Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Norms and validity of sixteen test variables for predicting success of foremen
(USC Thesis Other)
Norms and validity of sixteen test variables for predicting success of foremen
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
NORMS AND VALIDITY OF SIXTEEN TEST VARIABLES FOR PREDICTING SUCCESS OF FOREMEN A Thesis Presented to the Faculty of the Department of Psychology University of Southern California In Partial Fulfillment of the Requirements for the Degree Master of Arts hy Robert R^^ackie June 1948 UMI Number: EP63953 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. UMI EP63953 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 T/lis thesis, written by Robert R. Mackie / under the guidance of h ..X 3 .. Faculty Committee, and approved hy all its members, has been presented to and accepted by the Council on Graduate Study and Research in partial fu lfill ment of the requirements fo r the degree of Master of Arts Dean Faculty Committee Chairman TABLE OF CONTENTS CHAPTER PAGE I. THE PROBLEM................. 1 Statement of the problem ................... 2 Importance of the study.................... 3 II. REVIEW OF THE LITERATURE ...................... 4 Summary.................................... .. . 13 III. METHODS AND PROCEDURE..................... 1$ The subjects............................. 15 The criterion ......................... 16 The measuring devices .......... ....... 17 Statistical techniques ....................... 22 Summary..................................... 24 IV. RESULTS AND DISCUSSION..................... 25 Measures of dispersion and central tendency . . 25 Measures of correlation............ 33 Determining profiles........ .. • ......... 41 An interpretation of the validity obtained • . • 46 Summary..................................... 51 V. SUMMARY AND CONCLUSIONS........ 54 BIBLIOGRAPHY....................................... 58 .......................................... .. • ^3 LIST OF TABLES TABLE PAGE I. Means and Standard Deviations for Two Groups of Foremen in Sixteen Psychological Test Variables . .............................. * 26 II. Means, Standard Errors of Means, Differences Between Means, Standard Errors of the Differences Between Means, T-ratios, and Significance Where Obtained . . . . . . . 29 III. Combined Means, and Standard Deviations, Weighted in Terms of N, for 409 Foremen in Sixteen Test Variables 32 IV. Coarse and Corrected Pearson r for Sixteen Variables Measured in Two Groups Using Foremen Efficiency Ratings as a Criterion . . 35 V. Pearson r*s, Corresponding Eta’s, Chi Squares, Degrees of Freedom, and Significance of the Curvature where Found For Nine Test Variables 3 8 VI. Mean Scores of Foremen Rated ”5" and.”1” on Eleven Variables Which Correlated Significant ly with Criterion............ 42 VII. Proportion Who Will be Satisfactory Among Those Selected, for Given Values of the Proportion of Present Employees Considered Satisfactory, iv TABLE PAGE The Selection Ratio, and r. (after Taylor and Russell)............ . ... ........... 47 VIII. Proportion of Successful Foremen Which Would Be Selected by Several Tests Under Conditions of A yfo Selection R a t i o .......... . 48 IX. Percentage Increase of Successful Foremen Chosen Under Conditions of 5^ Selection Ratio Using the Several Significant Tests ... . . . . . 50 LIST OF FIGURES FIGURE PAGE I. Profiles for the Best and Poorest Foremen in Production Group ......................... 44 II. Profiles for the Best and Poorest Foremen in Maintenance and Tooling G r o u p ......... 45 CHAPTER I THE PROBLEM It has been the policy of many personnel men in the past to assume that the person best suited to fill the fore man’s position is the departmental worker who has had the most satisfactory work record in terms of experience and time on the job. Innumerable experiences have taught management that selection of foremen on this basis often leads to dis appointing results. According to Moore^, "The modern fore man is not necessarily a pace-setter, neither is he the most skilled and highest quantity or quality producer in the unit. His success as a foreman depends on other factors than job knov/ledge or job efficiency." It is true, of course, that the foreman should know the jobs that he supervises and should be able to assist operators v/ho need help. But in recent years management has realized more and more that today’s successful foreman must be equipped with certain other assets. Just Virhat these are, and just what degree of their presence or absence can be discovered has become a task of the psychologist. Toops has recently stated the essence of the ^ Herbert Moore, "Supervision I - Selection," Personnel Journal. 20, 1940, #10, 353-357. 2 problem . Job profiles should not be in terms of job character istics but in terras of the human traits needed for that job. The . job is created for the man. Thus occupations should be defined in terms of ideal sets of human, traits, and from empirical observations the human profile which best succeeds at the occupation should be determined. In other words, ideally one should be able to group the successful foremen according to certain basic psychologi cal dimensions. These dimensions should be isolated by factor methods and incorporated into suitable objective measuring devices. Then, applying such measuring devices to a group known to vary in the desired behavior, one should be able to produce ideal profiles in terms of those dimensions. Statement of the problem. It was the purpose of this investigation to make a comparative study of efficiency ratings of 409 foremen in a large manufacturing plant and scores in sixteen psychological dimensions as measured by the Guilford-Martin personality inventories, the Bennett Test of Mechanical Comprehension, the Minnesota Paper Form Board Test, and the Wonderlic Personnel Test. Norms were established for each dimension, measures of variability applied, and an effort made to establish a profile for the successful supervisor in these plants. Finally, the degree of correlation between ratings and the dimensions in question was ascertained. ^ H. A. Toops, "Some Concepts in Job Families and Their Importance in Placement," Educational and Psychological Measurement, 5, 1945, 195-216. 3 Importance of the study. Selection of a good foreman, in contrast to the selection of a poor one, means not the mere addition of one man’s greater efficiency to the working force, but rather a cumulative addition in terms of the greater morale which he inspires throughout his span of super vision. Enlightened managements are well aware of the extreme costs, even among regular workers, of excessive turnover due to faulty selection and/or placement techniques. Because of his strategic position, the personal relationships betvveen the foreman and his workers directly influence the efficiency of production throughout an entire unit. Turnover at the foreman level, or lack of turnover after faulty selection, makes labor costs progress geometrically. The value of determining the successful foreman’s job profile, then, together with the discovery of economical and objective measuring devices for that profile, cannot be over estimated. In general personnel selection in industry has taken great strides through psychological testing. To this date testing at the supervisory level has proved relatively ineffectual. It is hoped that some progress in this area may be made through the results of the present study. CHAPTER II I. REVIEW OF THE LITERATURE The psychological literature is particularly barren of studies on the prediction of supervisory ability through the application of paper-and-pencil tests. It is very likely that the results of many such studies have been kept confi dential by the organizations involved in view of the consider able competition on the job market for such leaders. It is also true that selection through testing, especially at the supervisory level, is yet at the experimental stage and there are no prominent landmarks in this area from which one may take special direction. A general suggestion of the problem is presented by Cowley’s study of 1928^. He measured leadership traits in three group situations: army, penitentiary, and college. Leaders were chosen from each group on the basis of case histories and ratings by the members in each group. Against this criterion he compared the results of twenty-eight psycho logical tests purporting to measure aggressiveness, self- confidence, intelligence, emotional stability, finality of judgment, tact, and suggestibility. Cowley’s conclusions 3 w. H. Covyrley, "Three Distinctions in the Study of Leaders," Journal of Abnormal and Social Psychology. 23: 114-157, 1 9 ^ follow: First, that leaders possess different traits from their followers, and second, that leaders in these dif ferent situations do not possess even a single trait in common. This amounts briefly to a demonstration of the fact that leadership is a function of a definite situa tion and that we- cannot talk about, leadership traits in general, but that instead we must talk about leadership traits in particular situations. A report by Jackey^ gives us a general suggestion of the essential traits we should attepipt to measure in estab lishing a testing program at the supervisory level. He lists the following as basic qualifications of the successful supervisor: 1. He must command the respect of his men. 2. He must have good health. 3. He must make a good appearance. 4. He should have knowledge, training, and experience in the work he is supervising. 5. He should have initiative. 6. He should be progressive, and should develop a certain amount of temperate aggressiveness. 7. He should be able to make prompt decisions. 8. He should have the kind of mind that can organize the work of his department in a logical and effi cient manner. 9. He should have the kind of mind that would be able to solve difficulties and problems by the proper treatment of facts. 10. He should be a good judge of human nature. 4 David E. Jackey, "Supervision II-Training," Personnel Journal. 20, #10, 357, 1942. 