20.0 and 55.0 C 10.0 10.0 5.0 20.0 20.0 below a 18.6 d 13.3 20.0 25.0 33.3 11. 0 e 2.0 2.0 1. 0 4.0 13.0 3.0 2.0 1.0 2.0 21.0 1. 9-61.9 14.3 9.5 4.8 9.5 21. 0 2.0 22.0 20.0 20.0 25.0 16.7 13. 0 3. 0 2.0 1. 0 2.0 19.0 3.0 3.0 1.0 3.0 29.0 2.1-65.5 10.3 10.3 3. 4 10.3 29.0 2.2 32.2 20.0 30.0 25.0 25. 0 19.0 3.0 3. 0 1.0 3.0 10.0 5.0 2.0 1. 0 2.0 20.0 2.3-50.0 25.0 10.0 5.0 10.0 20.0 2.5 16.9 33.3 20.0 25.0 16.7 10.0 5.0 2.0 1. 0 2.0 6.0 2.0 1. 0 0.0 1.0 10 o 2. 6-60.0 20.0 10.0 0.0 10.0 10.0 5.0 10.2 13 3 10.0 0.0 8.3 6.0 2.0 1. 0 0.0 1.0 Column 59.0 15.0 10.0 4.0 12.0 100.0 Total g 59.0 15.0 10.0 4.0 12.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. f c. Percent of those in this evaluation category contained in this independent 7ariable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category~ g Numerical total and percent of total sample in this independent variable category.
PAGE 59
50 Table 12 Age Related to Evaluation Assessment for Sample B Eval. Age cate9ory Unde r 26 26-30 31-35 36-40 Over 40 Row Total 50.0 b 14.0 3.0 4.0 1.0 72.0 1. 0 a 69.4 C 19.4 4.2 5.6 1. 4 72.0 72.5 d 82.4 60.0 57.1 50.0 50.0 e 14.0 3.0 4.0 1.0 10.0 3.0 1. 0 0.0 1. 0 15.0 1.1 66.7 20.0 6.7 0.0 6.7 15.0 14.5 17.6 20.0 o.o 50.0 10.0 3.0 1. 0 0.0 1.0 3.0 0.0 1.0 2.0 0.0 6.0 1. 2 50.0 o.o 16.7 33.3 0.0 6.0 4.3 0.0 2 0. 0 28.6 0.0 3.0 o.o 1.0 2.0 0 0 2.0 0.0 0.0 0.0 0.0 2.0 1. 3 100.0 0.0 o.o 0.0 o.o 2.0 2.9 0.0 0.0 0.0 0.0 2.0 0.0 0.0 0.0 0.0 4.0 0.0 0.0 1. 0 0.0 5.0 1.4-75.0 0.0 o.o 25.0 o.o 5.0 1.5 5.7 0.0 0.0 14. 3 0.0 4.0 0.0 0.0 J.. 0 0.0 Column 69.0 17.0 5.0 7.0 2.0 100.0 Total g 69.0 17.0 5.0 7.0 2.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. f c. Percent of those in this eval~ation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 60
represented 12 percent of the sample and r eceived 20 percent of the ratings i n the most p ositive (1.8 and below) category. 51 In Sample B found in Table 12, 69 p ercent of all teachers were in the under 26 age category and 72 percent of all teachers under 26 received the mos t positive rating (1.0). Tables 13 and 14 show graduates of universities in Florid and outside Florida related to evaluation assessment for Samples A and B respectfully. A s noted previously (see Table 2), Sample A contained 66 percent Florida graduates, while Sample B had 21 percent. In Sample A, Florida graduates (representing 66 percent of the teachers) accounted for 47.1 percent of the most positive ratings. In Sample B, Florida graduates (representing 51 percent of the teachers) accounted for 51.4 percent of the most positive ratings. In Sample A, 30.6 percent of all Florida graduates were rated 2.0 or lower. The percent of graduates of out of state schools who received a 2.0 or lower rating was 56.1 percent. Previous teaching experience is shown related to evaluation assessment in Tables 15 and 16. In examining Sample A (Table 15) it is found that the category none (no previous experience) contained 62 percent of the sample and 45 percent of the most positive ratings (1.8 and below). The category 6-10 years contained 10 percent of the sample and received 25 percent of the most positive ratings.
PAGE 61
Table 13 Graduates of Universities in Florida and Graduates of Universities Outside of Florida Related to Evaluation Assessment for Sample A Grad. Grad. of Evaluation University University Row Category outside total in Fla. Fla. 8.0 b 9.0 17.0 1.8 and 47.1 C 52.9 18.1 below a 12.9 d 28.1 8.5 e 9.6 11.0 9.0 20.0 1. 9-2. O 55.0 45.0 21. 3 17.7 28.l 11.7 9.6 22.0 7.0 29.0 2.1-2.2 75.9 24.1 30.9 35.5 21. 9 23.4 7.4 16.0 3.0 19.0 2.3-2.5 84.2 15.8 20.2 25.8 9.4 17.0 3.2 5.0 4.0 9.0 2.6-5.0 55.6 44.4 9.6 8.1 12.5 5.3 4.3 Column 62.0 32.0 94.0 Total g 66.0 34.0 100.0 52 f a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 62
Table 14 Graduates of Universities in Florida and Graduates of Universities Outside of Florida Related to Evaluation Assessment for Sample B Grad. Grad. of Evaluation University University Row Category in Fla. outside total Fla. 37.0 b 35.0 72.0 1.0 a 51.4 C 48.6 72.0 72.5 d 71.4 37.0 e 35.0 6.0 9.0 15.0 1.1 40.0 60.0 15.0 11.8 18.8 6.0 : 9. 0 3.0 3.0 6.0 1.2 50.0 50.0 6.0 5.9 6.1 3.0 3.0 2.0 o.o 2.0 1.3 100.0 o.o 2.0 3.9 0.0 2.0 0.0 3. 0 2.0 5.0 1.4-1.5 60.0 40.0 5.0 5.9 4.0 3.0 2.0 Column 51. 0 49.0 100.0 Total g 51.0 49.0 100.0 53 f a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluction category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. N~merical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 63
Eval. cate-gory 1.8 and below a 1. 9-2.0 2.1-2.2 2.3-2.5 2.6-5.0 Column Total g Table 15 Years of Previous Experience Related to Evaluation Assessment for Sample A Years o f Previous Experience Over Row None One 2-5 6-10 10 total 9.0 b 2.0 3.0 5.0 1.0 20.0 45.0 C 10.0 15.0 25.0 5.0 20.0 14.5 d 20.0 20.0 50.0 33.3 9 0 e 2.0 3.0 5.0 1.0 13.0 3.0 2.0 3.0 0.0 21.0 61. 9 14.3 9.5 14.3 0.0 21.0 21.0 30.0 13.3 30.0 ,; o ~ o 13. 0 3.0 2.0 3 0 o o 18.0 2.0 7.0 0.0 2.0 29.0 62.l 6.9 24.1 0.0 6. 9 29.0 29.0 20.0 46.7 0.0 66.7 18.0 2.0 7.0 0.0 2.0 15.0 2.0 2.0 1.0 o.o 20.0 75.0 10.0 10.0 5.0 o.o 20.0 24.2 20.0 13.3 10.0 o.o 15.0 2.0 2.0 1.0 o o 7.0 1.0 1.0 1. 0 o o 10.0 70.0 10} 0 10.0 10.0 0 0 10.0 11.3 10.0 6.7 10.0 o.o 7.0 1. 0 1. 0 1.0 o.o 62.0 10.0 15.0 10.0 3.0 100.0 62.0 10.0 15.0 10.0 3.0 100.0 54 f a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable categor y who were in thi s evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 64
Eval. Category 1.0 a 1.1 1. 2 1.3 1.4-1.5 Colum n Total g Table 16 Years of Previous Experience Re lated to Evaluation Assessment for Sample B Years of Previous Experienc e Over Ro w f Non e On e 2 5 6 -10 10 Total 38.0 b 12.0 17.0 2 0 3.0 7 2.0 52.8 c 16.7 23.6 2.8 4.2 72.0 69.1 d 75.0 81. 0 40.0 100.0 38.0 e 12.0 17.0 2.0 3 0 9.0 2.0 3.0 1. 0 o.o 15.0 60.0 13. 3 20.0 6.7 o.o 15. 0 16.4 12.5 14.3 20.0 o.o 9.0 2.0 3.0 1.0 0.0 3.0 1. 0 1.0 1. 0 o.o 6.0 50.0 16.7 16.7 16.7 .0.0 6.0 5.5 6.3 4.8 20.0 0.0 3.0 1. 0 1. 0 1 0 o o 2.0 o.o o.o o.o o.o 2.0 100.0 0.0 o.o 0.0 o.o 2.0 3.6 0.0 o.o 0.0 o.o 2.0 0.0 o.o 0.0 0.0 3.0 1. 0 o.o 1.0 o.o 5.0 40.0 20.0 0.0 20 0 .. 0.0 5.0 5.4 6.3 0.0 20. 0 o.o 3.0 1.0 o.o 1. 0 0.0 55.0 16.0 21.0 5.0 3.0 100.0 55.0 16.0 21.0 5.0 3.0 100.0 55 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in thi s c ell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 65