6 It is interesting to note that several of these requisites probably are correlates of the temperament traits investigated in this study. One of the earliest and more specific investigations of this problem was reported by Uhrbrock and Richardson in 1934^. In this study nine test forms, containing 820 items, were administered to 163 factory supervisors. The tests included were: (1) Modified Army Alpha (Intelligence) (2) Strong’s Vocational Interest Blank (3) McCall’s Multi-Mental Test, Form A (4-5) Minnesota Paper Form Board Test, Series A and B (6) Selected items from Thurstone’s Personality Sched ule (7-9) Three Company Information Tests In addition to the test results, fifteen items of personal history data were obtained for each supervisor. An item analysis was then conducted on the total 835 items in rela tion to efficiency ratings. Each supervisor was rated by four superintendents, the reliability of the ratings being .80 ^ .02. The authors report that only 85 of the 820 psychologi cal and interest items proved to be significant under the 5 R. S. Uhrbrock and M. W. Richardson, "The Basis for Constructing a Test for Forecasting Supervisory Ability." Personnel Journal, 12: 141-154, 1934* 7 item analysis. In addition four items of personal history were significant. Whereas the validity coefficient of the entire battery (in terms of ratings) had been only .49 ^ .04, use of the 85 significant items raised this coefficient to .71 - *03. Uhrbrock and Richardson concluded on this point that, ... the presence of approximately 90^ dead wood among the items analyzed suggests a possible explanation for the failure of certain industrial testing programs during the past decade. The available tests, with the exception of those especially constructed to measure company information, were prepared in university laboratories and represent collections of items designed to sample a wide range of experience and ability. For a given occupation, in a specific company, any test now on the market undoubtedly will contain a great amount of dead wood. Other interesting findings of this investigation in cluded the insignificant correlation of age with the criterion (-.18 - .06); a significant difference in the probability of successful supervisors with nine or more years of schooling as contrasted with those of eight or less (the advantage of college training over high school was inconsiderable); and the fact that the company information material showed a greater proportion of valid items than did the purchased tests. In 1940 Harrell^ applied several standard tests to cotton mill supervisors in five Georgia textile plants. In 6 Willard Harrell, "Testing Cotton Mill Supervisors," Journal of Applied Psychology, 1940, 24, 31-35* 8 this instance the criterion consisted of ratings 'by one to four qualified superiors. Only intelligence tests showed any appreciable relation to supervisory ability, as seen from the following table: (1) Bernreuter Personality Inventory (Self Confidence). .15 (Sociability) 09 (2) Otis S-A Intelligence Test (higher form)........37 (3) Moss, Hunt, and Omevake’s Test of Social Intelligence ............................. 18 The Strong Vocational Interest Blank indicated more interest in people and business than in science or language. The cor relation of ratings with age was -.03; with school grade, -.03* In 1944 Goodwin^ compared results of leadership ratings among military personnel with scores on the California Test of Personality. The coefficient of correlation between leader ship ability, and adjustment was -.02. Using the "group response" to each item as the scoring key the correlation was increased to .27 - .05. The men judged highest'on leadership ability showed the highest correlation with "group response" scores, indicating that the real leader does not deviate sig nificantly from his group in response to the items on this test. Only eleven of the 180 items proved capable of 7 Phillip Goodwin, "An Introduction Study of the Pre diction of Personal Leadership Ability Based on Group Atti tudes and Opinions." (Unpublished Master’s Thesis, University of Southern California Library, 1944.) 9 eliciting significantly different responses betv/een leaders and non-leaders. Shuman had some success in improving selection of foremen in aircraft engine and propeller industries . Two- hundred ninety-seven foremen, assistant foremen, set-up men, and group leaders were tested in small groups. These men were rated "poor-,7 "average," or "excellent." The following tests were used; (1) Otis Q,.S. Test of Mental Ability, Beta A (2) Minnesota Paper Form Board Test. (3) Bennett Test of Mechanical Comprehension Critical scores in each test were found which combined to give a 15% to 20% improvement in the selection of "excellent" supervisors. VogeL^ reports an attempt to measure the personality factor among supervisors at the Lockheed Aircraft Corporation in 1945. The testing division devised its own test in an effort to circumvent the weaknesses of the standardized tests available. Three hundred eighty-five supervisors were divided into standardization and validation groups. They were rated as good or bad supervisors on the basis of ratings of super visory, technical, and administrative knowledge. Only 20 of ^ J. T. Shuman, "The Value of Aptitude Tests for Super visory Workers in the Aircraft Engine and Propeller Industries," Journal of Applied Psychology, 1945, 29, I85-I9O. 9 E. M. Vogel, "An Attempt to Measure the Industrial Supervisor’s Ability by Means of a Self-Inventory Test. Unpub lished Master’s Thesis, University of Southern California, 1945.) 10 125 items on the test were significant in the standardization group. These items in turn failed to discriminate within the validation group. Vogel concluded that supervisor selection through testing personality traits is not a fruitful approach. Sartain recently made a correlational study at North American Aviation between ratings and several standard tests^^. Forty supervisors were given the tests listed below. Two ratings on each of two rating forms were secured for each man, and the sum of the four ratings, reduced to standard scores, became the criterion. The resulting correlations follow: (1) Otis S-A Test of Mental Ability (Higher Exam). .04 (2) Tiffin and Lawshe Adaptability Test (Form A)..-.07 (3) Revised Minnesota Paper Form Board Test......... 10 (4) Bennett Test of Mechanical Comprehension (Form AA).......................................-.15 (5) Remers and File How Supervise? (Form A)....... -.18 (6) Bernreuter Personality Inventory (Neurotic Tendency)...........11 ( Dominanoe-Submission)........12 (Self-Confidence)............. 01 (Sociability)................. 07 (7) Kuder Vocational Preference Record (Mechanical)................ .00 (Social Service)............ -.06 (Clerical)...... GO Differences between means of scores, although not statistically significant, suggested that the good supervisors A. Q,. Sartain, "Relation Between Scores on Certain Standard Tests and Supervisory Success in an Aircraft Factory." Journal of Applied Psychology. 30, 1946, 328-332. 11 were more stable, more dominant, more self-confident, and more sociable than the poor supervisors. Sartain concluded, "These low correlations may be due to a faulty criterion. It seems more probable, however, that the tests simply fail to correlate with supervisory success in the plant." During the war years, Stoclcford^^ made a study of the supervisory personnel in the Lockheed Aircraft Corporation. Items of personal information, and measures of personality, interest, and intelligence were obtained from 478 supervisors. The criterion of success was the rankings of these supervisors according to administrative ability, supervisory ability, and technical ability. This was done by immediate superiors. Four divisions of supervisors resulted--the "superior group" (26%), the "good group"(26%), the "fair group"(26%), and the "poor group"(22%). The items of personal information which were the best indicators of success proved to be age, marital status, number of dependents, residence in the community, training record, and previous experience. A personality test revealed no re lationship between individual test scores and job success. However, an employee adjustment scale revealed that rated personality "predicts success on the job 40 percent greater 11 Lee Stockford, "Selection of Supervisory Personnel." Personnel, 24, 186-199, 1947. 12 than chance would allow." Ratings were made hy men who knevf the supervisors, but v/ho had no knowledge of their success rankings. Intelligence as measured by the California Test of Mental Maturity proved to be "one of the most important factors in discriminating between the better and poorer men." Tests of vocational interest and mechanical aptitude were not correlated because of lack of time. The most significant items were later administered to a new group of supervisors which served as a validation group. Profile scores representing an average of the scores on the several items were assigned to each man and compared with his rating (in this case upper half versus lower half). Part of the results are indicated in the table belov/. A profile score of zero is most desireable, one of eight least desireable. Profile Evaluation Percent Given Above Score Average Ratings 0 70^ 1 72% 2 51% 3 37% 4 22% 5 12% 6-7-8 0% la 1944 Riohardson and Hanawalt^^ compared scores on the Bernreuter Personality Inventory of 258 business men divided into groups of office-holders and supervisors versus non-office-holders and non-supervisors. Office-holders and 12 H. M. Richardson and N. G. Hanawalt, "Leadership as Related to the Bernreuter Personality Measures." Iournal of Applied Psychology, 28, 308-17, 1944* 13 supervisors tended to be less neurotic, less introverted, more dominant, more self-confident and more self-sufficient than the control group. The authors reported statistically significant differences for the majority of the measures. Jenkins^^, after a recent review of many leadership studies of various sorts, concluded that no single trait or group of characteristics has yet been isolâteddwhich sets off the leader from the members of his group. He offered the following as the most likely hypotheses as a result of his survey; (1) Leadership is specific to the particular situa tion under investigation... (2) In practically every study reviewed leaders showed some superiority over members of their group in at least one of a wide variety of abilities. The only common factor appeared to be that leaders in a particular field need and tend to possess superior general or tech nical competence or knowledge in that area..... (3)^ Leaders tend to exhibit certain characteristics in common with the members of their group. Two of the more obvious of these characteristics are interest and social background. (4) there is a widespread but vague hint that certain poorly defined personality traits charac terize individuals holding positions of responsibility... II. SUMMARY A review of the pertinent literature suggested that 13 William D. Jenkins, "A Review of Leadership Studies with Particular Reference to Military Problems," Psychological Bulletin, 44, 54-79, 1947. 