56 The other categories contained minimal differences. In Sample B, shown in Table 16, 55 percent of all teachers had no previous experience and the category accounted for 51 percent of the most positive evaluations (1.0). Tables 17 and 18 show grade point average related to evaluation assessment. Grade point average was based upon a four-point scale and computed on all college credit recorded in the personnel file of the individual teacher. In examining the various categories of grade point average for Sample A, it is noted that 28.3 percent of all teachers achieved a grade point average of 3.1-3;5. This category accounted for 26.7 percent of the teachers who received the most f avorable evaluations (1.8 and b elow). In the category 2.2-2.5, which represented 22.2 percent of all teachers in Sample A, 26.7 percent received ratings in the evaluation category 1.8 and below. The results for Sample B were very similar. The category of 3.1-1~5 represented 28.3 percent of the sample and received 26.8 percent of the most positive ratings (1.0). The 2.2-2.5 category represented 18.2 percent of the sample and received 14.1 percent of the most positive ratings. Marriage status related to evaluation assessment for Samples A and Bis displayed in Tables 19 and 20 respectfully. In Sample A married teachers represented 60 percent of the sample and 65 percent of the most positive ratings (1.8 and below). Single teachers accounted for 34 percent
PAGE 66
Eval. cate9:ory 1. 8 and below a 1.9-2.0 2.1-2.2 2.3-2.5 2.6-5.0 Column total g Table 17 Grade Point Average Related to Evaluation Assessment for Sample A Grade Point Average Below 2.2-2.6-3.1-3.6-2.2 2.5 3.0 3.5 4.0 1. 0 b 4.0 6.0 4.0 0.0 6.7 C 26.7 40.0 26.7 o.o 50.0 d 22.2 16.2 17.4 0.0 1. 2 e 4. 9 7.4 4. 9 o.o 0.0 5.0 7. 0 5.0 1. 0 o.o 27.8 38.9 27.8 5.6 0.0 27.8 18.9 2 .1. 7 100.0 o.o 6.2 8.6 6.2 1.2 1.0 6.0 10.0 7.0 0.0 4.2 25.0 41.7 29,2 0.0 50.0 33.3 27.0 30.4 o.o 1. 2 7.4 12.3 8.6 0.0 o.o 3.0 6.0 6.0 o.o o.o 20.0 40.0 40.0 o.o o.o 16.7 16.2 26.1 0.0 0.0 3.7 7.4 7.4 0.0 0.0 0.0 8.0 1. 0 0.0 o.o 0.0 88.9 11.1 0.0 o.o o.o 21. 6 4.3 0.0 o.o o.o 9,9 1. 2 0.0 2.0 18.0 37.0 23.0 1.0 2.5 22.2 45.7 28.3 1. 2 Missj.ng Data -19 Row Total 15.0 18.5 18.0 22.2 24.0 29.6 15.0 18.5 9.0 11.1 81.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. f c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category. 57
PAGE 67
Eval. Table 18 Grade Point Average Related to Evaluation Assessment for Sample B Grade Point Average 58 cate-Below 2.2-2.6-3.1-3.6-Row f 2.2 2. 5 3.0 3.5 4.0 Total gory 5.0 b 10.0 33.0 19.0 4.0 71. 0 1. 0 a 7.0 C 14.1 46.5 26.8 5.6 71. 7 100.0 d 55.6 76.8 67.9 80.0 5.1 e 10.l 33.3 19.2 4.0 o.o 3.0 6.0 6.0 0.0 15.0 1.1 0.0 20.0 40.0 40.0 0.0 15.2 o.o 16.7 14.0 21. 4 o.o o.o 3.0 6. 1 6 T .J. 0.0 0.0 3.0 2.0 1. 0 0.0 6.0 1.2 0.0 50.0 33.3 16.7 o.o 6.1 o.o 16.7 4.7 3.6 0.0 o.o 3. 0 2 0 1.0 0.0 0.0 1.0 0 0 1.0 0.0 2.0 1.3 0.0 50.0 0.0 50.0 0.0 2.0 o.o 5.6 o.o 3.6 0.0 o.o 1.0 0.0 1.0 0 0 o.o 1.0 2.0 1. 0 l. 0 5.0 1. 4-o.o 20.0 40.0 20.0 20.0 5.0 1. 5 0.0 5.6 4.7 3.6 20.0 0.0 1. 0 2.0 1. 0 1.. 0 Column 5.0 18.0 43.0 28.0 5.0 99.0 Total g 5.1 18.2 43.4 28.3 5.1 100.0 Missing Data 1 a. b. c. d. e. f. g. The lower the numerical value, the more positive the rating. Actual number of teachers in the cell. Percent of those in t~is evaluation c ategory contained in this independent variable category. Percent of those in this independent variable category who were in this evaluation category. Percent of the tota l sample in this cell. Numerical total and percent: of total sample in this ~valuation category. Numerical total and percent of total sample in this independent variable category. f
PAGE 68
Eval. Cate9:ory 1.8 and below a 1.9-2.0 2.1-2.2 2.3-2.5 2.6-5.0 Column Total g Table 19 Marriage Status Related to Evaluation Assessment for Sample A Marriage Status Divorced or Married Separated Single 13.0 b 2.0 5.0 65.0 C 10.0 25.0 21. 7 d 33.3 14.7 13.0 e 2.0 5.0 13.0 1.0 7.0 61. 9 4.8 33.3 21. 7 16.7 20.6 13.0 1.0 7.0 16.0 2.0 11. 0 55.2 6.9 37.9 26.7 33.3 32.4 16.0 2.0 11.0 13.0 1.0 6.0 65.0 5.0 30.0 21.7 16.7 17.6 13.0 1.0 6.0 5.0 o.o 5.0 50.0 0.0 50.0 8.3 o.o 14.7 5.0 0.0 5.0 60.0 6.0 34.0 60.0 6.0 34.0 Row f Total 20.0 20.0 21.0 21.0 29.0 29.0 20.0 20.0 10.0 10.0 100.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. 59 c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 69
Eval. Cate9:ory 1.0 a 1.1 1.2 1.3 1.4-1.5 Column crotal g Table 20 Marriage Status Related to Evaluation Assessment for Sample B Marriage Status Divorced or Married SeEarate d Sin9:le 45.0 b 3.0 2 4.0 62.5 C 4.2 33.3 78.9 d 100.0 60.0 45.0 e 3.0 24.0 7.0 o.o 8.0 46.7 0.0 53.3 12.3 o.o 20.0 7.0 0.0 8.0 0 o.o 2.0 66.7 o.o 33. 3 7.0 0.0 5.0 4.0 o.o 2.0 o.o 0.0 2.0 o.o 0.0 100.0 o.o o.o 5.0 0.0 0.0 2.0 1.0 o.o 4.0 16.7 o.o 83.4 ::.1. 8 o.o ': 9. 0 1. 0 o.o 4.0 57.0 3.0 40.0 57.0 3.0 40.0 60 Row Totalf 72.0 72.0 15. 0 15.0 6.0 6.0 2.0 2.0 5.0 5.0 5.0 100.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 70
of the sample and receive d 25 percent of the most positive ratings. The same categories for Sample B revealed that 57 percent of the sample were married 61 and that married teachers receive d 62 percent of the most positive ratings. Forty percent of Sample B consisted of single teachers and this group received 33.3 percent of the most positive ratings. In examining the least positive evaluations as related to m arriage status, married teachers were given 30 percent of the ratings in the two least positive evaluation categories. Thirty-two percent of the single teachers were placed in the same two categories. The last variable in personal history d ata, race, is related to evaluation assessment in Table s 21 a n d 22. In Sample A, 78 percent of the teachers were white and this group received 75 percent of the most positive evaluations (1.8 and below). The 22 percent black teachers thus received 25 percent of the most positive ratings. Similarily, in Sample B, the white teachers represented 86 percent of the sample and received 86 percent of the most positive ratings while the 14 percent black teachers received 14 percent of the most positive ratings. Intervie w Assessment Related to Evaluation Assessment Interview assessment represented the numerical score attained by assigning values to the interview rating sheet.