14 leadership traits are peculiar to specific situations; that thus far, no general pattern of the leader has been isolated. The leader seems to maintain certain traits in common with his group, and to excell the group in others. For this reason it is likely that "tailoring" of available standardized tests to the specific test situation by item analysis or some simi lar technique would be useful. There seemed to be general agreement that leadership traits for foremen and supervisors are to be found in person ality and temperament areas. This line of attack has been moderately fruitful in some studies, not at all in others* There was some evidence that age is insignificantly correlated with foreman success, and that education beyond the high school level is of no advantage. CHAPTER III METHODS AND PROCEDURE It was the purpose of this investigation to determine what relationships hold between a criterion of foreman success and scores in several psychological tests that may yield meas urements of important traits in the profile of the ideal foreman. I. THE SUBJECTS The subjects were 409 foremen in two plants of a large eastern manufacturing concern^^. Of these, 289 were in charge of production groups, 70 in charge of maintenance groups, and the remaining 50 in charge of tooling groups. This suggested a natural grouping, but in order to enlarge samples and thus to make comparisons more significant the latter two groups were combined to make a total of 120. Rough breakdowns according to age, education, and job level also were considered, but again the resulting groups v/ould have been so decreased numerically that comparisons vrould have been much reduced in significance. The effects of age and education were determined later by correlating them sep arately with the criterion. Since 74 percent of the foremen were at one job level, and the remaining 26 percent divided between two higher levels, it was decided that a breakdown 14 Name withheld by request 16 in-this variable would have decreased the meaningfulness of the results. Consequently the entire group was treated as if job level were constant. According to a .personal letter received from the per sonnel director of the organization, the foremen had been called in for testing in small groups, during the period from May to August 1946. A conference room wns used for testing at each plant. The rooms were air conditioned, well lighted artificially, and free from any outside noises. The tests were taken in two different sittings: the Wonderlic Personnel Test, Bennett Test of Mechanical Comprehension, Minnesota Paper Form Board, and Guilford-Martin Personnel Inventory (0,Co,Ag) during the first sitting, and the Kuder^^, Guilford’s Inventory of Factors STDCR, and the Guilford-Martin Inventory of Factors GAMIN during the second. The men were informed that the test results would have no effect on their job status. Rather it was explained that the standards set by them might be used for the selection of others for similar positions. The tests were given on com pany time, the top supervision in each plant being among the first tested. There was practically no opposition to the testing program. II. THE CRITERION Three independent ratings were made of each foreman by 15 Results obtained on Kuder Preference Record were not reported in this study. 17 three of his immediate superiors. Each rater was asked to group his total ratings into an approximately normal distri bution (10%, 20%, 40%, 20%, 10%.) Ratings assigned were 1-2- 3-4-5, 5 being the most desirable. The final rating for any one foreman was taken as the average of his three ratings. A fraction of tvvo-thirds was counted as the next higher rating, and one of one-third as the next lower. In the majority of cases there was close agreement on the poorest and the best men. The final criterion is not a result of the first ratings obtained, but rather of a second group obtained several months after the first. No great differences være obtained; however, no coefficient of reliability was available at the source of the data. III. THE MEASURING DEVICES A. The Guilford-Martin inventories consist of three pencil-and-paper tests which, when combined, yield indica tions of temperament in thirteen different areas as measured by a total of 511 items. The Inventory of Factors STDCR^^ was developed as a result of the attempts to measure intro version and extroversion as defined by Jung. The traits may be thought of as representing bi-polar personality dimensions, and may be defined as follows: J. P. Guilford, ^ Inventory of Factors STDCR, Manual of Direction and Norms (Revised) (Beverly Hills, California: Sheridan Supply Company). 18 S-Sacial introversion-extroversion. Shyness, seolusive- ness, tendency to withdraw from social contacts, versus socia bility, tendency to seek others ‘and enjoy the company of others. T-Thinking introversion-extroversion. An inclination to meditative or reflective thinking, philosophizing, analysis of one’s self and others, versus an extrovertive orientation of thinking. D-Depression. Habitually gloomy, pessimistic mood, with feelings of guilt and unworthiness, versus cheerfulness and optimism. C-Cycloid disposition. Strong emotional fluctuations, tendencies towards flightiness and emotional instability, versus uniformity and stability of moods, evenness of dis position. R-Rhathymia. An inhibited, over-controlled, conscien tious, serious-minded disposition, versus a happy-go-lucky, carefree disposition, liveliness, impulsiveness. The Inventory of Factors GAMIN^'^ was developed partly as a result of the Freeman theory that individual differences in the reactivity of the nervous system furnish the psycho logical basis for important differences in personality. The factors are defined as follows: 17 j. p. Guilford and H. G. Martin, Inventory of Factors GAMIN, Manual of Directions and Norms (First RevisionT" (Beverly Hills, California: Sheridan Supply Company, 1943). 19 G-General pressure for overt activity, a tendency to engage in vigorous action. A-As0endanoe-8ubmission. Ascendancy in social situa tions as opposed to submissiveness. Leadership qualities. M-Masculinity-Femininity. Women typically score belov^r median and men above the median in this trait. I-Inferiority feelings versus self-confidence. N-Nervousness. Tendency to be jumpy, Jittery, easily distracted, irritated and annoyed, versus tendency to be calm and relaxed. The Guilford-Martin Personnel Inventory^^ was developed to measure the paranoid area of temperament on the thesis that many of the troublemakers in business and industry had tend encies in this direction. The traits being measured here are suppiciousness, fault-finding, belligerence, and personal reference. These are incorporated into the following scales: Co-Cooperativeness. This is opposed to fault-finding and over-criticalness. Ag-Agreeableness. This is opposed to belligerence or a domineering disposition. O-Objectivity. This is opposed to personal reference, a tend ency to take things personally and subjectively and to be 1^^. pT Guilford and H. G. Martin, The Guilford-Martin Personnel Inventory, Manual of Directions and Norms [Beverly Hills, California: Sheridan Supply Company, 19A3)• 20 hypersensitive. Substantial reliabilities have been demonstrated for all sub-tests. Alternate sixths and use of the Spearman- Brown formula gave estimated reliabilities as follows: 3——,92, T——.89» D——.91) C——.91» R——.89* For the other two inventories, split-half reliabilities were as follows: G——.80, A——.88, M——.82, I——.89> N——.90, 0——.83» Ag——.80, and Co——.91* B. The Wonderlic Personnel Test^^ is a,revised form of the Otis S-A Test of Mental Ability (higher form). It is comprised of fiftjr items and may be taken with a twelve minute time limit. The items include analogies, analysis of geometric forms, arithmetic problems, disarranged sentences, sentence parallelism with proverbs, similarities, logic, defi nitions, judgment, direction following, and spatial relations. The test w r s designed for testing adults in business and industrial situations. Wonderlic suggests that it should be useful as a selection instrument in hiring and placing. Norms are established for each situation to which it is applied. Reliabilities have been calculated on several popula tions. By alternate forms these range from ,82 to .94, and, by the odd-even method, from .88 to .94 for the twelve minute 19 E. F. Wonderlic, Wonderlic Personnel. Test Manual (750 Grove Street, Glencoe, Illinois, 1945)• 21 form of the test. 0 (\ C. The Bennett Test of Mechanical Comprehension is designed to measure the capacity of an individual to under stand various types of physical and mechanical relationships. Most problems are presented so as to require understanding of principles rather than measure one* s acquaintance with mechanical devices. There are sixty pictorial items. In this study the more difficult form (BB) was used. There is no time limit. Immediate test-retest reliabilities range from .90 to .93. D. The Minnesota Paper Form Board Test^^ is a modifi cation of the paper form board test prepared by Paterson, Elliot, Anderson, Toops, and Heidbreder at the University of Minnesota in 1930. The original test was found by its authors to have a relatively high validity for the measurement and prediction of mechanical ability, and to differentiate between various levels of ability. The revised form is essentially the same with the exception that it is easier to score and has simplified directions, equivalent alternate series, and other improvements. 