PAGE 71
Table 21 Race Related to Evaluation Assessment for Sample A Eval. Race Cate-Row f gory Black White Total 5.0 b 15.0 20.0 l 8 and 25.5 C 75.0 20.0 below a 22.7 d 19.2 5.0 e 15.0 3.0 18.0 21.0 1.9-14.3 85.7 21.0 2.0 13.6 23.1 3.0 18.0 5.0 24.0 29.0 2.1-17.2 82.8 29.0 2.2 22.7 30.8 5.0 24.0 8.0 12.0 20.0 2.3-40.0 60.0 20.0 2.5 36.4 15.4 8.0 12.0 1. 0 9.0 10.0 2.6-10.0 90.0 lC.O 5.0 4.5 11. 5 1.0 9.0 Column 22.0 78.0 100.0 Total g 22.0 '78.0 100.0 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. 62 c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 72
Eval. cate9:or;y: 1.0 a 1.1 1.2 1.3 1.4-1.5 Column Total g Table 22 Race Related to Evaluation Assessment for Sample B Race Row Black White Total 10.0 b 62.0 72.0 13. 9 C 86.1 72.0 71.4 d 72.1 10.0 e 62.0 1.0 14.0 15.0 6.7 93.3 15.0 7.1 16.3 1. 0 14. 0 1. 0 5.0 6.0 16.7 83.3 6.0 7.1 5.8 1.0 5.0 0.0 2.0 2.0 o.o 100.0 2.0 o.o 2.3 o.o 2.0 2.0 3.0 5.0 40.0 60.0 5.0 14.3 3.5 2.0 3.0 14.0 86.0 100.0 14.0 86.0 100.0 f a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. 63 c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 73
64 These scores ranged from a possible low of six to a possible high of 60, with the low score representing the more positive assessment. In Table 23, the relationship between interview assessment and evaluation assessment for Sample A is displayed. The i n t erview category 0-15 represented 9.5 percent of the sample and r eceived 15.8 percent of the most positiv e ratings. The least positive interview assessment category 31-60, r epresented 10.5 percent of the sample and this group received 10.5 percent of the most positive ratings. A comparison of the same data for Sample Bin Table 24, reveals the most positive interview category (0-15) represented 21.6 p ercent of the sample and this group received 23.2 p ercent of the most positive evaluation ratings (1.0). The least positive interview category (31-60) represented 1.0 percent of the sample and this group received 1.4 percent of the most positive ratings. Recommendations Assessment in Relationship to Evaluation Assessment Recommendations assessment related to evaluation assessment for Sample A is shown in Table 25. The same variables for Sample B appear in Table 26. In Sample A, of those teachers who received the most positive recommendations assessment (1), 37.5 percent received the most positive evaluation rating (1.8 and below). Of those teachers who received the second most positive recommendations
PAGE 74
Eval. Cate-'l'able 23 Interview Assessment Related to Evaluation Assessment for Sample A Interview Assessment 65 gory 0-15 16-20 21-25 26-30 31-60 Row f Total 3.0 b 5.0 6.0 3.0 2.0 19.0 1.8 and 15.8 C 26.3 31.6 15.8 10.5 20.0 below a 33.3 20.0 16.2 21. 4 20.0 3.2 5.3 6.3 3.2 2.1 1.0 9.0 5.0 2.0 2.0 19.0 1.9-5.3 47.4 26.3 10.5 10.5 20.0 2.0 11.1 36.0 13.5 14.3 20.0 1.1 9.5 5.3 2.1 2.1 4.0 5.0 11.0 4.0 3.0 27.0 2.1-14.8 18.5 40.7 14.8 11.1 28.4 2.2 44.4 20.0 29.7 ?. 8. 6 30.0 4.2 5.3 11.6 4.2 3.2 o.o 4.0 10.0 3.0 3.0 20.0 2.3-0.0 20.0 50.0 15.0 15.0 21.1 2.5 o.o 16.0 27.0 21.4 30.0 0.0 4.2 10.5 3.2 3.2 1.0 2.0 5.0 2.0 0.0 10.0 2.6-10.0 20.0 50.0 20.0 0.0 10.5 5.0 11.1 8.0 13.5 14.3 o.o 1.1 2.1 5.3 2.1 o.o Column 9.0 25.0 37.0 14.0 10.0 95.0 Total g 9.5 26.3 38.9 14.7 10.5 100.0 Missing Data 5 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 75
Eval. Cate.,.. Table 24 Interview Assessment Related to Evaluation Assessment for Sample B Interview Assessment 66 gory 016-20 21-25 26-30 31-60 Row f Total 16.0 b 27.0 20.0 5.0 1.0 69 0 1.0 a 23.2 C 39.1 29.0 7.2 1.4 71.1 76.2 d 71.1 66.7 71.4 100.0 16.5 e 27.8 20.6 5.2 1.0 4.0 7.0 3.0 1.0 o.o 15.0 1.1 26.7 46.7 20.0 6.7 0.0 15.5 19.0 18.4 10.0 14.3 0.0 4.1 7.2 3.1 1.0 0.0 o.o 2.0 4.0 0.0 o.o 6.0 1.2 0.0 33.3 66.7 o.o 0.0 6.2 0.0 5.3 13.3 0.0 0.0 o.o 2.1 4.1 o.o 0.0 1. 0 l.O 0.0 0.0 0.0 2.0 1. 3 50.0 50.0 0.0 0.0 0.0 2.1 4.8 2.6 o.o o.o 0.0 1. 0 l.C 0.0 o.o 0.0 0.0 1.0 3.0 1. 0 o.o 5.0 1.4-o.o 20.0 60.0 20.0 0.0 5.1 1.5 0.0 2.6 10.0 14.3 o.o 0.0 1.0 3.1 1.0 0.0 Column 21. 0 38.0 30.0 7.0 1.0 97.0 Total g 21.6 39.2 30.9 7.2 1.0 100.0 Missing Data 3 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 76
Table 25 Recommendations Assessment Related to Evaluation Assessment for Sample A Eval. Recommendations Assessment Cate-67 Row f 9"ory 1 2 3 4 5-10 Total 3.0 b 6.0 5.0 o.o 4.0 18.0 1.8 and 16.7 C 33.3 27.8 0.0 22.2 1~ 8 below a 37.5 d 31.6 16.1 0.0 20.0 3.3 e 6.6 5.5 o.o 4.4 0.0 5 0 7.0 5.0 3.0 20.0 1.9-0.0 25.0 35.0 25.0 15.0 22.0 2.0 o.o 26.3 22.6 38.5 15.0 0.0 5.5 7.7 5.5 3.3 2.0 4.0 11. 0 5.0 5.0 27.0 2.1-7.4 14.8 40.7 18. 5 18.5 29.7 2.2 25.0 21. l 35.5 38.5 25.0 2.2 4.4 12.1 5 5 5.5 3.0 3.0 4.0 1.0 7.0 18.0 2.3-16.7 16.7 22.2 5.6 38.9 1S.8 2.5 37.5 15.8 12.9 7.7 35.0 3.3 3.3 4.4 1.1 7.7 0.0 1.0 4.0 2.0 1.0 8.0 2.6-0.0 12.5 50.0 25.0 12.5 8.8 5.0 o.o 5.3 12.9 15.4 5.0 0.0 1.1 4.4 2.2 1.1 Column 8.0 19.0 31.0 13.0 20.0 91.0 Total g 8.8 20.9 34.1 14.3 22.0 100.0 Missing Data 9 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. 'Numerical total and percent of total sample in this independent variable category.