20 George K. Bennett and Dinah E. Fry, Test of Mechanical Comprehension, Manual of Directions [New York: Psychological Corporation, 1941)• R. Likert and W. H. Quasha, Revised Minnesota Paper Form Board Test, Manual of Directions (New York: Psychological Corporation, 1941TI 22 There are sixty-four problems of a visual-mechanical nature. The time limit is twenty minutes. The reliability as predicted by the Spearman-Brown formula is .85 for a single series end .92 for both series (AA and BB). The original paper form board test correlated .61 with mechanical ability according to the authors; the present revision cor relates .94 with the original test. 17. STATISTICAL TECHNIQUES The first step in working the data was to plot scatter diagrams of ratings against scores for each of the sixteen test variables. In this and subsequent operations production foremen were treated as one group, maintenance and tooling foremen as another. In addition, diagrams of ratings against age and ratings against education were plotted. Also a second set of diagrams for the Wonderlic test w^as plotted, this one op with a correction for age which that author recommends Diagrams of variables having significant correlations with the criterion will be found in the appendix. The mean score was then computed for each test variable, and the standard deviation was computed to ascertain the vari ability in each trait. These norms were compared with other reported norms for these tests. The standard error of the means were computed next. 22 Wonderlic, op. cit. 23 T-ratios were then determined and the significance of the differences between the means ascertained. Means for the two groups were averaged where this difference was insignifi cant. The next procedure was to determine the intercorrela tions betw'een ratings and tèst scores for each group in each of the sixteen variables. Correlations were also-determined for age and rating and for education and rating. Pearson r’s were computed first. These were then corrected for coarse grouping in the Y variable (ratings). Where the scatter diagrams suggested possible curvilinear relationships, the correlation ratio was also computed. The Chi Square test was then applied to the difference between the Pearson and eta coefficients to determine whether there was a significant departure from linearity. In an effort to represent graphically the difference in trait profiles between the best and the poorest foremen, test scores in those variables showing significant correla tion with the criterion were C-scaled. Mean scores for the best and poorest foremen, rated ”5” and ”1” respectively, were then computed and plotted against appropriate C-scores in each variable. This resulted in distinct profiles for the extremes of the rating scale. Finally an interpretation of the obtained coefficients in terms of the predictability they insure was made through 24 a technique developed by Taylor and Russell^^. V. SUMARY This chapter has presented the methods and procedure involved in the study: 1. Four hundred and nine foremen were divided into two groups: 289 in production and 120 in maintenance and tooling. 2. Each foreman was rated by three superiors, the average rating serving as the criterion. 3. The following tests were administered to each foreman: the Guilford-Martin Personality Inventory; the Wonderlic Personnel Test; the Bennett Test of Mechanical Comprehension; and the Revised Minnesota Paper Form Board Test. 4* Statistical techniques were selected with the pur pose of establishing norms, determining correlations betvveen test scores and the criterion, and establishing the validity of the tests. ^3 H. C. Taylor and J. T. Russell, "The Relationship of Validity Coefficients to the Practical Effectiveness of Tests in Selection,” Journal of Applied Psychology. 123: 565- 578, 1939. CHAPTER IV RESULTS AND DISCUSSION In this chapter the results of the data which were obtained by applying various statistical procedures are pre sented and.discussed. I. MEASURES OF DISPERSION AND CENTRAL TENDENCY In order to obtain a set of norms which might be used as a base for the future selection of foremen or for compari son of past and future studies of a similar nature, means and standard deviations were computed for each variable with in both the production and the maintenance and tooling groups. The standard deviations were obtained by the formula <T = i T h e s e measures of dispersion and central tendency are presented in Table I. It was considered desirable to-determine the reliabil ity of the means obtained, and to test the differences between the means of the two groups for significant differences. A combined mean and combined standard deviation, weighted in terms of N, would be justified providing the differences between the means were insignificant. 24 j. p. Guilford, Fundamental Statistics in Psychology and Education (New York: McGraw-Hill Book Ôompany, Inc., 1942), p. 54. 26 TABLE I MEANS AND STANDARD DEVIATIONS FOR TWO GROUPS OF FOREMEN . IN SIXTEEN PSYCHOLOGICAL TEST VARIABLES* PLANT TEST VARIABLE MEAN STANDARD DEVIATION P S 16,01 8.17 M & T S 17.80 9.22 P T 34.19 8.40 M & T T 34.52 8.44 P D 13.26 9.38 M & T D 13.08 9.19 P C 18.44 10.94 M & T C 17.25 10.44 P R 40.56 9.81 M & T R 38.95 10.01 P G 12.21 4.30 M & T G 11.66 4.12 P A 23.29 5.66 M & T A 22.13 5.78 P M 22.38 4.40 M & T M 22.08 4.02 P I 37.68 6.78 M & T I 37.62 7.47 P N 30.77 7.71 M & T N 31.70 6.66 P 0 50.24 11.69 M & T 0 51.00 12.12 P Ag 36.09 8.08 M & T Ag 38.35 8.45 P Co 60.55 16.58 M & T Co 62.25 16.79 Minnesota Paper Form Board Test. TABLE I (continued) MEANS AND STANDARD DEVIATIONS FOR TWO GROUPS OF FOREMEN IN SIXTEEN PSYCHOLOGICAL TEST VARIABLES PLANT TEST VARIABLE MEAN STANDARD D: P PFB 32.83 9.80 M & T PFB 33.10 9.40 P BMC 27.78 10.57 M & T BMC 30.00 9.90 P WONDERLIC 18.26 7.33 M & T WONDERLIC 15.92 7.92 28 The procedure here was to determine first the standard error of each mean by the formula (Tm = The stand ard error of the difference between means could then be com puted by the formula dd “ V Dividing the actual difference between means by the standard error of the difference between means produced the familiar t-ratio for a test of significance. This problem demanded a t-ratio of 1.966 for confidence at the five percent level and 2.588 for confidence at the one percent level. That is, only with t- ratios of such magnitude could one feel confident that the differences obtained did not arise by chance, the odds being nineteen to one and ninety-nine to one respectively. The means, standard errors of the means, differences between means, standard errors of the differences between means, t-ratios, and significance (where the standards were met) appear in Table II. An examination of Table II indicated significant dif ferences which favor the maintenance and tooling group on the Bennett Mechanical Comprehension and on Guilford*s- Ag (agreeableness), and a difference in favor of the production group on the Wonderlic Personnel Test. Means and standard deviations for all other test 25 Ibid., p. 128. 26 Ibid.. p. 135. 29 TABLE II MEANS, STANDARD ERRORS OF MEANS, DIFFERENCES BETWEEN MEANS, STANDARD ERRORS OF THE DIFFERENCES BET^VEEN MEANS, T-RATIOS, AND SIGNIFICANCE VfilERE OBTAINED PLANT VARIABLE MEAN STANDARD DIFFERENCE STANDARD T-RATIO SIGNIF- ERROR OF BETWEEN ERROR OF ICANCE MEAN MEANS DIFFERENCE p S 16.01 .481 1.79 .972 1.842 «W «0 ^ M & T S 17.80 .845 P T 34.19 .495 .34 .919 .366 M & T T 34.52 .774 P D 13.26 .552 .19 1.008 .185 M & T D 13.08 .843 P C 18.44 .645 1.185 1.155 1.026 M & T C 17.25 .957 P R 40.56 .346 1.615 .980 1.648 mm m m «0 M & T R 38.95 .918 P G 12.21 .254 .55 .455 1.220 M & T G 11.66 .377 P A 23.29 .328 1.15 .623 1.851 «■» mm ••m - mm M & T A 22.13 .530 P M 22.38 .259 .30 .452 .662 mm mm m m m m M 8 c T M 22.08 .369 P I 37.68 .400 .05 .793 .064 0 » mm m m m m M & T I 37.62 .685 P N 30.77 .454 .93 .811 1.114 mm mm m m < m m M & T N 31.70 .610 P 0 50.24 .689 .76 1.307 .581 mm mm m m mm M & T 0 51.00 1.111 30 TABLE II (continued) MEANS, STANDARD ERRORS OF MEANS, DIFFERENCES BBTV/EEN MEANS, STANDARD ERRORS OF THE DIFFERENCES BETWEEN MEANS, T-RATIOS, AND SIGNIFICANCE WHERE OBTAINED PLANT VARIABLE MEAN STANDARD DIFFERENCE STANDARD T-RATIO SIGNIF- ERROR OF BETWEEN ERROR OF ICANCE MEAN MEANS DIFFERENCE p Ag 36.09 .476 2.264 .903 2.493 .05 M & T Ag 33.35 .774 P Co 60.55 .977 1.704 1.322 .934 M & T Co 62.25 1.533 P PFB 32.33 .573 .266 1.033 .256 M & T PFB 33.10 .362 P BMC 27.73 .623 2.233 1.101 2.023 .05 M & T BMC 30.00 .903 P WON*Lie 13.26 .432 2.34 .345 2.769 .01 M & T WON* Lie 15.92 .726 01 variables could then be averaged. These are shown, weighted in terms of N, in Table III along with separate means and standard deviations for the two groups for the variables yielding significant mean differences. Norms reported by others were then compared with those obtained in this study. No real meaningful comparison can be made between scores obtained here on the Guilford battery and other available norms. Scores in personality;- variables are in part a function of the motivation of those answering the questions, and any comparisons between different types of populations would be risky. The real significance of these scores could be ascertained by administering the tests to other groups within the same industrial situation and under very similar conditions. Comparison of norms between such groups might indicate distinct deviations for the foreman or the other groups. Such a procedure was not a part of this study but should be supplemental to it in an effort to obtain maximum meaning from the results which are to be presented in the remainder of this chapter. No special deviations were indicated on the Minnesota Paper Form Board test t o t this group of foremen as a whole. In comparison with seventy-six "unclassified” males aged twenty-six to sixty^^, the mean score for all foremen fell at 27 Likert and Quasha, 0£. olt. 32 TABLE III COMBINED MEANS AND STANDARD DEVIATIONS, WEIGHTED IN TERMS OF N, FOR 409 FOREMEN IN SIXTEEN TEST VARIABLES TEST VARIABLE MEAN STANDARD DEVIATION 8 16.53 8.48 T 34.29 8.41 D 13.21 9.32 C 18.09 10.82 R 40.09 9.87 G 12.05 4.25 A 22.95 5.70 M 22.29 4.19 I 37.66 6.98 N 31.04 7.40 0 50.46 11.82 Co 61.05 8.19 Ag (P) 36.09 8.08 Ag (M & T) 38.35 8.45 PFB 32.91 10.37 BMC (P) 27.77 10.57 BMC (M & T) 30.00 9.90 WONDERLIC (P) 18.26 . 7.33 WONDERLIC (M & T) 15.93 7.92 33 the 60th percentile for the ”unclassified” males. However, in the maintenance and tooling group the best foremen scored at the 85th percentile, and the poorest at the 38th percentile There was little differentiation in the production group. Bennett has published norms for light mechanical work ers which might be somewhat comparable to the ones obtained in this study^^. The average production group foreman scored at the 45th percentile established by light mechanical work ers, and the average maintenance and tooling group foreman scored at the 55th percentile. In the latter group the best foremen averaged at the 68th percentile, the poorest at the 25th percentile. Wonderlic^^ stresses the importance of establishing new norms for each test situation, but suggests a minimum score of twenty-five for foremen. Using scores corrected for age, only the best foremen in this study made an average score that high. Wonderlic does not report the source of his norms, so no conclusion is possible. II. MEASURES OF CORRELATION Using the same scatter diagrams, Pearson ' r*-s- were next computed to determine the degree of relationship between 28 Bennett and Fry, o^. cit 29 Wonderlic, o£. cit. 34 ratings and test scores. The formula ^ ^ 7 “Z was used. The obtained r* s were then divided by the con stant .943 to correct for coarse grouping in the Y variable (ratings), as recommended by Peters and Van Voorhis^^. Table IV lists coarse r*s and corrected r*s obtained for the two groups in each of the sixteen variables. Included also are r*s obtained from Wonderlic scores corrected for age, and r*s obtained in correlating rating against age and education. According to appropriate tables^^, only the underlined coefficients deviated significantly from zero, enabling us to reject the null hypot-hesis of zero correlation. Some of the scatter diagrams suggested the possible existance of a curvilinear relationship between the rating and test scores. Eta coefficients were computed, therefore, according to the formula g* = In order to test the significance of any departures from linearity, as measured by the difference between the eta and coefficients, the chi square test v/as applied: = (N--W) ^ * 30 Guilford, 0£. cit., p. 206. 