PAGE 77
Table 26 Recommendations Assessment Rel a ted to Ev aluation Assess m e n t for Sample B Eval. Recommenda t i ons Assessm ent Cate-68 Row f 9:ory 1 2 3 4 5-10 Tota l 4.0 b 18.0 17. 0 9.0 16.0 64.0 1.0 a 6.3 C 28. 1 26.6 14.1 25.0 69.6 66.7 d 75.0 70.8 69.2 64.0 4.3 e 19.6 18.5 9 8 17.4 : : 1. 0 5.0 2.0 1. 0 6.0 15.0 1.1 6.7 33.3 13.3 6.7 40.0 16.3 16.7 20.8 8.3 7.7 24.0 1.1 5.4 2 2 1.1 6.5 1.0 1.0 1.0 2.0 1.0 6.0 1.2 16.7 16.7 16.7 '33. 3 16.7 6.5 16.7 4.2 4.2 15.4 4.0 1.1 1.1 1.1 2 2 1.1 0.0 0.0 2.0 o.o o.o 2.0 1.3 o.o 0.0 100. 0 0.0 0.0 2.2 0.0 0.0 8.3 0.0 0 0 o.o 0.0 2 2 0.0 0.0 0.0 0.0 2.0 1.0 2.0 5.0 1.4-0.0 o.o 40.0 20. 0 40.0 5.4 1.5 o.o o.o 8.3 7.7 8.0 0.0 0.0 2.2 1.1 2.2 Column 6.0 24.0 24.0 13.0 25.0 92.0 Total g 6.5 26.1 26.1 14.1 27.2 100.0 Missing Data 8 a. The lower the numerical value, the more positive the rating. b. Actual number of teachers in the cell. c. Percent of those in this evaluation category contained in this independent variable category. d. Percent of those in this independent variable category who were in this evaluation category. e. Percent of the total sample in this cell. f. Numerical total and percent of total sample in this evaluation category. g. Numerical total and percent of total sample in this independent variable category.
PAGE 78
69 (2), 31.6 percent received the most positive evaluation rating. In the entire sample, 19.8 p ercent received ratings in the most positive category. Teachers in the two most positive recommendations categories (1,2) represented 29.7 percent of the sample. These teachers received 50 percent of the most favorable evaluation ratings. Teachers in Sample A who received the least positive (5-10) recommendations represented 22 percent of the sample. Teachers in that category (5-10) received 22 percent of the most positive evaluations (1.8 and below). In Sample B, as found in Table 26, 64 percent of the teachers with the least positive recommendations assessment received evaluation ratings in the most positive category (1.0). The most positive recom~endations assessment category (1.0), contained only six teachers, but four (66.7 percent) achieved the most positive evaluation rating (1.8 and below). Infue second most positive recommendations assessment category (2.0), 75.0 percent of the teachers fell in the most positive evaluation category. The remaining recommendations categories (3,4) had 70.8 percent and 69.~ percent, respectfully, in the most positive evaluation categories. Results of Regression Analysis for Samples A and B Multiple regression is an extension of the use of the bivariate correlation coefficient to multivariate analysis.
PAGE 79
Multiple regression permits the study of the linear relationships between a set of independent variables and 70 a dependent variable while taking into account the interrelationships among the independent variables. The basic purpose of multiple regression is to produce a linear combination of independent variables which will correlate as highly as possible with the dependent variable. The linear combination can then be used to predict the values of the dependent variable.1 Regression equations were established for Sample A and Sample B incorporating the following independent variables: age, ~raduate of a university in Florida, years of previous experience, grade point average, marriage status, race interview assessment and recommendation assessment. The sUTIL~ary tables for the regression equations present the following information: multiple R, R-square, R-square change, Band Beta. Multiple R indicates the linear relationship between the independent variables and the dependent variable under investigation. R-square can be interpreted as the proportion of the variance in the dependent variable accounted for by the regression equation. R-square change is the change in R-square from the value of the previous step. The Band Beta values are the 1william C. Mitchell, "Multiple~Regression Analysis: Subprogram Regression,'' in Statistical Package for the Social Sciences, edited by Norman H. Nie, Dale H. Bent and C Hadlai Hull, (New York: McGraw-Hill Book Company, 1970), p. 175-95.
PAGE 80
71 regular and normalized regression coefficients, respect-2 fully. The summary table and prediction equation for Sample A is shown in Table 27. The correlation coefficients listed under multiple R represent the intercorrelation of the independent variables included to that point as well as the correlation between the independent variable and the dependent variable. The last dorrelation coefficient, .25153, represents the strength between the dependent variable and the eight independent variables taken together. R ~square, the proportion of the variance in the dependent variable accounted for by the regression equation is found to be .05834 percent. Multiple R for Sample B appears in Table 28. The total multiple R for Sample Bis .31898 The proportion of variance accounted for in the dependent variable for Sample Bis .10175. The prediction equation for Sample A appears below Table 27 while the prediction equation for Sample Bis below Table 28. The prediction equation is designed to predict the value of the dependent variable, evaluation assessment. A practical example is included to help in understanding the regression equation. A hypothetical applicant from Sample Bis used, with the following characteristics: 24 years old, a graduate of the University of Florida, no previous experience, a grade point average of 3.21, single, white, an interview assessment of 19 and a 2rbid., p. 176.
PAGE 81
72 recommendations assessment of two. Evaluation Assessment x9 = (.01968) (24) + (-.06769) (1) + (-.001099) (O) + (o.:oro99) (3.21) + (.46941) (2) + (-.37482) (2) + (.02097) (19) + (.09234) (2) + 9.20913 The actual value of the variable is used in the equation except for: graduate of a university in Florida where a graduate equals one and a non-graduate equals two, marriage status where married equals one and single, divorced or separated equals two, and race where black equals one and .white equals two. The 9.20913 is a constant as shown in the table. The predicted evaluation assessment for the hypothetical teacher would 1.3913 whereas the mean for Sample Bis 1.05. A Comparison of the Distribution of the Independent Variables for Samples A, B, and C -This section presents the comparison of the distribution of the independent variables: personal history data (age, graduate of a university in Florida, years of previous experience, grade point average, marriage status and race), interview assessment and recommendations assessment, for Samples A, B, and C. Sample A consisted of 100 teachers, randomly chosen, who were hired in 1971-72 and who were in continuous employment in the urban school district to January, 1974. This sample represented teachers who performed satisfactorily enough to be offered a third contract with the urban school district. Sample B consisted of 100 teachers, randomly chosen, who were initially
PAGE 82
Table 27 Multiple Regression Analysis for Evaluation Assessment for Sample A Variable x1 Age x2 Graduate of Univ. in Florida x3 Yrs. of Prev. Ex. x4 Grade Pt. Ave. x5 Marriage Status x6 Race x7 Interview Assess. x8 Recommendations Assessment Constant Prediction Equation R2-Proportion B Normalized of variance 2 Regression R e gression Multiple R accounted for R Change Coefficient Coef. f i c i e n t 0.08423 0.00709 0.00709 0.00125 0.00363 0.13970 0.01952 0.01242 -0.69066 -0.11617 0.16047 0.02575 0.00624 -0.08973 -0.11589 0.16610 0.02759 0.00184 0.01238 0.04282 0.17308 0.02996 0.00237 0.53363 0.078677 0.17359 0.03013 0.00018 0.34061 0.04246 0.22557 0.05088 0.02075 0.06705 0.15803 0.24153 0.05834 0.00745 -0.17113 -0.08830 19.70982 I\ x9 = ( ,0.00125)Xl + (-0.69066)X2 + (-0.08973)X3 + (0.01238)X4 + ( 0 5 3 3 6 3 } XS + ( 0 3 4 0 61 ) X 6 + ( 0 0 6 7 0 5 ) X 7 + ( -0 ,." l 7113 ) X B + 19 7 0 9 8 2 -..J w
PAGE 83
Table 28 Multiple Regression Analysis for Evaluation Assessment for Sample B R2-proportion B Normalized of variance 2 Regression R e gression Variable Multiple R accounted for R Change Coeff i c i e n t Coefficient xl Age 0.02029 0 0 0041 0.00041 0.01968 0.10364 x2 Graduate of Univ. 0.05175 0.00268 0.00227 -0.06769 -0.03131 in Fl01,ida X3 Yrs. of Prev. Ex. 0.10544 0.01112 0.00844 -0.01754 -0.04703 x4 Grade Pt. Ave. 0.10632 0.01130 0.00019 0.01099 0.05366 XS Marriage Status 0.24694 0.06098 0.04967 0.46941 0.21500 x6 Race 0.26961 0.07269 0.01171 -0.37482 -0.12032 X7 Interview Assess. 0.28889 0.08346 0.01077 0.02097 0.11082 x8 Recommendations 0.31898 0.10175 0.01830 0.09234 0.13886 Assessment Constant 9.20913 Prediction Equation -x9 = (0.01968)X1 + (-0.06769)X 2 + (-0.01754)X3 + (0.01099)X4 + (0.4694nx5 + (-0.37482)x6 + (o.02097)x7 + (0.09234)x8 + 9.20913
PAGE 84
employed in 1973-74 by the urban school district and who were in continuous employment to Januar~ 1974. This sample constituted the sample about whom no decision had been made regarding their continued employment in the 75 urban school district. Sample C consisted of 21 arbitrarily chosen teachers, who had not performed at a sufficiently satisfactory level to be retained in the urban school district and at the time of the investigation were no longer employed. The teachers in this sample were employed in the urban school district for varying lengths of time during the period 1971-73. In Table s 29-36, the comparisons of the three sampla s on each 0f the independent variables by frequency p ercents are shown. Missing data in Sample Chas rendered the comparison on grade point average relatively useless. The chi-square test of independence was performed on all independent variables except grace point average. The purpose of the chi-square test is to test the null hypothesis: Samples A, B, and Care from the same population. The chi-square obtained for each sample and the chi-square from a table in Guilford, at the probability level of .01 for the appropriate degrees of freedom, is shown in each table.3 A chi-square that is greater than that found in the table at the 01 level, provides th, basis for rejecting the null hypothesis of no difference in the three samples. A chisquare less than that found in the table in Guilford permits 3J._ P._ Guilford, Fundament.al Statistics in Psychology and Education (New York: McGraw-Hill Book Co., 1956) p. 145.
PAGE 85
76 acceptance of the null hypothesis.4 The variable~ age, graduate of a university in Florida, years of previous experience a n d marriage status contain chi-squares sufficiently low to accept the null hypothesis. The variables race interview assessment and recommendations assessment have chi-squares large enough to reject the null hypothesis. The comparison of age distribution is found in T able 29. In the age category 25 and under, Sample A had 59.0 percent, Sample B had 69.0 percent and Sample Chad 38.1 percent. In the category 40 and over, Sample A had 12.0 percent while Sample Band Sample Chad 2.0 p ercent and 19.0 percent respectfully. The value of chi-square computed on these distributions was 15.70, which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted. Table 30 shows the comparison of graduates of universities in Florida and graduates of universities outside of Florida for Samples A, B, and C. Graduates of universities in Florida accounted for 66.0 percent, 51.0 percent and 61.9 percent for Samples A, B, and C respectfully. The value of chi-square computed on these distributions was 4.69 which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted. 4Ibid., p. 145.
PAGE 86
'l.'able 29 Comparison o f Age Distribution for Samples A, B, and C Age Sample A Category Percent 25 and unde r 59.0 26-30 15.0 31-35 10.0 36-40 4.0 40 and 12.0 Over Total 100.0 Computed chi-square= 15.70 Critical value of chi-square Sample n Percent 69.0 17.0 5.0 7.0 2.0 100.0 at .01 level of significants= 20.09 77 Sample C Percent 38.1 23.8 9.5 9.5 19.0 100.0
PAGE 87
Table 30 Comparison of Graduate of Universities in Florida and Graduates of Universities Outside of Florida for Samples A, B, and C Sample A Percent Grad. of univ. in 66.0 Florida Grad. of univ. out-34.0 side of Florida Total 100.0 Computed chi-square = 4.69 Critical value of chi-square at .01 level of significance= Sample B Sample Percent Percent 51. 0 61. 9 49.0 38.1 100.0 100.0 9.21 78 C
PAGE 88
Years of previous experience is compared in Table 31 for Samples A, B, and C. The percent of teachers with no previous experience was 62.0 percent for Sample A, 55.0 percent for Sample Band 47.6 percent for Sample C. Teachers with 2-5 years of e xperience were found to constitute 15.0 percent for Sample A and 21.0 percent and 33.3 percent for Samples Band c. The value of chi-square computed on these distributions was 7.22 which was not significant at beyond the .01 level of significance. Thus at this significance level the null hypothesis is accepted. Table 33 shows m arriage status for Samples A, Band C. Sixty percent of Sample A teachers were married while 57 percent of Sample Band 57 percent of Sample C fell in this category. The largest difference occurred in the category divorced or separated where Sample A showed 6.0 percent, Sample B-3.0 percent and Sample C-19.0 percent. The value of chi-square computed on these distributions was .19 which was not significant at beyond the .01 level of significance. Thus, at this significance level the null hypothesis is accepted. The last personal history data variable, race, is shown in Table 34. As can be seen from the table, Sample A had 22.0 percent black teachers, Sample B had 14.0 percent black teachers and Sample Chad 57.1 percent black teachers. The value of chi-square computed on these distributions was 16.1 which was significant at the .01 level of significance. At this significance level the null hypothesis is rejected.
PAGE 89
Table 31 Comparison of the Distribution of Years of Previous Experience for Samples A, B, and C Yrs. of Sample A Sample B Sample previous experience Percent Percent Percent No previous 62.0 55.0 47.6 experience 1 year 10.0 16.0 14.3 2-5 years 15.0 21. 0 33.3 6-10 years 10.0 5.0 4.8 Over 10 years 3.0 3.0 0.0 Total 100.0 100.0 100.0 To compute the chi-square value, cell~ in the over 10 years category were combined with cells in the 6-10 years category. Using these combined cells the computed value of chi-square was 7.22. The critical value of chi-square for six degrees of freedom at the .01 level was 16. 81. so C
PAGE 90
Table 32 Comparison of the Distribution of Grade Point Average for Samples A, B, and C Bl Grade Sample A Sample B Sample C point average 2.1 and 2.5 below 2.2-2.5 22.2 2.6-3.0 45.7 3.1-3.5 28.4 3.6-4.0 1. 2 Total 100.0 Missing data: Sample A -19 Sample B -1 Sample C -10 5.1 9.1 18.2 72.7 43.4 28.3 18.2 5.1 100.0 100.0
PAGE 91
Status Married Separated or Divorced Single Total Table 33 Comparison of Marriage Status for Sample s A, B, and C Sample A Sample B Percent Percent 60.0 57.0 6.0 3.0 34.0 40.0 100.0 100.0 Sample C Percent 57.1 19.0 23.8 100.0 To compute the chi-square value, cells in the divorced or separated categor y were combined with the cells in the single category. Using these combined cells the computed valu e of chi-squar e was .19. The critical value of chi-square for two degrees of freedom at the .01 level was 9.21. 82
PAGE 92
Table 34 Comparison of Race for Samples A, B, and C Race Black White Total Sample A Percent 22.0 78.0 100.0 Computed chi-square= 16.04 Critical calue of chi-square Sample B Percent 14.0 86. 0 100.0 at .01 level of significance= 9.21 Sample C Percent 57.1 42.9 100.0 83
PAGE 93
84 The comparison of interview assessment is found in Table 35. The lower the numerical score the more positive the intervie w rating. In the most positive category 6-15, Sample A had 9.5 percent, Sample B had 21.6 percent and Sa.~ple Chad 11.8 percent. The greatest difference in interview assessment occurred in the least positive inter-view category 30-60. In this category Sample A had 10.5 percent, Sample B had 1.0 percent and Sample Chad 29.4 percent. The value of chi-square computed on these distributions was 29.12 which was significant at the .01 level of significance. At this significance lev~l the null hypothesis is rejected. Table 36 contains the distribution of recommendations assessment for Samples A, B, and C. The lower the numerical score the more positive the recommendations assessment. The total percent for the two most positive recommendations categories are as follows: Sample A -29.7 percent, Sample B -32.6 percent and Sample C -6.3 percent. In the least positive category 5 -10, Sample A had 22.0 percent, Sample B had 27.2 percent and Sample C 68.8 percent. The value of chi-square computed on these distributions was 22.20 which was significant at the .01 level of significance. At this significance level the null hypothesis is rejected.
PAGE 94
Table 35 Comparison of the Distribution of Interview Assessment for Samples A, B, and C Categories 6-15 16-20 21-25 26-30 Above 30 Total Sample A Percent 9.5 26.3 38.9 14.7 10.5 100.0 Missing Data: Sample A -5 Sample B -3 Sample C -4 Sample B Percent 21.6 39.2 30.9 7.2 1. 0 100.0 Sample C Percent 11.8 11. 8 29.4 17.6 29.4 100.0 85 To compute the chi-square value, cells in the above 30 category were combined with the cells in the 26-30 category. Using these combined cells the computed value of chi-square was 29.12. The critical value of chisquare for six degrees of freedom at the .01 level was 16.81.