31 C. C. Peters and W. R. Van Voorhis, Statistical Methods and Their Mathematical Bases (New York: McGraw-Hill Book Company, Inc., 1940), p. 398. 32 Guilford, op. cit., pp. 323-333- 33 Ibid., p. 233- 34 Ibid., pp. 236-237- 35 TABLE IV COARSE AND CORRECTED PEARSON '"r*8 ' FOR SIXTEEN VARIABLES MEASURED IN TWO GROUPS USING FOREMAN EFFICIENCY RATINGS AS A CRITERION GROUP VARIABLE COARSE PEARSON r»S'- CORRECTED PEARSON r»S P M & T S 8 — • o6 -.06 -.07 -.07 P M & T T T .03 .12 .04 .21 P M & T D D —. 18 -.ÔÏÏ P M & T C C -!w -.18 -.ÔÏÏ P M & T R R — • 01 — • o6 — » 02 —. 06 P M & T G G .02 .09 .02 .10 P M & T A A P M & T M M .06 .05 .06 .05 P M & T I I .08 ^14 .08 .14 P M & T N N .14 .# P M & T 0 0 -*èf P M & T Co Co .26 . 1 5 . 36 TABLE IV (continued) COARSE AND CORRECTED PEARSON T»S ’ FOR SIXTEEN VARIABLES MEASURED IN TWO GROUPS USING FOREMAN EFFICIENCY RATINGS AS A CRITERION GROUP VARIABLE COARSE PEARSON T»S- CORRECTED PEARSON "r*S' P Ag •il M & T Ag •Î2 P PFB .04 .04 M & T PFB P BMC — « 06 -.07 M & T BMC .28 •2 2 , P WONDERLIC .u M & T WONDERLIC •M P EDUCATION .24 •il M & T EDUCATION P AGE -.12 M AGE T AGE -.0# 37 In Table V will be found corrected Pearson r*s, cor responding eta coefficients, chi squares, and the signifi cance of the chi squares if at the five percent level or bet ter for those diagrams which suggested a possible curved relationship. It will be seen that significant curvature was present in only three instances; for the maintenance and tooling group in the case of traits agreeableness and objectivity, and for the production group in the case of the Wonderlic test. Inspection of the scatter plots suggests that in the former group the best men tend to score neither very low nor very high in agreeableness and objectivity, and that in the production group there is apparently an optimum score in intelligence for the greatest success. The fact that every eta coefficient is higher than the corresponding Pearson prob ably is due to some degree of curvature and also to the addi tional weight of the deviants at the extremes of the distri bution which gain their maximum influence when the correla tion ratio is applied. In general the tables of correlations indicate support for the following hypotheses: There was a significant positive correlation between cooperativeness and the criterion for both groups of foremen. Agreeableness correlated positively and significantly for both groups. In the maintenance and tooling group there 38 TABLE T PEARSON r*S, CORRESPONDING ETA'S, CHI SQUARES, DEGREES OF FREEDOM, AND SIGNIFICANCE OF THE CURVATURE WHERE FOUND FOR NINE TEST VARIABLES GROUP VARIABLE PEARSON r ETA CHI SQUARE FREEDOM SIGNIFICANCE P R .02 .18 9.105 10 M & T R .06 .34 14.182 15 P A .13 .26 15.461 9 M & T A .12 .32 10.696 12 ——— P M .06 .16 6.669 11 M & T M .05 .31 11.033 10 ——— P 0 .15 .24 10.598 8 M B « M M B M & T 0 .03 .40 20.245 11 .05 P Co .27 .32 8.847 15 M & T Co .26 .41 12.809 14 M B — M B P Ag .14 .22 8.851 13 M . M B M . M & T Ag .20 .50 29.309 12 .01 P PFB .04 .12 3.485 8 M & T PFB .26 .38 7.147 13 ——— P BMC .07 .17 8.440 8 M & T BMC .30 .40 11.033 9 M B M B . M P WONDERLIC .14 .29 19.769 8 .02 M & T WONDERLIC .34 .38 3.396 11 ——— 39 was considerable curvature. The best foremen are neither too agreeable nor too disagreeable. Intelligence as measured by the Wonderlic correlated positively and significantly in both groups. A correction for age on this test produced no appreciable change in the correlations. Age correlated negatively vvith success in all but the tooling group. This is as expected since it is likely, that for any particular job level, there will be some older men who have reached their limit of achievement and some younger men who have a capacity greater than the job calls for and are on their way to more demanding positions. The greater energy of the younger men also is probably related to the negative relationship. However, the correlation betv/een age and criterion is probably not high enough to influence cor relations with other trait scores appreciably. Thinking introversion correlated positively and signif icantly in the maintenance and tooling group. The correla tion was positive but insignificant for the production group. Possibly the difference here lies in the necessity for tool ing men to vievv their work in a meditative, analytical manner. The tendency toward depression correlated negatively and significantly with success in the production group. It was negative but insignificant in the maintenance and tooling 40 group. The better,foreman tends to be less depressed. Cycloid tendency also correlated negatively with suc cess. The correlation was significant in the production group. The better foreman is of more even temperament. There was a positive correlation between ascendancy and success in both groups, the relationship being signifi cant in the production group. Lack of nervousness was also an indicator of success, significantly so in the production group. The tendency to view things objectively correlated significantly with the criterion in the production group. In the maintenance and tooling group a significant curvilinear relationship was found, indicating that the.best men score neither very high nor very low in objectivity. Mechanical aptitude as measured by the Minnesota Paper Form Board correlated positively with success, and signifi cantly so in the maintenance and tooling group. Mechanical comprehension as measured by the Bennett correlated positively and significantly with the criterion for the maintenance and tooling group. However it was a slightly negative indicator in the production group. Evi dently the nature of the work demands mechanical aptitude In both groups, since both work with or on tools and machines. However, only in the maintenance and tooling group, where construction of machinery and application of mechanical 41 principles is important, are scores in mechanical comprehen sion significant. Education correlated positively with the criterion for both groups. The correlation was significant only in the production group. Indications from the scatter diagram for this group are that education beyond graduation from high school is not important. The following traits were insignificantly correlated with the criterion for both groups. The trends in every case were consistent however and may be suggestive: social introversion, slightly negative; self-restraint, slightly negative; general activity, slightly positive; masculinity, slightly positive; and lack of Inferiority feelings, slightly positive. III. DETERMINING PROFILES The next procedure was to make a graphic profile of scores for good and poor foremen in those variables which correlated significantly with the criterion. Because the correlations were generally low, only profiles for men rated ”1” and ”5” were constructed. These would .tend to demon strate the maximum discriminating power of the tests. Mean scores for men rated ”1” and ”5” in traits with significant correlations in one or both groups will be found in table VI. 42 TABLE VI MEANS SCORES OF FOREMEN RATED "5" AND "1" ON ELEVEN VARIABLES WHICH CORRELATED SIGNIFICANTLY WITH CRITERION GROUP VARIABLES M5 Ml P Co 69.9 47.2 P 0 52.8 44.7 P A 25.0 22.5 P WONDERLIC 21.5 16.3 P Ag 36.5 33.6 P N 33.1 29.4 P D 12.1 15.6 P C 16.8 21.8 M & T Co 76.1 52.0 M & T PFB 40.3 28.3 M & T BMC 34.7 21.3 M & T WONDERLIC 23.9 10.3 M & T Ag 40.8 32.3 M & T T 37.3 30.3 43 Raw scores in the variables concerned were next C- scaled according to the method described by Guilford^^. This procedure normalizes the distribution of raw scores into eleven intervals (0-10) which yield the following percentages for each 0—score; 1 — 3 ■ * 7 — 12 — 17 — 20 — 17 — 12 — 7 — 3 - 1. If this process is repeated for each variable, the C- soores in all tests are comparable and convenient profiles may be constructed. Figure I shows the profiles for the best and poorest foremen in the production group. Although there apparently is no discrimination offered by traits D and C, an inspection of the means on those traits will show the best foremen to be at the upper extreme of that C-interval for both traits, and the poorest foremen to be at the lower extreme. Since a C-score of ”5” includes the middle 20^ of the total group, scores at the extremes of that interval represent some separa tion. Figure II shows the profiles for the best and poorest foremen in the maintenance and tooling group. In only three instances were the traits the same as those graphed for the production group. Separations were somewhat more distinct. These profiles should be useful as indicators of the general areas into which good and poor supervisors fall in the plants under consideration. 