PAGE 95
Table 36 Comparison of the Distribution of Recorr~endations Assessment for Samples A, B, and C Categories Sample Percent 1 8.8 2 20.9 3 34.1 4 14.3 5-10 22.0 Total 100.0 Missing data: Sample A -9 Sample B -8 Sample C -5 A Sample B Sample C Percent Percent 6.5 o.o 26.1 6.3 26.1 6.3 14.1 18.8 27.2 68.8 100.0 100.0 To compute the chi-square value, cells in the recommendations category (1) were combined with the cells in the category (2). Using these combined cells the computed value of chi-square was 22.20. The critical value of chi-square for six degrees of freedom at the .01 significance level was 16.81. 86
PAGE 96
CHAPTER IV DISCUSSION OF THE DATA The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers in a single urban school district. Data were gathered on three samples of teachers who performed in the urban school district. Sample A consiste of teachers who had performed satisfactorily enough to be offered a third contract. Sample B consisted of teachers in their first year of teaching and about whom no decision had been made concerning their continued employment. Sample C represented teachers who did not perform to a sufficiently satisfactory level to be retained under contract in the urban school district. Regression equations were established for Sample A and Sample B to determine the relative predictive value of personal history data, interview assessment and recommendations assessment when principal's annual evaluation was used as the criterion measure. The regression equation for Sample A is shown in Table 27. Multiple R represents the intercorrelation of the independent variables as well as the correlation between the independent variable and the 87
PAGE 97
88 1 dependent variable, evaluation assessment. The value of multiple R for Sample A was .25153. A correlation of this size was described by Guilford as a definite but small relationship.2 R-square represents the proportion of the variance in the dependent variable accounted for by the regression equation. For Sample A, R-square was .05834, -meaning that the variables included in this equation accounted for less than 6 percent of the variance in the dependent variable, evaluation assessment. The results for Sample B were only slightly better. Multiple R for Sample B, found in Table 28, was .31898. A multiple R of this size is also described by Guilford as a definite but small relationship.3 R-square, the proportion of the variance in the dependent variable accounted fon by the regression equation, was slightly higher than in Sample A, but at .10175 it accounted for just over 10 percent of the variance. Due to the manner in which the correlation coefficients between the independent variables and the dependent variable were presented in this study (multiple R's), the exact correlation coefficient for any one variable can not be observed. The multiple R 1isted fo~ each independent variable 1J. P. Guilford, Fundamental Statistics in Psychology and Education, (New York: M~Graw-Hill Book Company, 1956), p. 145. 2rbid., p. 145. 3rbid., p. 145.
PAGE 98
89 represents the relationship between not only the dependent variable but the interrelationships of the various independent v ariables to that point. The difference s between the coefficients does give some idea of the relatively low coefficients obtained for each variable. Correlation coefficients below .40, as found in both Samples A and B, were cited in studies reported by Dunnette in 1963, as a 4 basis for disenchantment with prediction models. A study by Smith, Albright, Glennen and Owens reported achieving correlation coefficients as high as ~613 when comparing 484 personal history items to supervisory ratings.5 The findings in this study are in contradiction to the present study where no practical usefulness was found for personal history data. However, it is important to note that Smith, Albright, Glennen and Owens used 484 personal history items where as in the present study only six were included.6 Recommendations assessment added little to the intercorrelation of independent variablei or the correlation between it and the dependent variable, evaluation assessment. This fact supports findings by Bolton that letters of recommendation have not produced usable results in the .., selection of recruits.' 4M. D. Dunnette, "A and Selection Research," 317-23 (October 1963). There are at least two possible Modified Model for Test Validation Journal of Applied Psychology 47: 5w. I. Smith, L. E. Albright, J. R. Glennen, and W. A. Owens, "The Prediction of Research Competence and C reativity from Personnel History," Journal of Applied Psychol_ ?9Y 4 5: 59-62 (February 1961).
PAGE 99
90 reasons that the recommendat.ions produced such low correlations. First, the urban school district did not have a policy that governed who signed letters of recommendation and they accepted letters in all sorts of forms. On this point, the United States Civil Service Commission as the result of some of its work, recommended for improvement of the usefulness of letters of recommendation that it ought to be signed by a line executive in charge of the 8 employee. Secondly, even though there was a district-developed form available, its use was not required; as such, the district had not carefully defined discriminating terms to be commonly used. In this regard, as a result of a study by Peres and Garcia in which they synthesized adjectives used in recommendations, it was suggested that the identification of phrases would be helpful in establishing a more Q refined assessment procedure for handling recommendations.-Interview assessment produced correlation coefficients of disappointingly small magnitude just as recommendations 6rbid., p. 61. 7nale L. Bolton, Selection and Evaluation of Teachers, (Berkeley: Mccutchan Publishing Corporation, 1973), p. 65. 8u. s. Civil Service Commission, A Guide for Executive s~~ection, Personnel Methods Series, No. 13, Washington, D. C.: Government Printing Office, 1961. 9sherwood H. Peres and J. Robert Garcia, "Validity and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants," Personnel Psychology 16: 279-86 (Autumn 1962).
PAGE 100
91 assessment and personal history data. One problem with interviews, as was pointed out by Bolton, is that interviews can serve many purposes but a singular purpose must be bl. h d f b ff lO esta. is e or any one interview to e e_ ective. No evidence was found on the urban school district interview assessment sheet that indicated tha t the interview was d d d d b C ll structure or unstructure Stu ies reporte y rissy stressed that structured interviews were the only reliable form, while studies by Yonge12 supported structured or unstructured interviews as long as the purpose of the interview was clear. Bolton suggested that the validity of the interview depended: upon the skill of the interviewer and that interviewers needed to be trained.13 The measure of success and thus the basis for determining if an independent variable was an effective predictor of success was based upon the principal's annual evaluation of the teacher. There was a considerable dif ference in the mean of the evaluation assessments for Sample A and Sample B due to the urban school district changing the evaluation form for the 1973-74 school year. 10 Bolton, p. 67. 11w. J.E. Crissy, "The Areas, Methods and Results," 73-85 (Summer 1952). Employment Interview-Research Personnel Psychology 5: 12 K. A. Yonge, and a Pilot Study," (February 19 5 6) "The Value of the Interview: An Orientation Journal of Applied Psychology 40: 25-31 13 Bolton, p. 74.
PAGE 101
92 Sample B produced a higher multiple Rand R-square than Sample A, indicating a more favorable prediction equation. However, since general evaluation instruments were used to measure both Samples A and D, and this is contrary from what most experts advise, it. is not argued that the prediction equation for Sample Bis better tha n Sample A if the criterion is a successful teacher.: The prediction equation for Sample Bis better only if one uses the criterion measure used for this study.Dunnette advocated that the selection process should avoid assuming that there is a single all-encompassing measure of occupation~l 14 success. A study by Bass suggested that criteria change over a period of time, thus a person judged successful at one point in time may be evaluated lower in the future for h b h 15 t e same e avior. To avoid the situation where evalua-tors vary on the factors used to rate employees, Bolton suggested that: 1. The number of behaviors listed should be limited. 2. Behaviors should be related to the purpose of the situation in which they are to be exhibited. 3. Behaviors should be measurable. 4. Behaviors should include verbal interactions as well as non-verbal communications, and in-class as well as out-of-class action.16 14 Dunnette, p. 320. 15 Bernard M. Bass, "The Leaderless Group Discussion," Psychological Bulletin 51: 465-92 (September 1954). 16 Bolton, p. 58.
PAGE 102
93 The results of the joint frequency distributions did not provide prediction equations, but some observations of the data may assist personnel officers in the selection of teachers. The first variable related to evaluation assessment, age, was previously shown to have a v ery low correlation to evaluation assessment. Furthermore, data in Table 11, show that none of the arbitrarily chosen age categories received more than 33 percent of the most positive ratings and in the category with 33 percent (over 40) only four teachers were involved. As can be s een from Table 12, similar results were obtained for Sample B. A larger percent (28.1) of the graduates of universities outside of Florida received the most positive ratings (1.8 and below) than graduates of universities in Florida (12.9), for Sample A. But in Sample B, 72.5 percent of graduates of universities in Flo~ida received the most positive ratings (1. 0) while 71.4 percent of graduates from universities outside of Florida fell in the same category. Thus, like age, the variable grad~ate of a university i n Florida does not seem to offer any help to personnel administrators. The joint frequency distributions for years of previous experience also produced inconclusive results. In examining the variable marriage status, it was found that 21.7 percent of the married teachers received the most positive evaluation ratings for Sample A, while 14.7 percent of the single teachers were in this evaluation category. In Sample B, the percents for the corresponding classifications
PAGE 103
94 were 78.6 percent for Sample A and 60.8 percent forsample B. The distributions 011 this variable provided no greater assistance than previous variables. It was anticipated that intervi~w assessment would produce positive relationships with evaluation assessment. However, as is shown in Tables 23 and 24, no particular category of interview assessment received a disproportionately large share of the most positive evaluation ratings for either Sample A or Sample B. It should be noted that there was a difference of the mean interview assessment ratings between Sample A and Sample B of 3.61. Although not reported elsewhere in the study, it was known by the writer that a different set of interviewers rated the teachers in Sample A than rated the teachers in Sample B. (Crissy argued that intra-rater and inter-raer con-. f d 1 f 1 b"l' 17, sistency was a un amenta aspect o interview re ia i ity. 1 Thus it would seem that the variable interview assessment suffered from inconsistent interview ratings. Recow.mendations assessment related to evaluation assessment for Samples A and B also produced disappointing results. The relationship of the most positive recorrunendations assessment and the most positive evaluation ratings resulted in favorable percents in Sample A (see Table 25). However, in Sample B, 64 percent of the teachers who received the least positive recommendations assessment received the most favorable evaluation ratings.
PAGE 104
95 The comparison of the distribution of the independent variables for Samples A, B, and C was done to determine if significant difference s existed among Sample A which consisted of teachers who performed satisfactorily enough to be offered a third contract, Sample B which consisted of teachers about whom no decision had been made regarding their continued employment, and Sample C, consisting of teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district. The chisquare test of independence was performed on all variables except grade point average. Variables o~ age, graduate of a university in Florida, years of previous experience and marriage status contained chi-square values sufficiently low to accept the null hypothesis that the three samples came from the population. The variables of race, interview assessment and recommendations assessment had chi-squares large enough to reject the null hypothesis. In Samples A and B, the preponderance of teachers were white, but Sample C contained a majority of black teachers. There was no evidence in the joint frequency distributions that black teachers receive d proportionately poorer evaluations. During the period of time (from 1971-1973) that the teachers in Sample C were employed, the urban school district underwent e xtensive court ordered desegregation and many black teachers were forced to teach in predominately white schools. Interview assessment, displayed in Table 35 shows that an obvious disproportionate percent of Sample C received the
PAGE 105
least positive interview assessment. A simiiar situation existed with the variable recommendations assessment which appears in Table 36. Nearly 70 percent of Sample C received recommendations in the least positive category. 96 Considering the present investigation in its totality, it appears reasonable to draw three major generalizations. First, there is limited predictive value in personal history data, interview assessment and recorrunendations assess ment as used in this study. That is, these variables have little practical value to the personnel administrator concerned with developing a more sophisticated teacher selection model. Second, it becomes apparent when the present investigation is viewed in relation to the review of literature, that part of the problem in developing a prediction model in a school district such as the one involved in the present study is that data collection methods and criterion measures are unsophisticated and ignore previous studies. Third, the one practical implication for personnel officers that tends to emerge is related to interview assessment and recorrunendations assessmeYit. Although for Samples A .. arid B the predictive value of these two variables was nil, it can be concluded, based upon data from Sample C, that teachers with the least positive interview ratings and recommendations ratings, are more likely to be unsuccessful than teachers who receive the most positive ratings. Personnel officers who choose to select teachers who have received poor interview ratings and poor recommendations are quite likely to select teachers who have a high probability of failure.
PAGE 106
CHAPTER V SUMMARY, CONCLUSIONS AND IMPLICATIONS Summary Numerous methods have been used by school systems in attempting to select the best qualified teacher applicants. Research conducted on the effectiveness of these methods has not been conclusive. It would appear that there have been no universal guidelines developed that apply to a cross-section of school districts. In view of the wide range of techniques used and obvious need to improve the selection process, the present study was undertaken. The problem in this study was to determine the relative usefulness of personal history data, interview assessment and recorrunendations assessment when utilized in the selection of teachers in a single urban school .district. Specifically, the focus was on the following: 1. For a sample of teachers who had performed satisfactorily enough to be offered a third contract by the urban school district, to determine the relative predictive value of personal history data, interview assessment and recommendations assessment and any 97
PAGE 107
combination thereof, when the average of three principal's annual evaluation was used as the criterion measure. 2. For a sample of teachers in their first year of teaching within the urban school district and about whom no decision had been made relative to their continued status, to determine the relative predictive-value of personal history data, interview assessment and recommendations assessment and any combination thereof', when the principal's annual evaluation was used as the criterion measure. 3. To compare the two aforementioned samples of teachers in regard to their basic characteristics with a third sample of teachers consisting of those teachers who did not perform to a sufficiently satisfactory level to be retained under contract in the urban school district. 98 The study was confined to the three samples of teachers in the urban school district. The data for each individual teacher in each of the three samples were collected exclusively from the personnel records in the urban school district. The three samples did not include applicants who were considered but not employed in the urban school district.
PAGE 108
99 The external validity of the study was confine d to the single urban school district. The three samples included both elementary and s econdary teachers. Sample A consisted of 100 teachers, randomly chosen, who were hired in 1971-72 and who were in continuous employment in the urba n school district to January 1974. This sample represented teacher s who performed satisfactorily enough to be offered a thir d contract with the district. Sample B consisted o f 100 teachers, randomly chosen, who were initially employed in 1973-74 by the urban school district and who w e r e in continuous employment to January 1974. This sample constituted the sample about whom no decision had been made regarding their continued employment in the urban school district. Sample C, consisting of 21 arbitrarily chosen teachers, composed the group of teachers who had not performed at a sufficiently satisfactory level to be retained in the urban school district and were no longer employed at the time of the investigation. The teachers in this sample were employed by the district for varying lengths of time during the period from 1971-73. To determine the relative usefulness of personal history data, interview assessment and recommendations assessment when utilized in the selection of teachers for an urban school district, the following operations were performed. Frequency distributions were developed for Samples A and Bin regard to each of the nine variables to provide some idea of the basic characteristics of the data in terms of distribution and variability. Joint frequency distributions
PAGE 109
100 were developed to show how the dependent variable re-lated to each of the independen t variables. Two regression equations were computed using the eight independent variables to predict the dependent variable. The multiple regression analysis produced the highest possible correlation between the independent variables and the dependent. variable. Frequency distributions were developed to compare Samples A, B, and C in regard to each of the eight independent variables. Based upon the preceeding analysis, the following major findings emerged: 1. In Sample A, 59 percent of all teachers ~ere under 26 years of age and that age category received 55 percent of the most positive evaluation ratings. 2. In the most positive evaluation category for Sample B, graduates of universities in Florida received 51.4 percent of the ratings while graduates of universities outside of Florida received 48.6 percent. 3. Teachers with no previous teaching experience represented 62 percent of Sample A. Fortyfive percent of this experience category received the most positive evaluation ratings. 4. Teachers with a grade point average of 2.2-2.5 represented 22.0 percent of Sample A. This
PAGE 110
category received 26.7 percent of the most positive evaluation ratings. 5. In Sample B, 57.0 p ercent of all teachers were married. This category received 62.5 percent of the most positive evaluation ratings. 6. In Sample A, 22.0 percent of all teachers were black. This category received 25.5 percent of the most positive evaluation ratings 101 7. In Sample A, 37.5 percent of those teachers who received the most positive reco~~endations assessment received the most positive evaluation rating. Of those teachers who received the second most positive recommendations assessment, 31.6 percent received the most positive evaluation rating. The entire sample considered, 19.8 percent of all teachers in Sample A received the most positive evaluation rating. 8. The value of multiple R, the intercorrelation of the independent variables as well as the correlation between the independent variable and dependent variable for Sample A was .25153. The value of multiple R for Sample B was .31898. 9. R-square, the proportion of the variance in the dependent variable accounted for by the regression equation for Sample A was .05834. R-square for Sample B was .10175.