35 Ibid., pp. 104-107. FIGURE I PROFILES FOR BEST (BLUE) AND POOREST FOREMEN (RED) IN PRODUCTION GROUP C-Score 0 Co A WONDEPXIC N D C 10 9 8 7 6 5 4 3 2 1 0 77 96 55 37 76 97 54 36 71 90 51 34 70 69 50 33 65 81 47 31 64 80 46 30 59 73 43 28 56 42 54 64 V"36] 24 4 ^ 57 23 33 32 31 30 28 49 0 0 46 0 0 45 0 0 44 0 0 41 1 4 40 2 5 37 6 10 7 11 10 15 3?t 41 ■ 30 19 12 25 21 27 36 27 17 10 22 24 32 35 40 26 16 9 21 25 33 30 32 22 14 6 18 29 37 29 31 21 13 5 17 30 36 24 24 18 - 11 2 14 34 43 23 23 17 10 1 13 35 44 0 0 0 0 0 0 45 FIGURE II PROFILES FOR BEST (BLUE) AND POOREST (RED) FOREMEN IN MAINTENANCE AND TOOLING GROUP C-Score T Co ____ WONDERLIC PFB BMC 10 9 8 7 6 5 4 3 Z 1 0 54 53 50 46 20 19 16 15 0 100 99 92 91 84 33 32 25 24 0 58 57 54 53 49 24 23 20 19 0 34 33 30 29 26 2 0 0 0 55 54 50 49 45 17 16 13 12 0 53 52 48 47 43 41 21 18 28 28 18 22 6 21 13 12 8 7 0 46 IV. AN INTERPRETATION OF THE VALIDITY OBTAINED In 1939 Taylor and Russell^^ demonstrated that tests with relatively low validity coefficients may he very useful under certain conditions of selection. The usefulness of a test will vary with two conditions: the proportion of present employees considered satisfactory, and the selection ratio (the proportion of jobs to be filled divided by the number of applicants for the job). If one has two foremen’s jobs open and twenty eligible men, the selection ratio becomes .10. The two men scoring highest wdll. be chose# (assuming a straight line relationship). The chances that these men will be as successful as the present group of successful foremen will then vary with the size of r. These probabilities are given in table VII which is a partial reproduction of the Taylor and Russell Tables. Proportions of .10, .30, and .70 are included because in this study approximately 10 percent of ”5’ * ratings were given, 30 percent of "5” and "4" ratings given, and 70 percent of **5” and ”4** and **3” ratings given. The proportion .50 is presented for comparative purposes. Table VIII shows how these probabilities apply to the Pearson r’s obtained in this study. Since foremen are a select group, and the odds rather high that there will always be considerable competition for such jobs, only proportions 36 Taylor and Russell, 0£. cit. 47 TABLE VII TABLES OF THE PROPORTION WHO WILL BE.SATISFACTORY AMONG.THOSE SELECTED, FOR GIVEN VALUES OF.THE PROPORTION OF PRESENT EMPLOYEES CONSIDERED SATISFACTORY, THE SELECTION RATIO, AND ' r‘ -\ (AFTER TAYLOR AND RUSSELL) Selection Ratio r .05 .10 .20 .30 .40 .50 .60 .70 .80 .90 .95 Proportion Satisfactory - .10 .10 .14 .13 .13 .12 .12 .11 .11 .11 .11 .10 .10 .15 .16 .15 .14 .13 .13 .12 .12 .11 .11 .10 .10 .20 .19 .17 .15 .14 .14 .13 .12 .12 .11 .11 .10 .25 .22 .19 .17 . 16 .14 .13 .13 .12 .11 .11 .10 .30 .25 .22 .19 .17 .15 .14 .13 .12 .12 .11 .10 .35 .28 .24 .20 .18 .16 .15 .14 .13 .12 .11 .10 Proportion Satisfactory = .30 .10 .38 .36 .35 .34 .33 .33 .32 .32 .31 .31 .30 .15 .42 .40 .38 .36 .35 .34 .33 .33 .32 .31 .31 .20 . 46 .43 .40 .38 .37 .36 .34 .33 .32 .31 .31 .25 .50 .47 .43 .41 .39 .37 .36 .34 .33 .32 .31 .30 .54 .50 .46 .43 .40 .38 .37 .35 .33 .32 .31 .35 .58 .54 .49 .45 .42 .40 .38 .36 .34 .32 .31 Proportion Satisfactory I .50 .10 .58 .57 . 56 .55 .54 .53 .53 .52 .51 .51 .50 .15 .63 .61 .58 .57 . 56 .55 .54 .53 .52 .51 .51 .20 .67 .64 .61 .59 .58 . 56 .55 .54 .53 .52 .51 .25 .70 .67 • 64 .52 .60 .58 . 56 .55 .54 .52 .51 .30 .74 .71 .67 .64 .62 .60 .58 . 56 .54 .52 .51 .35 .78 .74 .70 « 66 .64 .61 .59 .57 .55 .53 .51 Proportion Satisfactory % .70 .10 .77 .76 .75 .74 .73 .73 .72 .72 .71 .71 .70 .15 .80 .79 .77 .76 .75 .74 .73 .73 .72 .71 .71 .20 .83 .81 .79 .78 .77 .76 .75 .74 .73 .71 .71 .25 . 86 .84 .81 .80 .78 .77 .76 .75 .73 .72 .71 .30 .88 .86 .84 .82 .80 .78 .77 .75 .74 .72 .71 .35 .91 .89 .86 .83 .82 .80 .78 .76 .75 .73 .71 4.8 TABLE VIII PROPORTION OF SUCCESSFUL FOREMEN WHICH WOULD BE SELECTED BY SEVERAL TESTS UNDER CONDITIONS OF A 5% SELECTION RATIO (10^ SELECTION RATIO IN PARENTHESES) VARIABLE PEARSON ' r ■ PRODUCTION GROUP PROPORTION SUCCESSFUL tt^tf "4"&"5" 50% A .13 •15(.14) .40{.38) .6K.59) .79(.78) Ag .14 .16(.15) .4K.39) .62(.60) .79(.78) N .14 .16(.15) .41(.39) .62(.60) .79(.78) WONDERLIC .14 .16(.15) .4K .39) .62(.60) .79(.78) 0 .15 .16(.15) .42(.40) .631.61) .80(.79) D .18 .18{.16) .44(.42) .65(.63) .82(.80) 0 .16 .18(.16) l44(.42) .65(.63) .82(.80) Co .27 .23(.20) ,52(.49) .72(.79) .87(.85) VARIABLE PEARSON ’ r" MAINTENANCE AND PROPORTION TOOLING GROUP SUCCESSFUL "4"&"5" 50% "3"&"4"&"5" Ag .20 .19(.17) .46(.43) .67(.64) .33(.81) T .21 .20(.17) .47(.44) .68(.65) .84(.82) Co .26 .23{.20) .51(.48) .71(.68) .86(.84) PFB .26 .23(.20) .51(.48) .71{.68) .86(.84) EMC .30 .25(.22) .54(.50) .74(.71) .88(.86) WONDEgLIC .34 .27(.24) .57(.51) .77(.73) .90(.88) 49 of successful selectees under selection ratios of 5^ and 10% (in parentheses) will be shown. Assuming rectilinear regressions, as the Taylor-Russell method demands, the most valid test for the production group was Guilford’s Co. Using that test alone, under the condi tions of a 5% selection ratio, one would be able to select 23 percent foremen rated "5," or 52 percent rated ”4” or "5," or 87 percent rated ”3” or ”4” or "5*" Stated another way, use of this one test alone under the above conditions, yields an increase of 13 to 22 percent of successful workers. The most valid test for the maintenance and tooling groups proved to be the Wonderlio Personnel Test. Under the ideal conditions mentioned above, this test alone would enable one to select 27 percent foremen rated "5," 57 percent rated ”5” or ”4>” or 90 percent rated ”5” or ”4” or "3." Thus this test yields an increase of from 17 to 27 percent of successful foremen depending on the level of success desired. It seems reasonable to assume that some combination of the several tests would yi&ld considerably higher validity. This is another supplementary problem arising from this study which should be investigated. Table IX below shows the per centage increase in successful foremen selected through the use of each test for the top three ratings in this study. It should be emphasized that there were two conditions operating which acted to suppress considerably the apparent TABLE IX PERCENTAGE INCREASE OF SUCCESSFUL FOREMEN CHOSEN UNDER CONDITIONS OF 5% SELECTION RATIO USING THE SEVERAL SIGNIFICANT TESTS 50 VARIABLE PRODUCTION GROUP RATING ”5” "3"&"4"&"5" A 5 10 9 Ag 6 11 9 N 6 11 9 WONDERLIC 6 11 9 0 6 11 9 D 8 14 12 C 8 14 12 Co 13 22 17 MAINTENANCE AND TOOLING GROUP VARIABLE I I ^11 RATING "3"&"4"&"5" Ag 9 16 13 T 10 17 14 Co 13 21 16 PFB 13 21 16 BMC 15 24 18 WONDERLIC 17 27 20 51 effectiveness of the tests. One is that the Taylor-Russell method is based on rectilinear relationships. That many of the relationships in this study departed from linearity was demonstrated in table V on page 38. The second suppressing factor was that the reported increases in selection effi ciency had been determined from a group that was already select. Should the same tests be administered to a more heterogeneous group of workers, such as the one from which foremen would ordinarily be selected, the correlation coef ficients would very likely be increased considerably with a corresponding increase in selection efficiency. V. SUMMARY Norms have been established in sixteen psychological variables for two groups of industrial foremen. These should serve as landmarks for future samples of men in the same test environment. No fruitful comparisons could be made with the available published norms on these tests. Product-moment correlations were computed for each test variable against the criterion, and for age and educa tion against the criterion. Significant correlations were found in many instances. In cases where the scatter diagrams suggested a curvilinear relationship, correlation ratios were were also computed. The chi square test of linearity was applied and revealed curvature existing for regressions in 52 the case of objectivity and agreeableness for the maintenance and tooling group, and for intelligence in the production group. In every case the eta coefficient was higher than the Pearson r. Raw scores in the variables correlating significantly with the criterion were C-scaled and profiles constructed which distinguished the best from the poorest foremen in each group. For the maintenance and tooling group, the best super visors excel the poorest on Guilford’s cooperativeness, agree ableness, and thinking extroversion, and on the Bennett Mechanical Comprehension, the Wonderlio Personnel Test and the Minnesota Paper Form Board Test. For the production group, the best supervisors excel the poorest on the Wonderlio Personnel Test, and on Guilford’s objectivity, cooperative ness, agreeableness, and ascendancy, and have less tendency to nervousness, depression and cycloid disposition. An interpretation of the validity of these tests as recommended by Taylor and Russell indicated that under selec tion ratios of five percent in the production group, the percentage of foremen rated ’ ’5’ ^ could be increased as much as 13 percent through the use of one test; those rated ”4’ ^ or ”5” could be increased as much as 22 percent; and those rated ”3” or ”4” or *^5” as much as 17 percent. Similarly in the maintenance and tooling group the use of one test could produce a 17 percent increase in the selection 53 of foremen rated ”5"; a 27 percent increase in those rated ”4” or ”5”; or a 20 percent increase in those rated ”3>” ”4>” or "5*" CHAPTER V SUMMARY AND CONCLUSIONS It has been the purpose of this study to investigate relationships between a criterion of foreman success and scores on several psychological tests. To this end, norms have been established, profiles constructed,and correlations determined for sixteen test variables. This chapter presents a summary of the study together with conclusions and recom mendations for further investigation* Previous studies in this area were unable to reveal any consistent pattern of traits which differentiated the leader from his group. The most frequent hypothesis encount ered was that the best foreman possesses certain traits to the same extent as his fellow workers, and certain other traits to a greater degree than they. Just what traits were signi ficant seemed to be a function of each particular situation. Obviously the foreman should be above average in job knowledge and proficiency. But there was widespread belief that the more important differences lay in the area of temperament, and the few previous studies which approached the problem through the measurement of personality had yielded only modest results. This study substantiated the hypothesis just mentioned to a considerable extent. For a group of 289 foremen in 55 production groups it was found that the best foremen differ significantly from the poorest foremen in eight respects: greater objectivity, cooperativeness, and agreeableness; less nervousness, depression, and cycloid tendency; and higher intelligence. All of these traits correlated signi ficantly with the criterion, which was the proficiency rat ing for each foreman. These ratings were established by averaging the independent estimates of three immediate super iors for each man. For a group of 120 foremen in a combined maintenance and tooling group, the superior foremen made better average scores in the following areas: cooperativeness, agreeable ness, thinking introversion, mechanical aptitude, mechanical comprehension, and intelligence. All of these traits showed significant correlations wdth the criterion. The following trait scores showed no differentiation between good and poor foremen in either plant: social extra version, rhathymia (happy-go-lucky), general pressure for overt activity, masculinity, and inferiority feelings. For these groups these traits may be regarded as areas in which both superiors and inferiors hold common qualities. Certain variables tended to depart from a rectilinear relationship with the criterion, some of them significantly so. In the maintenance and tooling groups, the best men scored neither very high nor very low in objectivity and agreeableness 56 In the production group there was an indication that there was an optimal score in intelligence above which few of the best foremen fell. Many other variables showed evidence of curvature, although they failed to show a significant depar ture from linearity. The increased correlation coefficients when eta was computed indicated the importance of recognizing and testing for non-linear relationships. Although the obtained validity coefficients were moderately low, considerable usefulness was demonstrated through a method of interpretation devised by Taylor and Russell. Under a five percent selection ratio, the tests which correlated highest with the criterion would produce up to a 2 2 percent increase in successful foremen in the pro duction group, and up to 27 percent increase of successful foremen in the maintenance and tooling groups. These per centages might be considerably increased if some form of composite score, successive hurdles, or multiple cut-off score were adopted. It appeared very likely that a fruitful approach toward greater validity of the test battery would be through the technique of item analysis as suggested by the Uhrbrock and Richardson study (page 5). It is possible that a much shortened selection device with considerable validity might be obtained through such a process. Since several of the sub-tests showed little discriminating power in one group or 57 the other, one should expect to ,find considerable "dead wood" in such a battery. Lack of discriminating power within a group of foremen, however, does not necessarily mean that such tests would fail to discriminate in a less select group. Should the same battery of tests be administered to the heterogeneous group of workers from which one would ordinarily select new foremen, considerably higher validity coefficients would be expected. In addition to this, a consideration of the curvature of the regressions would indicate validities greater than those which can be used vdth the Taylor and Russell method. In other words, had allowances for curvature and homogeneity been pos sible in this investigation, higher validity coefficients probably would have been obtained together v/ith a correspond ing increase in selection efficiency. It is believed that this study has demonstrated some progress in the problem of foremen selection throu^ psycho logical testing. A step has been taken in the direction of determining the human traits needed for success in the job of foremen. It is to be expected that as selection devices are refined and less fallible criteria established, the human profile for the excellent foreman will become better and better defined. BIBLIOGRAPHY BIBLIOGRAPHY A. BOOKS Bingham, W. V. D., Aptitudes and Aptitude Testing. New York: Harper Brothers, 1937• Buros, 0. K., Mental Measurements Yearbook* 1940. Guilford, J. P*, Fundamental Statistics In Psychology and Edu cation* New York: McGraw-Hill Book Company, Inc., I94S. _______ , Psychometric Methods. New York: McGraw-Hill Book Company, Inc*, 193?I Paterson, D. G., Elliott, R. M*_, Anderson, L. D., Toops, H. A., and Heidbreder, E*, Minnesota Mechanical Ability Tests* Minneapolis: University of Minnesota. Peters, C. C., and Van Voorhis, W. R., Statistical Methods and Their Mathematical Bases. New York: McGraw-Hill Book Company, Inc., 1940* Terman, L. M. and Miles, C. C., Sex and Personality. New York: McGraw-Hill Book Company, Inc., I936. B. PERIODICAL ARTICLES Bennett, G. K. , and Cruikshank, R. M., "Sex Differences In Understanding of Mechanical Problems," Journal of Applied Psychology, 26:121-127, 1942. Bennett, G. K., and Fear, R. A., "Mechanical Comprehension and Dexterity," Personnel J ournal. 22:12-17, 1943. Cowley, W. H., "Three Distinctions In the Study of Leaders," J ournal of Abnormal and Social Psychology, 23:114-157, 1928. Crawford, J. E., "Measurement of Some Factors Upon V/hich Is Based Achievement In Elementary Machine Detail Drafting," PhD. Thesis presented to the University of Pittsburgh, 1941. Ellis, A., "Validity of Personality Questionnaire," Psychologi cal Bulletin, 43:385-440, 1946. 59 Goodwin, P., "An Introductory Study of the Prediction of Personal Leadership Ability Based on Group Attitudes and Opinions," Unpublished Master’s Thesis, University of Southern California Library, 1944* Guilford, J. P., "A Simple Scoring Weight for Test Items and Its Reliability," Psychometrica, 4:367-374, 1941# ,editor. Printed Classlfication Tests, Report No. 5 AAF Aviation Psychology Program Research, Washington, D. C. Guilford, J. P., and Guilford, R. B., "An Analysis of the Factors in a Typical Test of Introversion-Extraversion," J ournal of Abnormal and Social Psychology, 28:377-399, 1934. , "Personality Factors D, R, T, and A," Journal of Ab- normal and Social Psychology, 34:21-36, 1939. _, "Personality Factors N and GD," Journal of Abnormal and Social Psychology, 34:239-248, 1939. _______ , "Personality Factors 8, E, and M and Their Measure ment," J ournal of Psychology, 2:107-127, 1936. _______ , and Martin, H. G., "Age Differences and Sex Differen ces In Some Introvertive and Extravertive Traits," Journal of General Psychology, 31:219-229, 1944. Harrell, W., "Testing Cotton Mill Supervisors," Journal of Applied Psychology, 24:31-35, 1940. Eoviand, C. I., and Wonderlic, 1. F., "A Critical Analysis of the Otis Self-Administering Test of Mental Ability," J ournal of Applied Psychology, 23:367-387, 1939. Jackey, D. E., "Supervision II— Training," Personnel Journal, 20:357, 1942. Jacobsen, E. E., "An Evaluation of Certain Tests in Predicting Mechanic Learner Achievement," Educational and Psychologi cal Measurement, 3:259-267, 1943. Jenkins, W. 0., "A Review of Leadership Studies With Particu lar Reference to Military Problems," Psychological Bulletin, 44:54-79, 1947. Johnson, R. H., "The Inheritance of Personality," Education, June 1941, 592-597. 60 Kuder, G. F. , and Richardson, M. W., "The Theory of the Estimation of Test Reliabilities," Psyohornetrioa, 2:151- 160, 1937. Lovell, C., "A Study of the Factor Structure of Thirteen Personality Variables," Educational and Psychological Measurement, 5:335-350, 1945. McMurray, R. N., "Psychological Problems in Industrial Supervision," J ournal of Consulting Psychology, 8:175- 187, 1944. Martin, H. G., "Locating the Trouble-maker with the G-M Personnel Inventory.” Journal of Applied Psychology, 28:461-467, 1944. Moore, H., "Supervision I— Selection," Personnel Journal, 20:353-362, 1942. Hosier, C. I., "A Factor Analysis of Certain Neurotic Tenden cies," Psychometrica, 2:263-287, 1937. Pfiffner, J. M., "A Pattern For Improved Supervisory Leader ship," Personnel, 24:271-280, 1948. Quasha, W. H., and Likert, R., "The Revised Minnesota Paper Form Board Test," Journal of Educational Psychology, 28: 197-204, 1937. " Richardson, H. M., and Hanawalt, N. G., "Leadership as Re lated to the Bernreuter Personality Measures," J ournal of Applied Psychology, 28:308-317, 1944. Sartain, A. Q., "Relation Between Scores on Certain Standard Tests and Supervisory Success in an Aircraft Factory," Journal of Applied Psychology, 30:328-332, 1946. Shuman, J. T., "The value of Aptitude Tests for Workers in Aircraft Engine and Propeller Industries, "J ournal of Applied Psychology, 29:156-160, 1945. Stockford, L., "Selection of Supervisory Personnel," Person nel, 24:186-199, 1947. Symonds, P. M., "Choice of Items for a Test on the Basis of Difficulty," Journal of Educational Psychology, 20:481-493, 1929. 61 Taylor, H. C., and Russell, J. T., "The Relationship of Vali dity Coefficients'to the Practical Effectiveness of Tests in-Selection," Journal of Applied Psychology, 23:565-578, Thurstone, T. G., "The Difficulty of a Test and Its Diagnostic Value," Journal of Educational Psychology, 23:335-343, 1932. ----- Toops, H. A., "Some Concepts In Job Families and Their Impor tance In Placement," Educational and Psychological Mea surement , 5:195-2l6y 1945* Uhrbrock, R. S., and Richardson, M. W., "The Basis for Con structing a Test Forecasting Supervisory Ability," Per sonnel Journal. 12:141-154, 1934* Vogel, E. M., "An Attempt to Measure the Industrial Supervisor’s Ability By Means of a Self-Inventory Test," Unpublished Master’s Thesis, University of Southern California Library, 1945. Wonderlic, E. F., and Hoviand, C. I., "The Personnel Test: A Standardized Abridgment of the Otis S-A Test for Business and Industrial Use," J ournal of Applied Psychology, 23:685 702, 1939. Wright, J. H., and Lang, D. M., "The Time Factor in the Admin istration of the Wonderlic Personnel Te'st," J ournal of Applied Psychology, 27:27, 1943. ________, Army Air Forces Aviation Psychology Program, Research Report No. 6, 179-217. , "How To Select New Supervisors," Personnel, 20:66-79, 1943. C. MANUALS Bennett, G. K., and Fry, D. E., Manual of Directions for Test of Mechanical Comprehension, j^orm BB. Psychological Corp oration, New York City. Guilford, J. P., An Inventors^ of Factors S, T, D, C, R. Man ual of PirectTons and Norms (Revised). Sheridan -Supply Company, BevsLy Hills, California. 62 Guilford, J. P., and Martin, H. G., Inventory of Factors G, A, M, I, N. ..Manual of Directions and Norms (1st Revision). Sheridan Supply Company, Beverly Hills, California. _______ , The Guilford-Martin Personnel Inventory. Manual of Directions and Norms. Sheridan Supply Company, Beverly ËÏTls," CaiiTSrnlaT” Likert, R., and Quasha, W. H., Revised Minnesota Paper Form Board, Series A A ahd_BB — Manual of Directions. Psycho- logical Corporation, New York,City. Otis, A. S., Manual of Directions and Key (Revised) for Otis Self-Administaring est of Mental Ability. World Éook, New York. Wonderlic, E. F., Wonderlic Personnel Test Manual, 1945. 750 Grove Street, Glencoe, Illinois. UniversiV of Sou^ern California Libravy APPENDIX 63 to r r \ (Q. vsl 02 vO CO to \o CO to to I —I C O rH fO to » o vO vO c f i h“ C £ < X o o 02 to 02 pH oO I —I T3 C r to CO CO in I —I DO z o _i W QC OC O O to 02 O I—f CO 02 CO rH o O rH ^ O • o O 02 O rH 02 vO to rH l O _J 6 \o o CO 02 uo rH OO 00 ON O to pH to VÛ rH w X. H « 0 I —I O to to 02 O 0 0 O uo *o rH K > 0 0 OO t“ 00 00 02 64 to W to CV2 vO o to \o to to m 683 C \ J to «0 0 0 r- vO r4 -m C \ 2 O C \ J H CC < X o to o CVi to to to to o oO !K ID O to to O O 0 0 o Z o !< w c r oc o o vO 90 to c v o \o r4 O O to to to 0 01 l O o H _ l 6 o ! ---I o 00 vO r4 -W - O u X H CVi to to to to o to o o to vo N ) OO •o 00 65 CV vO c v o o > 5 00 OO to » o (V I—I CD 00 to o lO to o o o C \ 2 h" C C < X o o o o OD to CV to o o o \o OO z o !