PAGE 111
102 10. In regard to the comparison of the distribution of personal histor y data for Samples A, B, and C, it was found that the distribution on the variables age, graduate of a university in Florida, marriage status, years of previous experience and grade point average was similar. The most obvious difference in the distribu-tion of personal history data for Samples A, B, and C was the variable race. Sample A had 22.0 percent black teachers, Sample B had 14.0 percent black teachers and Sample Chad 57.1 percent black teachers when the variable race was compared. 11. In the least positive interview assessment cate gory, Sample A had 10.5 percent, Sample B had 1.0 percent and Sample Chad 29.4 percent when the three samples were compared. 12. When the three samples were compared on recom-. mendations assessment, Sample A had 22. 0 percent in the least positive category, while Sample B had 27.2 percent and Sample Chad 68.8 percent in that category. 13. When the chi-square test of independence was performed on the variables age, gradua~e of a university in Florida, years of previous experience and marriage status, it was found that chi-squares were sufficiently low that the null hypothesis was accepted. That is, at the .01 level of
PAGE 112
significance, Samples A, B, and C came from the same population. The computed chi-squares for the variables race, inuerveiw assessment and recommendations assessment were large enough to reject the null hypothesis that the samples came from the same population. Conclusions and Implications 103 From the investigation the following conclusions seem justified. 1. The relative usefulness of personal history
PAGE 113
for less than 11 percent of the variance in the dependent variable. The results of the frequency distributions were equally inconclusive in relating the independent variables to the dependent variable. 2. The findings of the present investigation are consistent with findings of other studie s conducted under :similar circumstances. 3. Even though peripheral to the major intent of the present investigation, based on the data obtained when Samples A, B, and C, were compared, it seems justifiable to conclude that if persons are selected for teaching positions when these individuals have poor interview and recommendations ratings, then the likelihood of their success is quite limited. 104 In regard to implication for practice in the urban school district, it is suggested that more sophisticated methods of collecting data on applicants should be initiated based upon specific studies described in the literature. The criteria for success on the job should include a limited list of behaviors. These behaviors should relate to the purpose of the situation, should be measurable and h 1 d 1 d b 1 11 b 1 3 s on~ inc._u e ver a as we as non-ver a cornrnun1cat1ons. 3Dale L. Bolton, Selection and Evaluation of Teachers, {Berkeley: Mccutchan Publishing Co rpor~tion, 1973), p. 58.
PAGE 114
105 Extensive training and the rating of the interviewers should be instituted along with a determination of the type of interview that would be most effective. Further more, it is suggested that the urban school district conduct research in the areas of criterion measures, interviews, letters of recommendation and personal history data. This study should provide the basis for that research.
PAGE 115
SELECTED BIBLIOGRAPHY I. BOOKS Bolton, D ale L. Selection and Evaluation of Teachers. Berkeley: Mccutchan Publishing Corporation, 1973, Chruden, Herbert J., and Sherman, Arthur W., Junior. Readings in Personnel Management. Cincinnati: SouthWestern Publishing Company, 1961. Fear, Richard A. The Evaluation Interview. New York: McGraw-Hill Book Company, 1958. Guilford, J.P. Fundamental Statistics in Psychology and Education. New York: McGraw-Hill Book Company, 1956. Mandell, Milton M. The Employment Interview. Research Study number 47, New York: American Management Association, 1961. Mitchell, William C. "Multiple-Regression Analysis: Subprogram Regression." In Statistical Package for the Social Sciences. Edited by Norman H. Nie, Dale H. Bent and C. Hadlai Hull. New York: McGraw-Hill Book Company, 1970. Stone, Harold C., and Kendall, William E. Effective Personnel Selection Procedures. Englewood Cliffs: PrenticeHall, 1956. II. ARTICLES IN JOURNALS Adair, Charles D. "Relationship Between Undergraduate Grades and First Year Teaching Success." School and Communi t y 60: 22 (January 1974). Bass, Bernard M. "The Leaderless Group Discussion." Psychological Bulletin 51: 465-92 (September 1954). Bolster, B. I., and Springbett, B. M. "The Reaction of Interviewers to Favorable and Unfavorable Information." Journal of Applied Psychology 45: 98-103 (February 1961). Crissy, W. J. E. "The Employment Interview-Research Areas, Methods and Results." Personnel Psychology 5: 73-85 (Summer 1952). 106
PAGE 116
107 DeNelsky, Garland Y., and McKee, Michael B. "Prediction of Job Performance from Assessment Reports: Use of a Modified Q-Sort Technique to Expand Predictor and Criterion Variants." Journal of Applied Psychology 53: 439-45 (December 1969). Dunnette, M. D. "A Modified Model for Test Validation and Selection Research." Journal of Applied Psychology 47: 317-23 (October 1963). Fitzgerald, Paul. "Recruitment of Teachers -A Need for Reevaluation." Personnel Journal 49: 312-14 (April 1970). Guetzkow, Harold, and Forehand, Garlie A. "A Research Strategy for Partial Knowledge Useful in the Selection of Executives." In Research Needs in Executive Selection. Edited by Ranato Tagiuri. Boston: Harvard University, Graduate School of Business Education, Division of Research, 1961. Guion, Robert M., and Battier, Richard. "Validity of Personality Measures in Personnel Selection." Personnel Psychology 18: 135-64 (Summer 1965). Hollmann, Thomas D. "Employment Interviewers' Error in Processing Positive and Negative Information." Journal of Applied Psychology 56: 130-34 (April 1972). Langdale, John A., and Weitz, Joseph. "Estimating the Influence of Job Information on Interviewer Agreement." Journal of Applied Psychology 57: 23-27 (February 1973). Moriwaki, E. "Note on the Comparative Validities of Judgements of Intelligence Based on Photographs and on Interviews." Journal of Applied Psychology 13: 630-631 (November 1947). Peres, Sherwood H., and Garcia, J. Robert. "Validity and Dimensions of Descriptive Adjectives Used in Reference Letters for Engineering Applicants." Personnel Psvchology 16: 279-86 (Autumn 1962). Smith, W. I., Albright, L. E., Glennen, J. R., and Owens, W. A. "The Prediction of Research Competence and Creativity from Personal History.~ Journal of Applied Psychology 45: 59-62 (February 1961). Snedden, D. "Measuring General Intelligence by Interview." Psychological Clinic 19: 131-35 (July 1930). Springbett, B. M. "Factors Affecting the Final Decision in the Employment Interview." Canadian Journal of Psychology 12: 13-22 (March 1958).
PAGE 117
108 Yonge, K. A. "The Value of the Interview: An Orientation and a Pilot Study." Journal of Applied Psychology 40: 25-31 (February 1956). III. OTHER U. s., Civil Service Commission. A Guide for Executive Selection. Personnel Methods Series No. 13. Washington D. C.: Government Printing Office, 1961.
PAGE 118
BIOGRAPHICAL SKETCH Jon Thompson was born in St. Louis, Missouri on May 24, 1936. He graduated from Webster Groves High School in Webster Groves, Missouri in 1954. After he received a Bachelor of Arts in economics from the University of Missouri, he worked for five years for Ralston Purina Company in St. Louis, Missouri and Cincinnati, Ohio. He began teaching sixth grade in Greenhills, Ohio while pursuing a Mas~er of Education Degree at the University of Cincinnati. After he received the degree of Master of Education, and had taught three years in the Greenhills School District, Mr. Thompson and his family moved to Jacksonville, Florida. He taught sixth grade and served as assistant headmaster at Southside Day School during the 1968-69 school year. The following year he was employed in the Duval County Schools as principal of Norwood Elementary School. In 1970, he was assigned to Lackawanna Elementary School as principal. In January, 1971, he accepted the position of Supervisor of Elementary Staffing in the Duval County Schools. Mr. Thompson remained in this position until September, 1972 when he entered the doctoral program at the University of Florida. Upon completion of the doctoral program in August, 1974, he plans to return to the Duval County School System as a school administrator. 109
PAGE 119
I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Education. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Education. I certify that opinion it conforms presentation and is fully as a dissertation for the fl. &,J e ~ -~n Crews Professor of Education
PAGE 120
This dissertation was submitted to the Graduate Faculty of the College of Education and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Education. August, 1974 llJG Wt~ f Education Dean, Graduate School
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EM5C0BI7C_64FY1I INGEST_TIME 2017-06-08T21:35:22Z PACKAGE AA00055802_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
|
|