< p to o rH UJ cc c r o o n O O O CM to - 4 to rH CM H l O CM H 6 to O CM O CM f l O 00 to C M to UJ X H c o cMvn CM \o rH vO to C M * o i T \ C ^ CM o rH l O to to K > <o 0 0 OO DO w ONItLVa to rH 02 02 vq ON CO CO NO 00 ON VN NO NO rH ro m ro 02 % «D N H C C < X o 02 G O ro CO rH CO 02 z o !< _j IxJ c r c c o o o NO CO CO CO o r4 o C D 02 w O ON rH O 02 02 02 rH CO n 02 rH 6 00 CO O o IxJ X H « D CO CO 02 CO ( O CO o >0 eo OO 00 n rH 02 67 vq vO On v\ to NO to » o in 6 d Z On U - \ K“ C C < X o to o Q O' • NO <r\ o H to •H Csi O o CVi Z O !< d q: c r o o 00 CD ON cv m ON M I —I M Cr O CV to to tr CV2 00 NO C\i 6 o 00 00 W X H 00 00 0 3 00 0 01 o v> •o to to 0 9 09 90 M- 68 V v J 1 '4 - H t o (\2 ON o O « 4 ^ ON t o % ' f-N c T X I & rH . » « .> > X N £>- NO C N 2 It % *- ■ K l WN WN 1 —1 C \2 4 C \2 1 1 - ? N + vO ON Ov ON rH -4" - 4 cv % vO o \ ON VN NO NO t o NO . I f T M t o -<h ON iTN NO NO t ^ 4 1 + » o + j - + K 3 4 - cQ + «"4 4 - o 1 N 1 r o 1 j - 1 IT i I < 0 1 '4 ' C \2 ON W \ Cr\ C\2 -------- L NO NO !> - 1 —1 6 8 2 - / V «+- K l » T N 1 t o 4- A ;K C B m t» H 1 ? ? p a * 1 H t ^ + o 4 o 1 4 1 i r -<) & ;** N s s o f a ? ' ? O + o C o O - t 1 cv % k N b 5 ^ CO «0 -«4 1 9 b 1 ? ? rH iH lO 4 - UN W N W' I> C N J W N O N O 1 —1 H 1 O' * ,4 !S î i t o d' c r\ 1 —1 ■ s '” » a O 1 a t ? f ? cn C T N d - 4" C N 2 ON iH to C \2 x r 4 NO C N 2 ON C O c^ 1 VÛ fy \ Ï5 b 5 o* ^4- « t o n NO 1 —1 rH ? W N f r f - 4 K > + C \2 ON iH NO c r O ' t o 4 ON NO C f> 1 ^ rr> s « 0 « CO al* ON n 1 —1 1 —1 •4 1-4 1 1 —1 d> 1 >» 1 C D 1 ? A r O Î 4 - C \2 t o 4 NO rH O ON NO C N 2 Ph cv O r r \ E4 1 O O < « . to m . - 4 c 4 O 1 —1 -4" C\2 WN T CN •4 1 ? d - 1 b f NO - 4 kH + NO -4 - NO 4 t o rH 1 —1 1 —1 — kjtj " P ', u# Pq CV 1 ^ O - 4 NO ON rH NO - 4 ' ON c ^ O O ------------- t v I (\2 1 c!j 02 ^ v S i? d* r 1 — I •d NO rH NO 1 —1 A A d - « > 1 —1 ON - 4 1 1 —1 ON 1 rH O N t o ON Cvi 1 rH O C\2 f) T i s 1 Oft 1 ? iH d- w \ 1 - 4 C \2 c 4 O d - O 00 £ a C \2 iH 1 - 4 cv 1 t o 4 4 4 rH r v 1 t o r 4 ? ? » o C \2 1 0 NO C f\ 1 —1 d- a 2 2 C \2 iH r > 1 NO ON 1 t o o H WN 1 —1 NO 1 —1 1 ir> a t ? s ? 1 —1 0» O d > 1 0 ) NO d* C \2 A 9 S ON à- 1 NO ON 1 : | rH O 4 t o ' ; 1 ' " 1 C v 2 1 1 ? ; $ 1 H 1 —1 « 0 9 9 o N R % ON * o 1 WN iH ____f-j W N Dn . W N W N 1 —1 H 1 a g ? d . ? ? f C\2 C \2 « A d) 4 - p* % » ■4- 0 1 ■ 4 C N 2 1 H ( \f 1 —1 O t o 1 VO i t ; ? f r rH C N 2 ea « 4 8 a 4- ON 1 iH C N 2 1 1 —1 4 H O : t § L 1 ? ? s eo 1 «B 9 a g i 09 d - 0 0 I wo '4 - C^ C\2 H X ( + - a X ■+• — 69 to to I— I to o o " \ rH rH ro lO m o o o «0 oa O Oi rH o H C C < X o O' O O lO ir rH -<) 0 0 z o !< UJ (T (T o o to o o o C Q ^ o M o o C \ 2 to O i o <0 to to _ l Ô to M to to rH I> 00 o rH o > VÛ to rH O bJ X H C v i rH « o to \o 0 0 rH I —I C V 2 K ) o O to o o\ to to I —I N ) to rH rH 04 0 » • o to eo « 0 AX' 70 - î V s l 1 w V \ o fC) C V i C V i rH ON O O C V i 1 1 ■ 9 c w t o t o 1 -> > X N C vi H UN ON l V t o t o 1 1 1 - ? N + rH NO VN t'V t o ------ NU to C + - nO ON o w v nO NO t o NO ' O - i l ■5M t o ON trv / ' NO NO 1 c * " ' 1 : ------ 1 I I %4- ■ K l + » o + à- + K ) 4 - c 4 4 - tH 4 - o <H 1 r 4 1 r o 1 j - 1 m ! > 0 1 C V i ' ON WN c r\ C V i NO NO H 6 8 2 r / V 4 - K l A hÎ CD m r* rH r * 1 f ? ? r A = * " 1 + ON & g y* iS 5 3 o f â ? ? p + 1 W \ XT\ % k N & Z 5 *o rH U 1 O fe 1 . f ? rH lO 4 - W N W' (Ni 1 o y. (4 H ïS * % y. O ys ^ 00 1 A ? ; ? cr> à- 4 - C V i 1 —1 t o -«4 o C V i rH u \ 0 \ 1 t r \ Î5 !2 a O* fH • o m C V i O M f 5* 1 # f r o + C V i rH NO O N NO NO -4 ‘ 1 O <T\ S ** « O y* C V i n rH rH 1 m 1 ? A T t o r î 4 - NO H C V i O N -4 - C V i rH O r r \ 1 Ph ^ O "4 W y. •Q . O c 4 fH T rH H 1 ? f Y f (VN rH ▼4 4 - < n rH ON H rH NO E -1 o c \ 1 k o rH wv NO UN rH t o rH o O v\ (J \ ? <v C5 VN !? y* ? r r \ A UN C V i rH r r \ rH A A » . > * > 4 > C V i ^-4 1 C V i 1 -st C V i W N rH f —1 O j 1 o C \2 O 1 ? e o •O i W y> 1 NO rH NO r r \ sK O 00 o o - e n n i 1 -d * r - i t o " < > " fH ->à' C V i - t ON r 4 1 U -\ rH $ ? 0» 1 C V i f t o to C V i (Ni rH m rH a 2 2 £>- -H" r > 1 rH 1. ON C V i t o Hh NO ON r4 1 O rH 1 o r s ? *rs M ON y. rH 1 C *N < r\ fH * r o« A 8 W N o - d - 1 O o ON O O (Ni H o o H NO rH rH ON 1 » r\ f ü ; 5 , H ^ rH O C V i 1 * 1 rH s ^ 2 o N 3 % rH W N » o 1 1 WN WN C V i WN C n- C V i o t o W N O rH 1 O s , 8 y . ? C V i f H f CN C V i A A y - r» % & t o > û 1 t o - t 1 1 —1 t o t o (Ni C V i H O e n ; ? f r t» 0» *4 8 A :*- r - 1 % t f f ? ? 00 1 60 5 c 4 W î 00 ;h 00 1 / W < T \ C V i H ( f X X A X s - -f — O.MIiVH o vsl to — Of 1 1 vs) w \C ro m Q D P* • O lO o O «0 a O ro o w o * O CD O oi (M m UJ (T O X Û to m § O 0 0 to o < 5 0 O CD C 8 0 0 \D m o O f HI O K) OO OO n m 72 CO o ON i r s vO On O CO I —I to to » o m r4 c O VO Cv2 CD CO p* » r > NO rH CNÎ vO W VL) vO O NO o H C H < X o o CO vO ON o o O £ > vO vrs w 02 to o o o o r—I o • o z o CO o CO rH CO -I W cc cr o o o CO o I —I o Of o 6 CO Of CO fH H e o > 0 0 0 O TXT O LJ X H co « 0 ON 02 Of Of fH H >o 0» t - Û 2 . 0 0 OO Xx 73 CO rH O rH I-- f vO CO 02 ro » o 02 OB CO 02 02 02 I CO o H a : < X o CO 02 VO eO lO o > o> ON 02 02 O Z o VO I—! o CD vO W tr ÛC o o o. g CO o i r v rH O ■COHS^ 02 n Ô CO 02 O 02 0 0 p Ov Ov ÜJ X H co M < o O vO ! —I 02 O O 02 O CO rH O 0 0 K > I> O c- «0 H- 74 W vO vO to vO K) VÛ OZT rH c O CD O o C\2 O O to O o to to C V 2 O C V J O to o o O rH ■ Ü N £ > ■ O C\2 0 0 Z o !< w ce cr O o cv o o to t o o o to to O i 00 l * > _J ô H c v î *r\ O C V 2 e o 0 0 O 00 vO OO *r\ üJ I H C V 2 On W W vO O « 0 O O C V J w « o vO O O K5 O 0 » O O o OO 00 ON «e OMIiLTO 75 os cv Os 02 00 02 02 ro m OZT « o 0 0 o 00 02 O rH O rH O 02 02 rH 00 o cO lO 02 00 00 e o >o rH 00 02 O 02 rH 00 00 00 n LJ O O rH O n O » o 00 o tH e o 00 to 02 02 00 O H CO < r \ I —I 02 p CO »o O O o o o o o oc I —I K > (O OO «e i * î •OMIJiTa 76 o o vq rH vO vO 1 T \ Ç V i fO VÛ ro » o 021 o rH «fi CR O o o M H ÛC < I O rH t x > aO C\2 ro rH rH rH « O ■O C\î 02 Z O !< -I UJ a : cr o o o to o o o m O M to to O l O 6 I —I 02 02 to rH 02 e o O 00 rH 02 UJ X H o> « to <r\ rH 02 &0\ vO CO •o O U V . vO O 02 02 O < t > OO BO 77 o CO I —I t r \ cv CO to to m OZT o C V i I ---I H O o CD cv C V i o ( T V o CO C V i H c r < X o < r \ o oO I—I to C V i 0 0 C V i o c^- co z o _1 L aJ c r cr o o o o vO o v r \ rH CVi CO C V i cv ( • > rH C V i I— I o t o O Ô o CO C V i CO 0 0 L aJ X h" C V i CO CD o CVi CO CO CO o o o o o o < n CO CO o N ) CO ON 0 » c ~ OO • o n 78 N vO to cv vD o\ to cv to ro 021 c O 00 e o M VN v r \ H (T < X O T o VT\ \o to o o lO o o z O ! < _j ÜJ c r c r O o cv o cv I —f cv o > 9 -- T o\ O f O O O cv cv o to to _J 6 M e o 0 0 to O I —I UJ X H "P to I —I o 0 0 «O o >0 o > o K > OO OO n 79 -j? vsl 1 O CV C r \ * r \ ' rH to en cv to I I é * ■ c Vsl VÛ VÛ -\l ->> X N u %: tvi cv i vû VÛ 1 i l ■? N cv f - l VÛ --cH_ O CV % t+_ . t 5 0 CV VÛ rH <0 rH o cv CV Nv -Il c^: VSJ * -t rH VÛ rH «0 rH 1 o 1 —1 1 + lO + à- + K3 + (4 4* fH 4* O eH 1 N 1 to 1 j- 1 m 1 vO 1 C n- VÛ H ’ ' cv to rH i r v oz, - N 4- KI H ----—— — C^ 1 ON A O B m è @ r - r~ 1 ? ? r & 1 rH 4 C^ o O VÛ 1 VÛ UN\T) g : * • N s s 5 o f a ° f O H- O o O O VÛ 1 VÛ C N i% k S b 5 * o la 1 rH s f e 1 ? f r rH lO 4- t r \ * C N CV O O rH VÛ 1 O VÛ % î i rH rH cv 3K ^ t 00 1 A T ? f ? ±- 4 - to CV cv rH CV rH VÛ 1 —1 0> « T N 1 ! > - % b 5 « • O > 0 CV n o cv 1 0 rH f f ( Q r b f V\ to 4- tn rH rH v\ - < h VÛ CV rH v> 1 Cf\m s « a < » 1 —1 3 » . cv n c r \ o A 1 > » t o 1 ? A r VÛ oî 4- CV 1 —1cv to O v> A iT\0 l a m . C ° cv C r \ rH T 1 T d~ 1 V f VÛ rH 4- vO VÛ CV rH ■ > i “ H 1 â 3 o o O CV CV O O O O ..- 1 I T \ s » ! ? » ? f ^ rH H rH c ) l O H- « > > Û - . t f4 1 1 - 4 - rH en -4 1 (\ 1 —1 ~ < J f ) T ? e o 0 1 o ü . O 4 1 C r \ VÛ t ) rH zK s O 00 a a O H r< 1 O cv 1 o - 4 - VÛ rH o 1 On C^tO ? ÿ ? » . r cv f K > r- c r \ m o » a 2 2 VÛ rH to 1 to 1 :î H ON ~ 4 - cv C V ' 1 VÛ a 1 ? s 1 ? o y * CV 1 H O sh A 1 2 a ü. a en 1 cv rH to ' . t O to ---* r \ - ' < T \ 1 ? ; 5 ° f c r \ 1 —1 w e 2 o N 3 % w\ »o 1 1 m cv VN i f \VN c r 1 C c r g ? ? H ? H f O < a « H- r * % A CV >0 1 cv 1 —! 1 cv c^ O to H On ■ ■ CV 1 f ; ? r r rH t * oo r» S a : l - CV I / - < h rH 1 to ON O ~4 1 —f ----- 6V- s 1 f ? 5 , 1 , ? eo 1 oo > a w i ? 00 1 ' * r \ - 4 - c r \ CV rH X X H- a X <4- -f — m
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The development of a performance rating scale for the evaluation of shipboard performance of enlisted naval personnel
PDF
The relative value of fourteen test variables for predicting success in executive and supervisory positions
PDF
Prediction of temperamental traits from aptitude test scores
PDF
Comparison of the Wherry-Doolittle and successive cutting-score methods in the selection of tests in a clerical aptitude battery
PDF
A correlational study of the efficiency of certain aptitude tests in predicting the success of supervisors in an aircraft factory
PDF
A study of the role which several factors play in determining the attitudes of judges toward internal organs and reliability of these attitudes as expressed on a rating scale
PDF
Toward the development of a test to predict maintenance and termination of love relationships
PDF
The construction and validation of an optometric aptitude test
PDF
Computer-analyzed EEG findings predictive of schizophrenia
PDF
A validity study of a test battery for food market checkstand operators
PDF
A compilation of norms on several tests of learning
PDF
A study of scores designed to measure intra-trait variability on a test of temperament
PDF
The detection and measurement of faking on personality inventories
PDF
The use of biographical data in predicting success and duration in psychotherapy
PDF
A validation study of some factor tests of verbal fluency
PDF
Number of training trails and variations in stimulus cues as factors effecting extinction in children
PDF
The validation of an orientation test
PDF
An integration of perceptual and behavioral variables in martial adjustment
PDF
Effects of focus of attention on the causal attributions of actors and observers
PDF
A study of stimulus response times in the localization of sound
Asset Metadata
Creator
Mackie, Robert Ramsay
(author)
Core Title
Norms and validity of sixteen test variables for predicting success of foremen
Degree
Master of Arts
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,Psychology
Format
application/pdf
(imt)
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c39-42088
Unique identifier
UC11310472
Identifier
EP63953.pdf (filename),usctheses-c39-42088 (legacy record id)
Legacy Identifier
EP63953.pdf
Dmrecord
42088
Document Type
Thesis
Format
application/pdf (imt)
Rights
Mackie, Robert Ramsay
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA