THREE ATTITUDE SCALES IN RELATION TO THE ACADEMIC ACHIEVEMENTS OF MALE COLLEGE FRESHMEN OF MODERATE ACADEMIC POTENTIAL Thcsls Ior Ike Degree oI pk. D. MICHIGAN STATE UNIVERSITY Walter Russell Stevens Jr. 1958 This is to certify that the thesis entitled THREE. ATTITUDE SCALES IN RELATION TO THE ACADEMIC ACHIEVEIENTS OF MAIE- COLLEGE FRESHIEIN OF MODERATE ACADEMIC POTENTIAL presented by Walter Russell Stevens Jr. has been accepted towards fulfillment of the requirements for _Bh4D_._ degree in _P_sy.c.h.ology Major professor 0-169 1. I B R A R Y Michigan State University _ ___ ___,_,_..-. ——_-....-—..__. . THREE ATTITUDE SCALES IN RELATION TO THE ACADEMIC ACHIEVEMENTS 0F MATE COLLEGE FRESHIIIEN OF MODERATE ACADEMIC POTENTIAL By WALTER RUSSELL STEVENS JR. A THESIS Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1958 ABSTRACT Stevens, Walter Russell Jr. Ph. U., Michigan State University, August, 1958. Three Attitude Scales in Relation to the Academic Achieve- ments 2; Male Collggg Freshmen g£IModerate AcademIE Potential. Major Professor: Albert I. Rabin. Three five-item attitude scales were empirically derived from the responses made to 135 rating-scales by 63 unmarried, white, male college freshmen, students who had diSplayed moderate academic aptitude on the Michigan State University Freshman Orientation Test Battery (the MSU English Placement Test, the MSU Arithmetic Proficiency Test, the "Quanti- tative" and "Linguistic" scores of the ACE Psychological Examination?- l9h0 edition, and the “Vocabulary" and "Comprehension" sections of the LSU Reading Test). All the rating—scales had been selected by the writer and three other counseling psychologists as logically representative of universally-eXperienced aSpects of college life. Each of the scales met criteria for unidimensionality, and scalability was largely confirmed in cross-validation. The attitude scales were labelled as representing: (a) Institur tional Identification, the extent to which a student considers the acar demic enterprise to be compatible with his own needs and aspirations; (b) Self-Confidence, the degree to which the student feels comfortable-- even stimulatedr-in his performance of the student role, particularly of its more public aspects; and (c) Achievement Valuation, the extent to which working for recOgnition and enjoying positions of prestige are cone sidered worthwhile by the student. Comparisons were made between students' scores (scale types) on the attitude scales and their cumulative grade point averages (CGPA) for three terms of the freshman.year (with a.minimum of AZ course credits carried and.a minimum.of 39 credits earned). No statistically significant relations (coefficient of contingency) were found between the attitudes ii ._ -__.. and achievement, either when the attitude scores were combined in various three-scale or two—scale profiles or when they were individually compared with the achievement criterion. The small size of the sample and the unreliability of the achievement criterion probably contributed to the essentially negative results. The strongest relationship was that between low (scale types 0 and 1), moderate (scale types 2 and 3): and high (scale types A and 5) Self— Confidence Scale (808) levels and low (below 2.15), moderate (2.15-2.5h), and high (above 2.5h) CGPA levels. The obtained .325 contingency coef— ficient was significant at the .12 level of confidence. High achievement was associated with moderate self—confidence; moderate achievement, with low self-confidence; and low achievement, with high self—confidence. Knowledge of 505 levels increased efficiency of forecasting CGPA levels by 30.h per cent beyond that obtained without such knowledge. Although the statistic may not have been fully justified, a curvilinear regression of CGPA on 508 scale types was cmnputed and reported (§3§=.h73, signifi- cantly greater than zero at the .01 level). When the distribution of three-scale attitude patterns was super- imposed on the low, moderate, high three-by—three contingency table com- paring 863 level and achievement level, a number of interesting trends appeared. The trends were not worth testing statistically because of small sample size, but may be worth pursuing with further research. The observed pattern trends failed to support the widely-held contention that Optimally-achieving students (high-achievers) are inclined toward exag- gerated identifications with authority—figures. NOSt students with atti- tude patterns commensurate with such identification were low-achievers. Conversely, among the high—achieving students the majority exhibited pat- terns essentially contradictory to such identification. The only stu- dents with patterns in keeping with the "identification" thesis who were iii found achieving above the lowest level were those with 808 scale types 0 or 1. It was suggested that this last group of students is perhaps more typical of counseling center and psychological clinic clientele than of students in general, thus possibly accounting for the inclination of many clinically~trained personnel to predict high achievement for most students with strong tendencies toward identification with authority-figures. An attempt was made to eXplain the trends in terms of the concepts of "feelings of personal inadequacy," "defensiveness" v. "tolerance," and "premature identification" v. "independence and flexibility." Optimal achievement for the students of the sample was viewed as symptomatic of an essentially realistic awareness of and tolerance for self and surround- ings, with no particular concern for the conventional signs and symbols of prestige per se. Failure to achieve was seen as a concomitant of: (a) premature narrowing of interests (relative to the demands of first year liberal arts curricular requirements); (b) a straining after relar tively superficial appearances of academic success as compensation for felt but poorly tolerated personal inadequacies; and/or (0) genuine satis- faction with the personal status quo such that no need to strive is felt. Further research was recommended to test the very tentatively-held and essentially ES hoc inferences of the present study. Approved W J Signature of major Professor Date 45$45€;(//<§;2?> /h_fi // iv ACKNOWLEDGMENTS I wish to express my sincere thanks to the members of my guidance committee, M. Ray Denny, Alfred G. Dietze, and Donald M. {phnson, and my particular appreciation to the committee chairman, Albert 1. Rabin. Their contributions extend far beyond the mere mechanics of committee functions. Individually and collectively they have provided very much of the ideational base upon which the study was founded, and have helped to determine the intellectual and emotional climate in which it was brought to fruition. Their patience and support have been truly sustains ing in some of the more trying moments of the total experience. To Harold Dahnke of the Michigan State University Board of Examiners special thanks are due for his invaluable help in screening the 1955 male freshman pOpulation for the homogeneous sample of the study; and I am sure the final instrument received a sound beginning through the generous contributions of time and professional judgment so freely given by my former colleagues at the MSU Counseling Center: Gwen Norrell, Harry Grater, and Paul King. Cross-validation of the attitude scales was made possible by the contributions of class time by Roberta Koons and John Nagle at MSU and by Lark Daniel, Jeseph Sidowski, Wolcott Treat, and Merle Turner at San Diego State College. Without the kind assistance of all these friends and colleagues, the research could not have been done. Finally, I am deepLy indebted to the many students whose willing participation in the study was the most direct basis for anything the ensuing pages may contribute to the understanding of relations between attitudes and academic achievement. CONTENTS LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . . . . . . . IAIST OF TABIJE o O O a o o o c o o o o o o o O I o O I o o I 0 Chapter I. REVIEW OF THE LITERATURE AND STATEMENT OF THE PROBLEM . General Assessment of Past Research: A Confused Liter- t we...............'....'.... VagueObjectiveS............o..o.. Empiricism and Evaluative Biases . . . . . . . . . . 9 _ Self-Report Techniques: The Most Widely Used Approaches. Representative Findings - o o o o o o o n o o o o o o A Problem Peculiar to the Method: Stereotyped Responses 0 O O O O O 0 I O O 0 O O O O O I O O O O Criterion Problems: The Cumulative Grade Point Average Variability of Student Perfonnances . . . . . . . Variability of Grading Practices . . . . . . . . . Scale Characteristics of Letter Grades . . . . . . Orientation to the Present Study: Attitudes and A Chi evem ent O I O D O C O O I O O O I O I I O I I O Attitudes Toward Authorities: Identification . . Attitudes Toward Self . . . . . . . . . . . . . . Self-confidence . . . . . . . . . . . . . . . . Self-satisfaction v. Need for Achievement . . . Review Summary and Specific Hypotheses . . . . . . Aims C O D O O C O O O O O O I 0 I O O O O O O I O 0 vi Page encounter-4 I: 13 13 l6 17 18 19 19 20 2h 25 CONTENTS, continued Introduction . . . . . . . . . . . Problems of Type of Item . . . . . e Determination of Content Areas . . The Student Opinion Survey . . . . Subjects . . . . . . . . . . . . . . Cross-validation Sample . . . . . . . Administration of the Student Opinion III . MASULTS o 0 0 h I o 0 C c o o o I O o 0 Item Selection . . . . . . . . . . Scoring and Scale Analysis . . . . cross-Validation o o o a o o o o o Naming the Scales . . . . . . . . Achievement o o o o o o o o o u o o Vil a Survey Derivation of the Three Attitude Scales Construction of the Student Opinion Survey Scale Analysis as an Approach to Homogeneity The Riley Revisions of Scaling Techniques II. METHOD AND PROCEDURE . . . . . . . . . . . . . . The Theory of Scale Analysis . . . . . . . O C Q I Attitude Patterns and Academic Achievement General Considerations . . . . . . . . . Each Attitude Scale Compared with Academic O O O O Attitude Levels and Achievement Levels Attitudes and Achievement . . . . . . . swam o O O o o o o o O O o O 0 O 0 o Self-confidence and Achievement . . . . . Applications of the Attitude Scales to the Problem of Academic AChievement . . . . . . . . . . . . . . . . Problems of Attitude Pattern Designation . . . . . Levels: 0 0 O O D O 0 O 27 27 27 3h 39 39 ho A2 145 h? 51 53 SS 55 SS 56 62 614 6h 68 68 69 70 71 E IV. V. Appen A. CONTENTS , continued Introduction . . . . . . . . . . . . . . . . . . . Self—Confidence Scale Levels and Achievement Levels U 0 Q C C O C O O O I O I O O O O C I C O . Self-Confidence Scale Levels, Three—Scale Attitude Patterns, and Achievement Levels . . . . . . . . . swam- o O o 0 o b o I o 0 t o o o 0 O I O o O C 0 DISCUSSION AND SUGGESTIONS FOR FURTHER RESEARCH . . . . . . Implications of Present Failure to Demonstrate Statis- tically Significant Attitude-Achievement Relations . . . Attempts to Rationalize the Trends of the Study: Feelings of Inadequacy, Institutional Identification, and Academic Achievement . . . . . . . . . . . . . . . . Suggestions for Further Research . . . . . . . . . . . . SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . SHIECTED WNGES o a t I I o I o o o o 0 o o n c o o o O O o O O dixes Fifty Polar Adjective and Present-participle combinations . The student Opinion survey. 0 0 I O I O I 0 O O O O I O C O 0 IBM FORM I. T. S. 1100 A 151 . . . . . . . . . . . . . . . . Fonm Letter Used to Invite Selected Students to Participate in the Stuw O O O O C V O I O Q 0 O I I O Q U C O I U . 9 Face Sheet for Subjects of the Select, Homogeneous Sample . Face Sheet for Cross-validation Subjects; MSU Students . . . Face Sheet for CrOSs-validation Subjects; SDSC Students . . Terms Which Defined the a Priori "Positive" Extreme for Each Item of the Student Opinion Survey, as Detennined by the Judge 5 I O O O C O O O O O O O 0 O O O O O O O O O O 0 Items Approximating Desired Scale Distributions in Original Forty-five-item Matrices; Also Data on Which Final Item Selection Was Based . . . . . . . . . . . . . . . . . . . . Final Item-Score Matrix (5 Dichotomized Items); Attitude. viii 71 71 72 7h 76 76 78 82 85 88 93 9A 101; 105 106 107 108 109 110 1. i" K. GONTEN'I'S , continued Area I O C C O C O O C C O Q O O O 0 O O O I O O O O O O O 112 Final Item—Score Matrix (5 Dichotomized Items); Attitude Aran 0 o O O 0 o I s a o c O t Final Item-Score Matrix (5 Dichotomized Items); Attitude AI‘ ea 2 O O D O O O I O O I C O O O D 0 . C O O C C O . C . 116 Data Summary: Achievement, Intellective, and Attitude Id ea sure S O I O O O O I O I O 0 I O O O O I O C I O O O O O 118 Data Summary: Time Spent in Outside flork, Field of Specialization, and Credits Carried . . . . . . . . . . . . 120 LIST OF ILLUSTRATIONS Figure Page 1. Item-Score Matrix for a Perfectly Reproducible Scale, in Which Each Item Is Scored l or 0 and the Total Score Is the Sum of the Item Scores . . . . . . . . . . . . . . . . . 28 2. Examples of the Three Principal Types of Error Distribu- tions in Predicting Item Responses from Subjects' Ranks . . 32 \ \ 3. Derivation of All Possible "Unique Scores” and "Scale Types" from Weighted Scoring of Responses to a Five-item Scale; All Items Diohotomized . . . . . . . . . . . . . . . . . . . 36 \ LIST OF TABLES Table A Page 1. Frequency Distribution of Cumulative Grade Point Averages Earned by Male Freshmen; One Year's Performance . . . . . . 52 2. Dichotomized Items Comprising Three Scales, Each Represent— ing an Area of Student Attitude . . . . . . . . . . . . . . 58 3. Summary of the Extent to Which the Three Attitude Scales | Meet Criteria for Scalability (Unidhmensionality) . . . . . 59 | I h. Summary of the Extent to Which the Three Attitude Scales Meet Criteria for Scalability (Cross-validation Sample) . . 60 5. Contingency Coefficients Representing Relationships Between Pairs of the Three Attitude Scales . . . . . . . . . . . . . 66 6. Comparison of Eight Basic Three-Scale Attitude Patterns with Three Levels of Academic Achievement . . . . . . . . . 67 7. Composite Table for Comparisons of Each of the Attitude Scales with Academic Achievement Levels (Entries for Each Attitude Scale to be Considered Independently of the Others. 68 Scales with Academic Achievement Levels . . . . . . . . . . 69 9. Relations Between Each of the Attitude Scales and Academic Adlievelnent O O O O O O O O O O U 0 O O O O O O O O O O 0 Q 70 10. Frequencies of Occurrence of the Eight Basic Three—Scale Attitude Patterns (IIS-SCS-AVS) When Levels on the Self— Confidence Scale Are Compared with Academic Achievement 8. Statistical Summary of Comparisons of Each of the Attitude Levels (Table 6 Superimposed on the SCS Portion of Table 7). 73 Chapter I REVIEfi OF THE LITERATURE AND STATEMENT OF THE PROBLEM General Assessment of Past Research: A Confused.Literature A problem which has commanded increasing attention and research efforts of educators and psychologists over the past three or four dec- ades is that of identifying so-called "non-intellective" factors which are purported to relate to academic achievement. The search has come in response to the persistent failure of intelligence or achievement tests to account for*more than one-quarter to one-third.of the variance Observed in college students' academic performances. Published attempts to find.non-intellective correlates of academic achievement are legion, many of them superficially repetitious of earlier works, some of them compellingly suggestive in their implications, and a distressing mass of them weakly conceived, awkwardly deve10ped, and abortively brought forth to add little more than their number to the scene. In 1931 (31) and again in 19h0 (32), Harris made eXhaustive sur- veys of all the available literature pertaining to the quest for none intellective correlates of academic achievement, and in 1950 Lord (L2) performed a similar service. At least in part the widespread confusion in this research area is traceable to ignorance or lack of methodological rigor. In his later review Harris (32) bemoaned the continuing widespread failure of researchers to profit from the errors of the early studies. The most serious shortcomings have been failures to control for intelligence or acadenic aptitude, failures to select sufficiently homogeneous samples, and failures to report findings in terms of apprOpriate statistics. Unfortunately, Harris' objections are applicable to most of the research published since 1910. It is virtually impossible to draw from the mass of studies unequi- vocal conclusions concerning relations between non-intellective aspects of personality and academic achievement. The bulk of the relevant liter- ature represents a hodgepodge of hunches, hypotheses, and plain and fancy curiosities, pursued or attacked with all manner of tools and tech- niques, variously applied to a great range of pOpulations, analyzed and interpreted in a facinating—not infrequently bizarre--variety of ways, and only too often generalized to apply to all students everywhere. A majority of published studies have used small samples, most num- bering less than one hundred subjects, and many with fewer than twenty- five. An appallingly small minority of researchers appear to have been even faintly cognizant of the necessity for cross-validating empirical findings, and this reviewer has discovered only one study, by Cronbach (1h) , which represents a careful attempt to repeat an earlier study, the well~known Monroe Rorschach study (148). In this instance the statisti- Cally significant results of the earlier study were not substantiated. As Cronbach himself pointed out (1h), many studies using so complex an instrument as the Rorschach yield "siglificant" results by virtue (or Viee 3) of inapprOpriate statistical reasoning (cf. 149 and 69). Like chrLbach's study (11;), that of ItcCandless (A3, M) was carefully con— cei‘red and executed and disclosed no Rorschach indices capable of differ- entiating between high— and low-achievers among Maritime Service officer can(-‘lidates, all of whom were of very superior acadenic potential. \ 3; A ; Vague Objectives Often researchers appear to have had only the moat general of objec- tives: finding some basis for differentiating between high- and low- achievers. A common approach has been the firing of a "shotgun blast" at a readily available student sample, apparently in the hOpe that some vulnerable spot might be hit. Representative of such studies is that of Burgess (10), in which )40 engineering students were bombarded with the Rorschach test, the Thematic Apperception Test (TAT), the Minnesota Multiphasic Personality Inventory (WI) , the Rosenzweig Picture-Frustra- tion (P-F) Study, the Strong Vocational Interest Blank, the Bernreuter Personality Inventory, and the Borow College Inventory of Academic AC3~.‘]‘I:lstment. Half of Burgess' subjects were "under—achievers" and half were "over-achievers," the categories being defined with reference to discrepancies between actual academic performances and levels of achieve- ment predicted from several intellective indexes. Many kinds of "scores" were obtained for each subject: frequencies, sums of various weighted frequencies, ordinal rankings, ratios of various kinds, etc.; and then means were cOmputed for each type of "score" for each of the two samples. One hundred fifty-one t tests (for testing the significance of the dif- ferences between the means) were made, and 11 of then were found to be "significant," five at the .05 level, and six at the .01 level. Sad to Say, the most "significant" results appeared where the statistic used was least- apprOpriate. Empiricism and Evaluative Biases Very few researchers, with the notable exceptions of Ryan (59) and Klugl'l (39), seemed even faintly aware of the theoretical limitations they assumed when they employed devices which were themselves empirically h standardized on other populations. For example, one would be hard put to provide a logically sound defense of Morgan's contention that low— achievers are perhaps "insensitive, callous, self-centered, and irre- sponsible" (50, p. 295) because they score significantly higher than do high-achievers on the :3 scale of the MIPI. Such an assertion betrays the false logic of "guilt by association“: low-achieving students and "psychOpathic deviates" (the P_d of §_d_ scale) tend to respond similarly to some of the same items on the MPI; "psychOpathic deviates" tend to be “insensitive, callous, self-centered, and irresponsible," therefore, low-achieving students must also tend to be "insensitive, callous, self-centered, and irresponsible." (Besides, Morgan himself adds, "The greater percentage of nonachievers with profile elevations on Pd is partly a reflection of an absence of higher scores on the neurotic scales Which were more often obtained by the achievers." (50, p. 295) In other words, Morgan's definition of "high" on any of the MMPI scales was a relative one within each individual's profile rather than an absolute one relative to the normative standard score scale of the test itself.) Similarly, the meanings of results obtained with another empirically derived test, the Strong Vocational Interest Blank, must remain highly Speculative. Both Morgan (50) and Ryan (59) found that high-achievers earned higher scores on the Group V--Social Service—~00cupations than did low-achievers, and although Morgan made no attempt to eacplain the fin¢'Lng, Ryan suggested that its significance might be that it reflects a tendency toward conformity or identification with authority figures! How such a conclusion was reached is not clear, but it is perhaps worth nOting that the general evaluative tone accompanying Ryan's discussion is not particularly favorable to the achievers. The evaluative bias against "over-achievers" is carried forward in another study (58), this one by Rust and Ryan. The instrument employed l—A was the Harrower—Erickson group method Rorschach, scored according to K10pfer's method, and their subjects were Yale College juniors and sen- iors, excluding engineers. The findings are summarized in the following quotation: An admittedly highly speculative portrait of the over- achiever begins to emerge. He is overconventional or a con- foxmist (high P); he is practical—minded, tending to "see what's there" and exhibiting "stereotypy in thinking" (high A per cent); he shows little "introversion or self-preoccu- pation" (low M) and is probably emotionally immature (Fm greater than M . (58, pp. 2514-255) Another pair of investigators (52) used the same Rorschach method With a somewhat different sample of college students. Their "non- achievers" were students on academic probation at the University of Georgia, while their "achievers" were psychology students who had never been on academic probation. Osborne and Sanders noted that the two groups Were not even equated for "intelligence or previous educational exPerienc:e" and that the achiever group was "somewhat" older than the non-achievers. (52, p. 685) A fairly long quotation of the findings of this Study follows, for the variations between it and the previous quota— tion dramatically illustrate the sort of chaos which characterizes much or the Literature, especially where the results of well—known tests nses to individual items; thus, all subjects earning a score of five on the test represented by Fig. 1 would have responded positively to items d, e, f, g, and h and negatively to items a, b, and c. This amouIlts to stating that "persons who answer a given question favorably all haVe higher ranks on the scale than persons who answer the same ques- ti‘m unfavorably." (66, p. 9) 27 sub jects h Total score Hmm'fimbomb‘ OOOOOOOOH ooooooor—‘t—l OOOOOOHI—JH OOOOOHHI—‘H OOOOHI—‘Hi—‘H OOOHHHHHH OOHI—IHHHHH OHHI—IHHHHH OHNwt‘mo-qoo Fig. 1. Item-score matrix for a perfectly reproducible scale, in which each item is scored 1 or O and the total score is the sum of the item scores. In practice, the ideal is seldom if ever attained. For example, some subjects with a score of five on the scale represented by Fig. 1 might have responded positively to one or more of items a, b, or c and negatively to one or more of items d, e, f, g, or h, leading to erroneous predict-.1011 of item responses from total score ranks. Guttman (25, 26, 27) has prOposed a 10 per cent error of reproducibility as the maximum allowable if a set of items is to be considered representative of a scalable (i. e., unidimensional or homogeneous) universe of items. That is, if a set of items indeed represent a range of affirmation (as in attitude Scaling) or of difficulty (as in ability testing) within a single common factor, errors in reproduction of item responses from subjects' ranks Should result from chance factors and should not exceed 10 per cent of the total number of item response predictions called for. The extent to ' . which this criterion is approached is given by the coefficient of 29 reproducibility (27, p. 117): Coefficient of _ 1 _ number of errors Reproducibility m A 10 per cent error in reproduction of item responses from total score ranks results in a .90 coefficient of reproducibility. Lower coef- ficients identify sets of items which are not unidimensional; i. e., scales which represent more than one factor. Reproducibility is but the most easily quantified of four criteria prOposed by Guttman (25, 26, 27) for appraising the unidimensionality of a. set of items; another is that of the number of reSponse categories, combined with what falls under the heading of "item difficulty" in abil- ity testing. That is, greater confidence can be placed in a given coef- ficient of reproducibility obtained with items for which there are at least three possible response categories (e. g., responses weighted 2, 1: and 0) than in the same coefficient obtained with dichotomous items (Weighted l and O). In Guttman's language, The more reSponse categories for items included in a scale, the greater is the assurance that the entire universe is scalable. . . . For example, four dichotomous items with high reproducibility do not provide as dependable an infer— ence concerning the scalability of an area as would four trichotomous items which were equally as reproducible. It is especially important to keep as many reaponse cate ories as possible when the total number of items is small. 26, pp. 79—80 The primary reason for this recommendation is the fact that the "rePrOducibility of any individual item can never be less than the per- centage of respondents falling into a single answer category of that item, regardless of whether or not a scale exists" (26, p. 78), and com- bining Scale categories often produces items which do not discriminate among r e Spondents. If combining response categories produces items which have been responded to affirmatively (or negatively) by as many as 90 per 30 canni; of subjects, a spuriously high (at least .90) coefficient of repro- diicsinility is automatically obtained. In reality, all that has been axzczcnnplished in such an instance is to compose a scale of items about which there is little or no disagreement among respondents, some items being almost universally accepted and some being almost universally rejected. Thus, it is necessary that as many scale categories as possible be retained, or if dichotomization of item responses must be resorted to, i that "attanpts . . . be made to include in the sample [of items] as wide .5 a range of marginal distributions as possible, and specifically to attempt 15<> :include items with marginals around 50-50." (26, p. 78) This is equiv— EfiLenit to stating that item "difficulties“ (borrowing from the terminology (If' ability testing) should cover a wide range and should include a fairly sizeable preportion of items which are answered affirmatively ("correctly" ill ability testing) by approximately 50 per cent of the reSpondents, and negatively ("incorrectly") by the other 50 per cent. A third criterion for unidimensionality is presented as a recommen— clattion to be followed if the obtained coefficient of reproducibility is to be accorded confidence. The advice is that, unless it is possible to iJiclude among the items several whose marginal distributions "are in the ITange of 30 per cent to 70 per cent“ (26, p. 79), it is necessary to use at least 10 items to represent the scale being deve10ped. Riley, 23.21- (56) have found as few as five or six dichotomized . items to be useful for classifying respondents with respect to broad attitudes, and contend that more precise classification is usually not warranted by the nature of the attitude being studied; e. g., it is unlikely that "attitudes toward academic authorities" is subject to meaningful analysis into more than five or six gross strata. Guttman (26) further states, An important property of a scalable universe is that the order- ing of persons based on a sample of items will be essentially the same as that based on the universe [of items from which the items in question are assumed to constitute a sample]. If the universe is scalable, the addition of further items merely breaks up each type [i. e., rank] given by the sample into more differentiated types. But it would not interchange the order of the types already in the sample. (p. 81) Therefore, provided the marginal frequencies of the scale-limiting ijaeanls fall generally within the 70 per cent - 30 per cent range, with a rrtnnflaer of the intervening items approxhnating the 50-50 Splitting of the sanple of respondents, an obtained coefficient of reproducibility of .90 CI!” lrigher leads to a reasonably confident conclusion that the items 1?€31>resent a scalable universe of similar items. Even should the coefficient of reproducibility be less than .90 (5.. e., in the .805), if the pattern of errors of prediction of item 1?EBSponses from total score rank diSplays a gradual interlacing or overlap (see Fig. 2), it may be inferred that the sample of items in question 3?€npresents a quasi scale rather than a truly scalable universe. A quasi iscale represents a principal factor, plus a number of lesser factors Vflaich have introduced error variance-—beyond that attributable to chance ££lone—-into the attempt to order subjects according to their ranks along the scale continuum. The absence of such a pattern of error in prediction of item responses stands as the fourth criterion for unidimensionality. Should such a pattern of error obtain in conjunction with an acceptable coefficient of reproducibility, it discloses the presence of a useful-- albeit less than unidimensional-—scale. In fact, Guttman maintains that, The importance of a quasi scale lies in how it is used for exter- nal prediction problems. . . . The prediction of the external variable rests essentially on the dominant factor that is being measured by the quasi—scale scores. Thus a quasi scale has the full mathematical advantages of a scalable area. (28, p. 162) Should it be found, however, that a set of items defines a quasi Item A Item B Item 0 Subject‘s rank order (random scale (grouped nonscale (gradient quasi— (based on errors) errors scale errors) total score) 32 1: Scores Scores Scores 0 O 0 A Fer—v; ea:- NNNNNN N \ooaw Oxmu—w to H N 10 NNN NNN KNNNNM N NM NNNN H U1 NHH NNNWNN HKNNNNM NNH NHNNH KN R E" N NRNNNHNHNNMN KN U'l NRNNK w NRNNNNNNNHN H N Errors : \A) Fig. 2. Examples of the three principal types of error dis- tributions in predicting item responses from subjects‘ ranks. (Of course many more items would be needed to determine so many ranks.) Each X represents a respondent's reply to an item; an affirmative reply being represented by an X in the "1" column; a negative reply, b an X in the "0" column. (Adapted from Guttman [28, p. 1611.), ‘ \ \ Scale rather than a truly unidimensional scale, it would be necessary to inelude a larger number of items for stable differentiation among sub- ‘ \ jeCtS. Guttman (28) states it as follows: While the single darninant variable of a quasi scale cannot be represented by means of a small number of items due to the amount of error involved, increasing the number of items which contain this dominant variable makes this error assume a gra— dient pattern, and permits an invariant rank order [of respond— ents]. (p. 163) In Guttman‘s Cornell technique of scalogram analysis (25), subjects are first ranked according to their total scores on the attitude ques— tionnaire being used, and their responses to each item of the question- naire are then recorded as in Fig. 2, above. Visual inspection of the resulting chart permits location of the point in the distribution of responses to each item at which it would be necessary to alter prediction 0f subjects‘ responses to that item, based on their total score ranks. In Fig. 2 these points are identified by horizontal lines in among the x3- They must be so located as to minimize errors of item response pre- diction, but as Edwards (17) and Edwards and Kilpatrick (18) point out, the cutting—points must fall between ranks and not within a set of rank ties as Gut’tman (25) has been inclined to place them. In Fig. 2, "1" responses would be earpected on Item A frOm all sub- jects whose ranks among the 30 reSpondents are 1).; or higher; and ”0" re Sponses, from all subjects whose ranks are 15 or lower. This predic- tion results in three errors (10 per cent of the 30 predictions for the itfin): a "l" re3ponse by a subject with the rank of 18, and "Os" by suba‘ects with ranks of 7 and 11. It will be recalled that predictive erI‘or of this magnitude is the maximum allowed by Guttman's approach to Scgale homogeneity if a given set of items is to be considered to repre- Sent a scalable (unidimensional) universe. The errors in prediction for the other two items of Fig. 2 are greater than the 10 per cent allowable, but those for Item 0 exhibit a gra~C1ient pattern of error similar to that of the scalable item; i. 6., there is no clearly definable clustering of errors, as there is in the 7 3h reaponses to Item B. A set of items including some with error patterns like that of Item c would thus compose a quasi scale, which could be retained as a reasonably reliable basis for predicting an external crite- rion. The Riley Revisions of Scaling Techniques Riley, it. a}. (56) have improved Guttman's approach to scale analysis, largely by minimizing the subjective judgnents required and by maximizing quantification. Their method requires items with but two response cate- gories, or items whose response categories have been combined into dichot- omies by splitting the obtained item response distributions as near their medians as possible, scoring "0" all categories below the cutting-point (toward the "negative" end of the response continuum) and assigning a differential scoring to all categories above the cutting-point. The differentially weighted scoring of reSponses on the "positive" or "affirmative" side of the respective cutting-points assumes an ascend- ing geometric progression from the item receiving the highest preportion 0f positive response. In this method scales are usually limited to five or six items, and in a five-item scale the differential weights would be 1: 2, h, 8, and 16. In other words, with such dichotomous scoring, the item to which the highest prOportion of affirmative response is obtained Carries the scoring weight of l and earns 0 points for all subjects below 1 Point for all subjects above the division point. Similarly, the next mOS‘t affirmative item, with a scoring weight of 2, yields 0 or 2 points credit; the item weighted )4 yields 0 or )4; etc. Total scores, called "unique scores," are the sums of the weighted iten scores, but these are not the direct basis for ranking of subjects frem which the coefficient of reproducibility is to be computed. Obtain- the median division point in the obtained item response distribution and 35 ing unique scores is but the first step in the deriVation of the best possible ranking of subjects. The most efficient ordering of respondents -—the ordering which leads to the smallest possible error in reproduction of item responses from subjects' ranks—is that based on what both Riley, _e_t_ Q. and Guttman call "scale types." The scale type concept can per- haps best be defined by means of illustration. Figure 3 presents all the pOSsible patterns of item responses under the Riley differential weighting method, the corresponding unique scores, scale type assignments, and "error" designations for a scale having five dichotomized items. In Fig. 3 "errorless" scale types would be represented by subjects earning unique scores of 31, 15, 7, 3, l, and 0. Such respondents have not "erred" by giving a negative reply to an item with a high preportion °f POSitive responses and/or a positive reply to an item having a high Proportion of negative responses. Converting their unique scores to scale types, "pure," errorless scale assignments are obtained. Subjects emng unique scores other than these six are said to have "erred" in responding to one or two items of the five-item scale, and thus belong to "error" scale types. (Riley, e_t é. use the term "non-scale scores" interchangeably with the term "non-scale types," both terms referring to uni-Cine scores which yield error scale types.) For example, a subject earning a unique score of 29 is said to belong to scale .ype S on the premise that his negative response to the item carrying a weight of 2 was an "error" of omission, in view of his positive response to all the other items, particularly the three items which were responded to negatively by a higher preportion of the subjects. Similarly, a subject with a unique score of 20 is assigned to scale type 0 0n the premise that MS positive responses to the items weighted h and 163 after re5ponding negatively to items with lesser weights, are "errors" of commission. Scoring weights Unique Scale Errors N egative Positive score type (No.) 16 8 h 2 l X X X X X 31 S 0 x x x x ( ) 30 5 l x x x ( ) x 29 5 l x X x ( ) ( ) 28 S 2 x x ( g x x 27 S l x x ( x ( ) 26 5 2 (x (xg x 25 _1_ or 5 2 (x (x 21. o 2 (x x X x 23 2 or S l (X X X ( ) 22 2 or 5 2 x x ( ) X 21 g or s 2 EX (2:) 20 o 2 X X X 19 2 1 X X ( ) 18 2 or O 2 X X 17 1 1 (x 16 O 1 X X X X 15 )4 O X X X ( ) lb. h l x. x ( g x 13 h 1 X X ( ( ) 12 h or O 2 X X X . ll 2 or u 1 Ex 1: ( ) 10 g or o 2 X X 9 1 1 Ex 8 0 1 X X X 7 3 O x x ( ) 6 3 1 x ( ) X 5 3 or 1 1 (X) h o 1 X X 3 2 O X ( ) 2 g or 1 1 X 1 1 O O O 0 Fig. 3. Derivation of all possible "unique scores” and "scale types" from weighted scoring of responses to a five-item scale; all items dichotomized. Each X indicates a positive response to the item carrying the designated weight; blank Spaces indicate negative responses. "Errors" in item responses are indicated by parentheses. Perhaps it is evident that the Riley method implies a somewhat dif- ferent interpretation of "error" from that of Guttman. No longer is the emphasis on errors made by someone else in predicting subjects' item I‘eSponses from their ranks; rather, the implication is that scale errors are those the subjects themselves have made in responding to the items 37 of the scale. (However, whichever way errors are viewed, their number in. 3133* given scale is the same.) The latter "error" concept is made clearer by an analogy borrowed from Riley, gt 31.. (56): We might ask each of 100 men, "Are you at least 5' 1;" tall?" If each says "Yes," we can ask, "Are you at least 5' 5" tall - . . 5' 6" tall . . . ," etc. Eventually, each should say "No." The tallest man is the one who continues to say "Yes" the great— est number of times, provided of course that the men know how tall they are (and that none has a sense of humor). But it is in an identical sense that we try to find which peOple take the mOSt extreme response to attitude questions. Although we never know the "unit" difference from one question to the next in atti- tude research, we feel that the rank order into which people fall is meaningful and that such a measure is potentially useful. This analog gives us another chance to see what is meant by "scale error." If a man checks "No" to the question asking tmether he is at least 5‘ )4", but then proceeds to check "Yes" to greater heights until he reaches 5' 9", then he made an error somewhere. In Guttman‘s approach, the first reply of "No" would be counted as the error and the ”Yes" to the others would be accepted as correct responses. The man would be classified as S ‘ 8”, with one response error. The alternative is to say he is less than 5' LL", and that the four IrYes" replies after the “No" were all errors. In counting errors and classifying peOple, Guttman always assigns a person to the class which minimizes the errors. The "error" just discussed could have arisen from careless- ness on the part of the respondent, misunderstanding of the task at hand, misinterpretation of the question, etc. This is the problem of question unreliability. But scale error can arise in other ways. If, after the first question about height, we inter— Polate a question which reads, “Are you married, " or "Do you have red hair," it is obvious that the "Yes" or ”No" response does not belong on our yardstick. In attitude scaling, a count of the errors by questions helps us to decide which are not useful as Scale questions, either because of unreliability or because they are off the continuum. If we were able to phrase our height Questions no more accurately than the following, we could well lmagine that the "Yes" and the "No” answers would be helter- Skelter: Are you taller than a table? Are you taller than the head of a pony? _ Are you taller than a good-sized bookcase? m[hile these are ridiculous, we often wonder whether attitude questions do not throw up just as vague images for the respondent 0 cepe with. In fact, the size of a pony may be a lot more d-e:E'inite! The Guttman approach certainly offers no final answer to questions of reliability or validity, and his error criteria or scale acceptability are still characterized by rules-of-thmb. Nonetheless, they are a big improvement over the sheer intuition W'hich has guided us through many a survey. (PP. 281—282) 38 Several times it has been mentioned that assignment of subjects to scale types is made such that errors are kept to a minimum. When more than one assignment can be made with the same amount of error (e. g., a cutting-point between ranks 19 and 20 for Item C of Fig. 2, p. 32), Riley, e_t a_l. (56) have suggested rules for arbitrary designation of scale type. In Fig. 3 the scale type chosen in such cases is indicated by underlining; e. g., subjects with unique scores of 25 are assigned to scale type 1, even though assignment to scale type 5 would result if the negative responses to items weighted h and 2 had been considered the errors, a number of errors no greater than that resulting from the recom- mended designation. All such quandries for a five—item scale are resolved by applying a rule that assigns unique scores to the one of the pair of scale types which is nearer the center of the distribution of scale types. The distribution for a five—item scale is, of course, 0, 1, 2, 3, h, and 5’ With types 2 and 3 nearest the center. Riley, _e_t 31.: state that the effect of the application of this rule is to distribute error as evenly as POSsible over the five items and the six scale types. Irnproving on Guttman's (25) visual inepection technique (see p. 33, above) , Riley, Q 31. (56) make the following suggestion for quantita- tively evaluating the randomness of error in a scale which has been derived from ranking based on scale types. (It should be recalled that the t'erm "non-scale score" in this quotation refers to any unique score which yields error in assigning scale type.) If there is a high concentration in one or more particular non- scale scores, this indicates that error is not randomly distrib— uted, as is required for an acceptable scale. . . . There is no exact measure as yet for judging the randomness of a distri— bution. Erupirically, if a non-scale score contains over 5 per cent of the sample pepulation, the scale should be viewed with Suspicion. (pp. 293—2914) These authors further state that if any one item of a scale contains 39 more than 15 per cent error responses, that item "is undoubtedly not suitable and the scale as it stands must be rejected. " (p. 295) It should be remembered, however, that a set of items which include such an "msuitable" item may still represent a quasi scale (see pp. 31-33, above) of potential value in predicting an outside criterion. The foregoing summary of Guttman's approach to attitude scaling, particularly as modified by Riley, g _a_l. (56), provides a methodological introduction to the present investigation of possible relations between the attitudes held by male college students and their academic achieve- ments . Construction of the Student Opinion Survey Introduction Skepticism concerning the apprOpriateness of extant instruments for testing the hypotheses of the present study dictated the construction of a more adequate device; specifically, a device or devices suitable for l‘eli ably measuring three areas of student attitude. At the outset the approach taken was rational rather than empirical, in general following the Medal procedure described by Suchman (67) for sampling a universe of itmlS. The concept of an aggregate of characterizing items . . . Conceives of a saxnple from an unlimited number of items which may be used to characterize any social phenomenon. The char- acterizing universe consists of all items which can be used to exemplify the social concept. The determination of whether or not an item belongs to a certain universe, however, remains a matter which must be decided upon by common agreement. A characterizing item belongs to a universe on the basis of some arbitrary decision as to its content. The universe itself is decided upon arbitrarily as the content of interest to the investigator. Some additional means, such as the consensus of judges, might be introduced to help the investigator, but the final decision of whether or not this item characterized the Universe or phenomenon of interest, must be a subjective one. (67, Po 8’4) 'Problaas of Type of Item Prior to deciding upon attitude area contents, a number of consider- ations entered decisions concerning the exact type of item to be used. Although the typical questionnaire item which must be answered "yes," "2 , " or "no" has the apparent advantage of ready quantifiability, it has the disadvantage that it often arouses Oppositional tendencies in subjects of average or higher intelligence. That is, many subjects feel frustrated by their inability to represent the variety of their experiences or atti- tudes when they are forced to make a series of responses so narrow in scope. They sometimes feel that they have been asked to give "yes" or “no " replies to questions of the "Have—you-stepped-beating-your-wife-yet" Variety, and becane resistive and disinterested in the task, seeking refuge in the "’2" column and producing a total score of questionable VEECL‘Ldity. ‘ A subgoal of the present study was the devising of an instrument Which could tap attitudes toward a fairly wide variety of attributes of any one aspect of the academic environment; e. g., attitudes toward academic authorities. For example, knowing what Opinion a student holds °°ncerning instructors was considered of potential interest. An incom- Plete sentence offered the advantage of giving a great potential latitude of reSponse, but if a student were asked to complete a sentence, "I think "1°31? instructors are . . .," he could be expected to give a one-word or ShoI'l‘o--answer reaponse, such as ". . . interesting" or ". . . inclined to be too demanding." Such responses quite probably neglect very many other E"t'it'i‘tundes the student has concerning instructors, attitudes which would add fuller perspective to an appraisal of how he feels about instructors. Furthermore, like most projective techniques, incomplete sentences are extremely cumbersome for testing large groups (a goal held with ref- erence to predicting academic achievement in college), and are subjective 1:1 in their scoring and interpretation (30). Objectivity of scoring tends to re strict breadth of interpretation, thus reducing the value such instruments are purported to ha'fe when interpreted somewhat more intui— tively by a skilled clinician. Returning to the idea of a questionnaire, the student might have been asked to answer "yes," "2," or "no" to questions such as, "Do you think most instructors are interesting? " or, "Do you think most instruc- tors are too demanding? " However, in this manner of presentation the focus of the respondent's attention is so concentrated on the single relationship; _e. g., "instructor—-interesting" or "instructor-too demand- iné, " and upon the fairly obvious social implications of either a "yes" or a "no" answer; that he would be expected often to give the socially acceptable answer or resort to the "’1". A partial solution to the problems just enumerated might have been to Provide a broader range of "socially acceptable" responses (and also °f responses on the "socially unacceptable" side of a theoretically "neutral" position). However, the usual Likert—type scale (e. g., State Whether you "strongly agree," "agree," "mildly agree," "mildly disagree, " "disagree," or "strongly disagree" with the following statement: "Most in‘S‘t‘d‘uctors are interesting.") has the disadvantage of requiring a sepa- rate pairing of—in the example—instructor-trait and rating scale, resulting in a rather unwieldy test format. Perhaps a more serious obstacle from a clinical viewpoint is that items like the one just suggested ask the subject to qualify the extent of his own agreement or disagreement with a given statement. It was thought preferable to ask him to qualify his response in keeping with his ilnpression of the frequency or intensity of the presence of the par- ticumin trait in the instructor. The consequence of such an emphasis m‘ 13h“; well be to give a naive subject the impression that he is rating 7.. .‘vfgr. .w: some‘bl’flmg "out there" in the instructor rather than something in himself, thus tending to minimize the need to be defensive. That such an impres- sion is logically fallacious should not reduce the Operational value of so directing the subject's attention away from an evaluative appraisal of himself. In an attempt to blend the worthwhile features of the several types of 1.th discussed above, but without introducing insurmountable obsta- cles to administration or analysis, an incomplete sentence, such as "I think most instructors are . . .," was combined with a series of quali- fying adjective or participial Likert—type rating-scales, roughly similar to the scales used by Osgood (S3) . Although this method could hardly be expected to eliminate entirely the problem of stereotypical reaponses, the hope was that it would tend to reduce somewhat a subject's tendency to conform to social stereotypes in his responses. Embedding a given rating in a context of many different ratings, all pertaining to the same initial referent (the same incomplete sentence), was viewed as tending t° Push a reapondent to assume his own personal frame of reference within the test structure, at the same time drawing him away from making compar- isons between his own attitudes and those he believes others hold. For a given content area, the use of such a variety of qualifying rat>13ngu-scales gave a breadth of Opinion, which was then readily amenable to cornpletely objective scoring. Such objectively scored rating-scales (hereafter called items) then provided a pool from which unidimensional a(obi-tude scales could be derived (see pp. 56—62, below). 1) eteIlamination of Content Areas F our incomplete sentence fragments were cOmpOSed to represent vari- 0 us ragets of each of the three attitude areas involved in the hypotheses study. An attempt was made to have each sentence refer to some 1:3 universally experienced aSpect Of the college situation or to future aspirations common to college students, but content areas were sought about which it is logical to expect a difference Of Opinion to exist among college freshmen. The four sentence fragments for each attitude area were: Attitude toward academic authorities (”Area X" sentences): A. I think most instructors are . . . B. I think most textbooks are . . . C. I think mOSt academic requirements are . . . Most student intellectual leaders are . . .2 Attitude toward self; self-confidence ("Area Y“ sentences): D. When I am studying, I feel . . . E. Just before a test, I feel . . . F. When called on to contribute to a class discussion, I feel . . . Just before I receive my grades, I feel . . .2 Attitude toward self; striving or achievement orientation ("Area Z" sentences): G. Having a college degree would be . . . 11. Working for recognition from others is . . . J. Being an outstanding success would be . . . Being an influential person would be . . .2 Fifty sets Of polar adjectives or present-participles (APPENDIX A) Were composed as a reservoir of verbal limiting-points for rating—scales, from which sets of items were to be selected to represent a range of logical completions for each of the sentence fragments. The writer and each Of three other counseling psychologists on the Staff of the Michigan State University Counseling Center independently \ 2Fragments discarded by the method described below. 34).; selected from the list of 50 polar combinations the 25 which he considered logically mOst apprOpriate for completing each of the 12 sentence frag- ments. He then crassed out the five Of that 25 which he believed to be the least suitable of those he had selected (those five to be used only as additional items, if needed, in selecting the final set of items for each sentence fragment, according to the criterion and procedure set forth below). Also, he indicated which direction for each item (respon- ses toward which of the poles) he expected to be positively related to academic achievement, that direction to become the a priori "positive" end of the response continuum for the item. Finally, he checked which three of the four sentence fragments he thought most representative of the attitude area under which it was subsumed. An arbitrary criterion of three-fourths agreement among the psychol- °gists was first used to select three sentence fragments to represent each attitude area. COmplete agreement was obtained among the four raters on the inclusion of sentence fragments A, B, C, D, E, and F. Three‘fi‘E‘O‘urlshs agreement was obtained on the remaining three sentence fragments. One rater thought that sentence fragment G was too specific and represented something about which there is likely to be little dif— ference of Opinion among college students (a contention empirically con- firmed later). Three of the raters felt that the sentence fragment, "Being an influential person would be . . .," was a less satisfactory expres'sion of essentially the same content area as that represented by Sentence fragment J. Fifteen sets of polar adjectives or present—participles were selected for each Of the incomplete sentences, in all cases the criterion of three‘Zfi‘ourths agreement being met, though it was twice necessary to resort to use of the crossed-out items among the 25 selected by some of t he raters. When this was necessary, the practice was to choose first 16 those items in which no more than one cross-out selection was needed to produce three-fourths agreement. Poorest agreement among the raters was found for fragment J, for which six of the 15 items were based on the inclusion Of cross—out selec- tions- For sentence fragment A, three of the final 15 items were depend- ent on the inclusion Of cross-out selections. All items for all other sentence fragments were based on the raters' best-20 selections only. For six of the nine sentence fragments, exactly 15 items were Obtained by original three-fourths agreement, without resort to the cross—out listings. For sentence fragment B, 16 items were obtained, forcing the writer to make a further subjective judgment as to which item Was the least suitable of the nine on which three Of the four psycholo- giStS had agreed. The item "orderly-chaotic" was rejected. For sentence fragment G, three-fOurths agreement was obtained for 17 items, and here the items "calm-exciting" and “liberating—restricting" were subjectively rejected by the writer from among the seven agreed upon by three of the four Psychologists. The item "calm-exciting" was similarly rejected from among the nine sentence fragment H items which carried three-fourths agreement, Three-fourths agreement was easily obtained for "positive" direction or all the items, most directions in fact being decided unanimously. The StIlcient Opinion Survey Test booklets, entitled Student Opinion Survey (APPENDIX B), were prepared with an incOmplete sentence heading each of nine pages, placed in the Order (according to sentence fragment designations): A, D, G, B, E’ H’ C: F, and J. The appr0priate 15 sets of polar adjectives or pres“-'r"‘l""'participles were placed below each sentence fragment according to a t’able of random numbers (38), each item's location determined by a 1:6 mnnber assigned to it which corresponded to its rank among the other 15 for the page as the items had been ordered originally in the list of 50 sets (APPENDIX A); and with the direction of "positive" response randomized by a coin toss for each item. A six-category forcing scale, selected with the purpose of attempting to reduce what Guilford (23) has referred to as the "error of central tendency," was employed to separate the polar extremes of each item. Bendig ()4) has demonstrated that an even-category rating—scale (his had four categories) was more reliable than scales having an odd number of categories (three or five categories in his study). However, such an even-numbered scale was somewhat more prone to rater bias than were odd- numbered scales. Bendig (2, 3) also showed that there is no significant change in rater reliability in scales having anywhere from five to nine categories. Each category of the items was anchored by the words, "Always," "Often, " "Sometimes," "Sometimes," "Often," and "Always" at the t0p of the Six columns on each page of the test booklet. Verbally anchoring each Position on the scale continuum in terms of frequency of occurrence Of a giVen trait or eXperience (e. g., "Always," "Often," or "Sometimes" for 33-011 direction on the continuum from "interesting" to “boring" with reference to "I think most instructors are . . .”) provided for each response to an item, a measure of both the direction and the intensity of a s"lbject's attitude concerning a particular subject—trait pair (e. g., "Instructors-—often-interesting" or "Instructors—sometimes" too clemahcling"). According to Bendig (2), such verbal anchoring has the further advantage of increasing somewhat the reliability of self—rating scales. The six steps for each continuum between the polar extremes were 1‘. yped in and numbered l-6 to represent the first six pOSitions on IBM h? Form I. '1‘. S. 1100 A 151 (APPENDIX C). These numbers, l—6, were merely to guide subjects to the appropriate answer sheet spaces, and are not to be construed as having any necessary relationship to eventual scoring, the scoring weight actually being reversed for items having their "posi- tive" extreme to the left in the test booklet. Two answer sheets, prominently numbered in red: 1, 1R (for reverse side of 1), and 2, were used in actual testing; and careful written and vocal instructions were given, so subjects would place responses to sen— tence fragments A, B, and C on the front of answer sheet one, responses to sentence fragments D, E, and F on the reverse side of answer sheet one, and. responses to sentence fragments G, H, and J on the front of answer sheet two. By this means, the three attitude areas, split up in the test booklets, were once again integrated on the answer sheets, facilitating scoring. A11 example for illustration and practice was included on the face ‘ sheet of the test booklet, practice responses being recorded by the sub- jeCtS On the reverse side of answer sheet two. Subjects 0n logical grounds, but with the implications of the Bendig and Sprague study (5) as support (see p. 114, above), it appeared probable that ait'rtazitudes would be more directly related to academic achievement in a group 0f students of average academic potential (as measured by an orientation test battery and compared with college freshman norms) than they W0uld be in either a group with high potential (similarly defined), Whose intelligence and/ or superior preparation might be eXpected to carry them through at least some of their first year college courses with reasonably good grades, even with relatively "unfavorable" attitudes; or in a group of students with low potential, who might be somewhat predis- h8 posed to relatively low achievement no matter how "favorable" their atti- tudes. Students of average potential were selected also because there appeared a good chance of finding in that large group representing the middle 240 to 60 per cent of entering freshmen a reasonably large homoge- neous sample. In order to eliminate problems stemming from various selective fac- tors which might accompany marital, racial, and sex differences, the sam— ple was limited to single, white, male subjects. The entire freshman male pepulation for the fall quarter, 1955, at Michigan State University (MSU) was sorted for a homogeneous group of students having the following characteristics: 1 - Never married 2- No previous college experience 3 . Admitted to the University under regular conditions (not required to take entrance examinations, no irregularities in secondary level preparation) 1;. Scores on the MSU Freshman Orientation Test Battery meeting the following criteria: Raw score MSU DeriVed range Scores a. MSU English Placement Test (E) ll-23 14—7 b. MSU Arithmetic Proficiency Test (A) 26-h0 14-7 c. ACE Psychological Examination; 1910 Ed. (1). Quantitative (Q) 38—148 5—6 (2). Linguistic (L) 52—67 5—6 (1). Vocabulary (V) 12-21 5—6 2). Reading Comprehension (C) 17-23 5—6 \ of 1.1.345 BU Derived Scores comprise a 10 point division of the distribution e an scores for each test or subtest of the Orientation Battery, “that the following prOportions of the total entering freshman class and female combined) are represented for each Derived Score: 1 and 10, .01 each, scores 2 and 9, .03 each; scores 3 and 8, .08 ’ Scores )4 and 7, .16 each; and scores 5 and 6, .22 each. (mal ale scores each i \ \ \ d. MSU Reading Test 149 The indicated test score criteria were selected to minimize scatter of performance, on the premise that students having all their test scores within a fairly narrow normative range can more validly be labeled "aver- age" in their general academic potential than can students exhibiting wide divergence of performance on the tests. The latter type of perform- ance might reflect differences in potential in separate aspects of aca- demic endeavor (e. g., "linguistic" v. "quantitative") and/or the differ- ential effects of a factor like anxiety on their specific test perform- ances, where the tests are administered differently (e. g., timed v. untimed) . The narrower score range was selected for the ACE and the MSU Reading Test because the subtests used were known to be better predictors of academic achievement at MSU than are the scores of the other two tests. Out of the total male freshman pOpulation of 2,993, a homogeneous sample of only 77 was obtained (far fewer than had been anticipated). This ntmi‘ber was further reduced by‘ the loss of nine students who had left ”Rage during the freshman year, leaving 68 to be contacted for testing. Letters (APPENDIX D) were sent to each of the 68 men the last week of the Spring quarter, 1956, inviting them to participate in a research study. An incentive was provided by offering two prizes, one of 20 dollars and one of 10 dollars, to be awarded to two participants on the basis of a blind drawing to be held following the final day of testing. The incen- tive was thought necessary in view of the fact that the men were being asked t"23 take time out from their accustomed routines just before the beginning of final-examination week. Follow-up telephone calls were made to students who had not yet appeared after the first few test sessions. The only additional informa- tion giVen in such calls was that the research involved a survey of stu- dent 0pinion, that it was in no way related to the students' work at the U ‘ mverSity, and that it would require less than the one hour of their 5o tine originally estimated. 01' the 68 men selected, 65 participated in the study. When their complete freshman year records became available at the end of the spring quarter, it deve10ped that eight of the men had failed to earn an average of 15 course credits over the three terms of their freshman year, and one of them had completed orfly two terms in school, having drOpped out for the winter quarter. At MSU a lS-credit minimum is necessary if a student (in courses other than engineering, where the requirement for graduation is 20 credits higher than the total in other curricula) is to graduate after 12 quarters (four school years) in college. However, following the definition of "full-time attendance" set down by the Veterans Administra— tion 1‘ 0r compensation purposes under the "G. I. Bill," an average of 1).; credits per term was accepted, giving a minimum total for the three quarters of h2 credits. This latter criterion restored to the sample three of the nine subjects 10st under the more stringent lS—credit minimum. Further examination of the course records of the other five students (it Will be remembered that one of the nine had completed only two terms ; in SChool) disclosed that they had repeated one or more courses in which they had received grades below "C," thereby earning fewer total credits along With a spuriously high three-term CGPA. In such cases the grade earned in repeating the course replaced the original grade on the stu- dent! S transcript. Such students were retained in the sample provided they had carried at least 11; credits per term, thus having academic expelT'ience as extensive as that of many of the other students in the sam- ple. However, their criterion scores (CGPA) were computed on the basis of their original performances, in a few instances with a credit total of 39 units. Four more of the students qualified for inclusion in the sample 0n the hZ-credits-carried criterion. To recapitulate, two students out of the sample of 65 were excluded 51 from the data. (even though they had participated in the study) on grounds of having had less academic experience than the majority of the sample, leaving a homogeneous sample of 63 unmarried, white, male freshmen of average academic potential, whOSe total collegiate experience consisted of three terms at MSU, during which time they had carried a minimum aver- age course load of 114 credits per term and had earned a. minimum of 39 course credits. In view of the problems attendant upon the use of CGPA as the cri- terion of academic achievement (see pp. 13-17, above), it was considered desirable—if not necessary-to use a three-term CGPA in an effort to minimi ze random error factors such as would have relatively greater effect on only one term's performance. At MSU further stability of the criterion is contributed by the presence of a core curriculum of so—caJled "BaSic College" courses (Communication Skills, Natural Science, and for some Students, Social Science), which, for the subjects of this study, provided a minimum common base of approximately ho per cent of their tOtéfl- Course credits. Cross-validation Sample Concurrent with the examination of the sample of 63 male freshmen in the Specially selected homogeneous sample, all the students enrolled in the five sections of an MSU psychology one-credit "service" course (methods of Effective Study) were asked to respond to the Student Opinion Survesr. The Survey was administered by the writer during regular class time. In April, 1957, seven sections of General Psychology, a required lower division (including freshmen) "General Education" course at San Diegc’ State College (SDSC), were similarly asked to respond to the Stu- dent Opinion Survey, again administered by the writer. In all cases it was posSible to maintain an administrative foxmat comparable to that for 52 the men in the hounOgeneous sample. From the several hundred students tested in the 12 psychology classes, it was possible to obtain a second sample of—coincidentallyn63 white, male freshmen, all nearing the end of their first year in college and with no previous college experience. Unlike the specially—selected homogeneous sample, the second sample contained students representing a wide range of academic aptitude and included both married and single men. Only the variables of race, sex, and educational eJCperience were con- trolled. All the men had carried roughly a full—time course load for the year, but this variable was not as strictly controlled as it had been in the first sample. Incidentally, the CGPA was based on three quarter's work at MSU, but on two semesters at SDSC. Table 1 Frequency Distribution of Cumulative Grade Point Averages Earned by Male Freshmen; One Year's Performance ‘ - Frequency ‘ Achl 1:336“ CGPA limits Exact limits Sample 1 Sample 2 \_ H - 2.8 and above 2.75 and above 9 9 n‘gflh 2.6-2.7 2.55-2.7h lo 7 eax‘ame 2.2-2.3 2.15-2.3h l3 7 ];<)vV 2.0—2.1 1.95—2.1h l2 8 1. 9 and below 1.9).; and below 9 26 \ Table 1 presents the frequency distributions of CGPA for both samples, ) and . . . . readily discloses the practical import of the differences between the two gr 01.1135. In the homogeneous sample (Sample 1), roughly equivalent ' nmnb . ers of subjects fell into each of the six class intervals, but in the 53 cross—validation sample (Sample 2) a very heavy plurality of the men fell into the lowest class interval. When the distribution of Sample 1 is trichotomized—-as is frequently done in the analysis of the data reported in the next chapter-it results in 19 subjects being classified "high- achieverss" 23 subjects, "moderate-achiever53" and 21 subjects, "lowe achievers." The corresponding frequencies in the second sample are 16, 13, and 3h, respectively. Administration of the Student Opinion Survey All the participating students were asked to fill out a face sheet (APPENDIX E for the 63 men of the homogeneous sample, APPENDIX F for the MSU men in the cross-validation sample, and APPENDIX G for the SDSC men in the cross-validation smnple) and were told that the information called for and everything else about the Opinion Survey would be held in confi- dence, being for the examiner's personal use as part of his doctoral research. When the students had completed the face sheet, they were told that the Survey asked for strictly personal Opinions concerning a number of factors pertaining to their experiences in college. They were told that there are no "right" or "wrong" answers to the ratings they were to make, that all answers are a matter of personal Opinion, and that the entire procedure is experhnental and previously untried, thus not even providing the examiner with a conception of how the ratings would be made. During each testing session, the examiner read the instructions aloud, especially emphasizing that each item forces a response in one direction or the other with reference to the incomplete sentence at the tOp of a given page, and that the subjects were to mark only one Space on the answer sheet corresponding to each scale item, the mark to repre- sent both the direction of their choice and the frequency with which they Sh considered the trait or experience they were rating to apply to the sub- ject of the sentence fragment in question. Considering that the students had to record their responses on three of the sides of two separate answer sheets, and with the necessity of changing to a new side and location for each consecutive page of the Survey, it was gratifying to find only a very small prOportion of papers spoiled by clerical errors (not more than two or three per cent out of the total number of over hOO papers). Students completed the Survey in from 10 to 140 minutes, with the median time somewhere around 20 minutes. It was quite easy to administer the entire procedure—introduction, face sheet, and Survey—~in a 50-minute class period; and adnirfistrations in the SInaller groups of the homogeneous sample seldom exceeded 35 to ho minutes. There appeared to be a good deal of student interest in the inStI'uInent and there were no overt signs of distress or defensiveness during the test sessions, as are so commonly elicited when tests such as the Elsi-PI, the Bernreuter, or the Guilford—Zimmerman Temperament Survey 31‘ 8 administered. Chapter III RESULTS Derivation of the Three Attitude Scales Item 8 election Riley, _e_t' £1: (56) have preposed that items for an attitude scale be selected from a pool of items like that represented by each of the 15-1th attitude areas of the Student Opinion Survey. According to these authors , selection should provide that: C a) the item having the nearest approach to unanimity of response ”Positive". or "negative") should certainly be no more extreme than an 80 per cent - 20 per cent Split; (1)) each itan—by—item change from maximum affirmation to maximum negatiOn along the scale continuum should represent more than five per cent of the respondents; e. g., in the homogeneous sample of 63 men, if the item receiving the greatest prOportion of affirmation were so replied to by ’45 subjects (a 71 per cent affirmative - 29 per cent negative Split) 3 the item chosen to represent the next step in the scale continuum toward maximum negation should be responded to affirmatively by no more than 141 subjects (bl being 65 per cent of 63, representing a drop in afSfir‘na‘tion equal to six per cent of the total sample of subjects); and (c) the items bracket the full range of response, from those items to WmCh a sizeable majority of subjects (within the requirement of [a], above) respond in the affirmative to those eliciting positive response from a. minority (again within the requirement of [a], above). 55 56 Adherence to the Riley criteria for item selection tends to insure that all scale types will contain approximately equivalent percentages of the respondents and that there will be no inversion of order among scale types in successive samples. Scoring and Scale Analysis The answer sheets for each of the 63 men in the homogeneous sample were scored according to the a priori assignment of scale values for the original six rating—scale categories (APPENDIX H), the most "posi— tive" category for each item receiving a score of six; the category at the opposite extreme, a score of one. The minimum possible total score fOI' the 245 items of each attitude area was therefore 1453 the maximum PCssible score, 270. Item-score matrices were constructed, one for each of the three attitude areas, based on the scoring outlined above and with subjects ranked from high to low in each matrix according to their total scores for that attitude area. Visual inspection of the matrices disclosed WhiCh of the 15 items in each attitude area most nearly approached the criterion, that "persons who answer a given question favorably all have higher ranks on the scale than persons who answer the same question unfavolt‘ably" (66, p. 9), and the six—category scoring for these selected items Was then reduced to dichotomous scoring. APPENDIX I lists the selected items by number for each of the three attitude areas, and indicates the item response categories which were combined to represent "positi‘fe" response for each item according to dichotomous scoring. The remaining data of APPENDIX I were the bases for several trial- anduerrOr attempts to satisfy the criteria for item selection set forth abOVe (See pp, 31,-39 and p. 55, above). To illustrate: for Attitude Are . a X, the two itans responded to positively by the lowest and highest S7 percentages, respectively (items 10 and 8), might have been selected to define the limits of the Area X scale continuum, with three additional items being selected to represent the intervening steps in the continuum, each step to include "more than five per cent" of the subjects. Approxi- mately equal steps might be represented by items 33, 15, and 25, responded to positively by ht, 56, and 65 per cent of the subjects, respectively. These three items appear to provide a better scale continuum than would items 11, to, and 2-which were also responded to positively by M, 56, and 65 per cent, respectively——because those of the former set yield con- sistently smaller error ratios (the last column of APPENDIX I) along with roughly cemparable numbers of subjects above or below the cutting-points. Sets of five items, selected in the manner just described, were then submitted to scale analysis, following the method of Riley, _e__t _al. (56) outflirted. on pp. 3h—39, above. Initial failures to achieve the minimal S1"and-Etill‘ds for unidimensionality necessitated substitutions of alternate items, new scoring, re-ranking of subjects, and reapplication of the techniques for appraising the scale characteristics of each new set of items- Obviously, the data of APPENDIX I could not serve as more than a rough guide to item selection, based as these data are on the hS—item, Six“38-‘l'aegory matrices. When only five dichotomous items were selected and subjects were ranked according to the new total score distributions, it was inevitable that new error patterns would appear. Only then was it pC’S'Si‘ble to identify and reject an apparently suitable item, now shown to have contributed too high a preportion of error to the new ordering of Subjects, or one contributing to the occurrence of too high an inci- dence of a single non-scale (unique) score. Table 2 lists those items which were finally selected to represent th e three attitude area scales. It may be noted that on three occasions, Table 2 Dichotomized Items Comprising Three Scales, Each Representing an Area of Student Attitude Attitude areas and items Per cent of subjects reSponding: "Positive" "Negative" Area X 31L. I think most academic requirements are 6 7 . . . unishing - rewarding. 3. 3 35. I think most academic gequirements are . . . sensible - senseless. 59 141 143. I think mOSt academic requirements are . . . realistic — unrealistic. 5h 146 33- I think most academic requirements are . . . inconsistent - consistent. M; 56 145- I think mOSt academic requirements are . . . encouraging — discouraging. 35 65 Area Y 5h- “Then I am studying, I feel . . . enthusiastic - apathetic. 67 33 86. mien called on to contribute to a class 8 ._ discussion, I feel . . . timid - confident. 56 ml )4. ifllen called on to contribute to a class 8 discussion, I feel . . . sad - happy. h9 51 9‘ When called on to contribute to a class __ discussion, I feel . . . orderl - chaotic. hl 59 93‘ "hen called on to contribute to a class discussion, I feel . . . elated - depressed. 37 63 Area Z 21‘ WOrking for recognition from others is 19 _ o o o Valuable - Worthless. 71 29 ‘ W(Dr-king for recognition from others is 36 B _ . . . immature - mature. 56 ml ‘ elng an outstanding success would be 314. B . . . . sensible - senseless. Ml 56 elng an outstanding success would be 141 B _ . . . immature — mature. 36 6).; ° elng an outstanding success would be . . . good - bad. 2h 76 \ appeagote'f'whe items are numbered and the polar Opposites for each item underlias in the Survey booklets (APPENDIX B). r’he terms which are ned define the "positive" extremes of their respective items. div ‘I." 59 twice in the Area I scale and once in the Area Y scale, the recommenda- tion that steps along the scale continuum encompass "more than five per cent" of respondents was not met. In fact, the step from the "most favorable" item to the next "most favorable" itan in Area X, and from the "least favorable" to the next "most favorable" item in Area Y, each include only four per cent of the subjects. Although this failure to meet the criterion for differentiation of response within the scale is of no immediate concern, scale instability can be expected in future applications of the weaker scales. (Such instability did appear in cross-validation-—see next page-~but it was not in conjunction with the abovedmentioned items.) Table 3 Summary of the Extent to Which the Three Attitude Scales ‘ Meet Criteria for Scalability (Unidimensionality) Obtained % of Ss with Least possible coefficient of most frequent . coefficient of reproducibility non-scale score poorest $teT reproducibility (Criterion: (Criterion: (Criterion. .90 or more) 5% or less) 13% or leSS) % of error in Attitude area ' x .59 .91 6.1:} 12.7 Y .59 .90 h.8 1m 2 .6h .93 3.2 9.5 *Only one non-scale score was obtained for so many subjects (N=h). Had this score been earned by only three subjects, the criterion would have been met. 0f considerably greater consequence to the present study is the fact that all three scales meet the requirements for scalability, in spite of the small size of the pOpulation sample. It is customary in such research to work with a random sample of several thousand subjects, and Guttman (25, 26) has stated that it is very difficult to attain unidimensionality , .‘ '3';- mm? 60 with fewer than 100 subjects. Quite probably the very homogeneity of the present sample has favored the appearance of unidimensionality. Appen— dixes J, K, and L present the item-score matrices and scale analysis information for the three attitude scales as finally constituted. Table 3 summarizes the formal characteristics of the three scales. Cross-validation The answer sheets of the second sample of male freshmen were scored for the three five-item attitude scales as a cross-validation of the scale characteristics just reported. Table h summarizes the results of that procedure. T able )4 Summary of the Extent to Which the Three Attitude Scales Meet Criteria for Scalability; Cross-validation Sample Obtained . % of $5 with . , Least possible coefficient of most frequent % or error 1n Attitude coefficient of re roducibility non—scale score poorest itan area reproducibility Criterion: (Criterion: (Criterion: .90 or more) 5% or less) 15% or less) x .62 .92 6.11 11.1 Y .66 .90 9.5 22.2 2 .61 .89 7.9 12.7 For the second sample of male freshmen, the Area X items were found to have quite different levels of affirmation from thoae obtaining in the homogeneous sample. That is, in the second sample the items receiving the lowest and next lowest prOportions of affinnation, respectively, had received the next lowest and the lowest, respectively, in the first sample; and similarly for the next two items along the scale continuum. The values reported for Area X in Table h are those obtained with scoring , ‘er'i -, .i 61 revised to conform to the new scale order in the second sample. As in the original sample, only one non—scale score was found to exceed the maJdJnurn frequency criterion in Area X. The scale continuum of Area I was maintained in the second sample, but with the three least affirmative items very close together; i. e., with nearly equal prOportions of the student sample responding in the same direction. In addition to the most frequent error score reported in Table 1;, another non-scale score was obtained for 6.); per cent of the subjects. It should perhaps be mentioned that the second poorest item of Area Y contained only 11.1 per cent error. As with Area Y, the original Area Z scale continuum was maintained in the cross-validation sample. However, in addition to the most fre- quent error score reported in the table, two other non-scale scores were each obtained for 6.11 per cent of the subjects. Considering that such small samples are involved in both student samples, and also that the cross-validation sample was markedly different from the homogeneous sample from whose responses the scales were derived, the r esults just reported are encouraging. At the very least, it appears that the attitude scales which have been derived are quasi scales, and thus qualify as useful predictive indexes, particularly for samples of male freshmen of "average" academic potential. flthough it may appear that the reduction of the six—category scor- ing to dichotomous scoring and reduction of a [LS—item scale to one of only fiVe items have eliminated much useful information, in reality these procedures have primarily eliminated error contributed by extraneous (non‘scale) items. Stating it more concretely, the probability is much greater that Student A is more favorably diSposed t0ward the content of At - titude Area I than is Student B if that contention rests on their r esPfiictive ranks (scale types) on the five-item scale than if it is based Naming the Scales The original goal was the measurement of what were somewhat vaguely called "attitude toward academic authorities," ”attitude toward self; . self-confidence,” and "attitude toward self; striving or achievement 7 orientation." with the final selection of unidimensional five-item scales, a revision of scale names seemed called for, in order to reflect _. on their ranks on the original hS-item scales. 1 E . Returning to Table 2 (p. 58, above), it will be observed that the 1? five items of the Area X scale all complete the sentence fragment, "I 5 think most academic requirements are . . .," and may be summarized as referring to a student‘s appraisal of the extent to which such require— ments are "objectively" apprOpriate ("sensible,” "realistic," and "con- sisten‘t") and "subjectively” satisfying ("rewarding" and "encouraging”). V More concisely, it appears that the Area A scale measures the extent to “hi-Ch a student considers academic requirements compatible with his needs and aspirations; or, emphasizing the other side of the coin, the extent to Which he considers himself to be in essential accord with academic requirements. The five items defining what has been called Attitude Area I have been I‘enamed the Institutional Identification _S_g_a_lg (IIS) . It should not be asSAimed that the former "attitude toward academic authorities” notion is hereby discarded. Rather, the scale continues to include attitudes toward academic authorities, but with an emphasis on identification with geneI‘al institutional representations of such authority, in contrast to a possible identification with specific persons holding positions of authority. The Area I items have a decidedly emotional tone, reflecting the more accurately the principal emphasis of each set of scale items. 63 extent to which a student feels "enthusiastic," "confident," "happy,“ “orderly," and "elatedll as he performs in roles assigned him in college, particularly in the public roleof student—in—class. In them there is self—appraisal, but with at least an implication that the emotions are the con sequences of social experiences, reflections of the ways in which the student has been (and/or anticipates being) received by fellow stu- dents‘ and faculty members as he performs his role publicly. Conceivably he feels confident, enthusiastic, and happy, and tends to derive contin- ued satisfaction in situations where his efforts have been well—received (rewarded). The Area I scale carries a heaVy self—confidence connotation and has been renamed the Self-Confidence agile (SCS). The five items of Area Z appear to represent the degree to which a student, considers "working for recognition from others" and "being an outstanding success" to be worthy goals. Consequently, this area scale has been renamed the Achievement 1a_l_1_1_e _S_ca_l_§ (AVS). Although the scale names have been revised to reflect the particular contents of their respective items, it should be remembered that all the items of the original h5—item scales helped determine the final five—item scales and therefore are implicitly represented by them. Different ini— tial Sets of items would have resulted in different original total scores and. :- a1":ld.ngs, and therefore in the selection of other "most representa- tinn items than those which were selected (APPENDIX I). It is quite possible that there are other sets of five (or more) items from the original matrices which would yield scale characteristics as Satisfactory as those which have been selected. The item combinations Which compose the three scales were merely the first to meet the criteria for uhidimensionality in their respective content areas. And finally, among the remaining ho items of each original matrix, ‘1‘. here may be one or more additional unidimensional scales, representing : ,v:e-cv7?¢~x~ fly . p. .. a. "nu-3.1% 6h smewh-at different item universes than those portrayed by the three scales which have been derived. In other words, the selection of three unidimensional five-item scales should not be viewed as necessarily exhausting the reliable scale information provided by the original Survey booklets. Only further manipulation of the data--beyond the scepe of the present stuc‘tywcould establish or refute an empirical basis for these speculations. APPENDIX M lists CGFA, scores on the Orientation Tests, and attitude scale types for the 63 men of the specially selected homogeneous sample. In APPISNDIX N will be found background data of possible interest. Applications of the Attitude Scales to the Problem of Academic Achievement Problems of Attitude Pattern Designation Attitudes and achievement could be compared in a variety of ways, but there are two practical considerations which here served as guides to the most apprOpriate comparisons: (a) the student sample is regret- tably Small, and (b) both achievement and attitude data are in terms of ordinal scales. It will be recalled that the original plan had been to use the three attitude scales simultaneously as profiles or patterns from which to pre- dict achievement. With six scale types in each of the three attitude areas, there would be 216 possible profile combinations. It could hardly be maintained that each of these profiles represents a psychologically unique attitude pattern, but if a "high—low" dichotomy of each of the attitude scales is made, eight basic patterns result, under which the 216 specific profiles may be subsumed. Several questions arose concerning the efficacy of the three-scale pattern approach. First was that of definition of "high” and "low," 65 whether to use an absolute definition in which scale types 0, l, and 2 would be "low" and scale types 3, h, and 5 would be "high" (in which event, for example, the profile 2-h-hh would be labeled ”low-high—high" [L-J-I—HJ]4 but the profile 3-5—5 would be labeled ”high—high-high" [H—H-H], although both profiles have the same internal relationships among the subject‘s scale types); or to use a more relativistic labeling which would emphasize the relations among an individual's attitudes rather than the normative relations of his attitudes compared with the standardiza— tion sample. A second problem encountered in the three-scale pattern approach, even when the many combinations of scale types are reduced to eight basic patterns, is that imposed by the small total sample of students. A com— parison of eight attitude patterns with three achievement levels (”low," "mOdeI'ate," and "high") results in a three-by—eight contingency table. With a total sample of only 63, the eXpected cell frequencies in the 21; cells of the table are too small to inspire confidence in any result obtained. Finally, comparisons of the three attitude scales with each other disc10 sed that a hypothesis that they are essentially independent of each other (the null hypothesis) would have to be rejected for one pairing. Table 5 presents the analysis of the interrelations among the attitude scales. ,4 The contingency coefficients of Table 3 were derived frOm chi squares Of three—by-three contingency tables of low, moderate, and high scale types (two scale types to each of the categories) in each of the attitude areas. For three-by-three tables, the maximum attainable coefficiency of contingency is .816 (22). The confidence levels referred to are those of SCS hAll such pattern designations list the IIS scale type first, the Seal ale type second, and the AVS scale type last. 66 the respective chi squares from which the contingency coefficients were derived. Table 5 Contingency Coefficients Representing Relationships Between Pairs of the Three Attitude Scales Contingency Confidence Scales coefficients levels IIS—AVS .285 .25 scs-Avs .h37 .01 113-503 .263 .35 Although there is a significant relationship between SCS and AVS, it is not large enough to negate the value of three-scale profiles. Of course, a relatively large prOportion of obtained differences in subjects' responses to the SCS and AVS scales would have to be attributed to chance, thus tending to reduce the levels of confidence accorded conclusions dram from profiles employing both scales. A“waltude Patterns and Academic Achievement Levels; General Considerations COntingency tables were drawn up for the eight basic three—scale patterns, but no matter how the question of pattern designation was met 1 (absolutistic or relativistic), no evidence appeared to warrant rejection or the null hypothesis; 1. e., that no general relationship obtains between the three-scale attitude patterns and academic achievement. The most nearly significant chi square obtained was for a three-by- eight Contingency table (Table 6) derived from attitude patterns in which a relativistic approach was made to pattern designation. Here chi square was 17-59 (without Yates correction for continuity), significant at only 6? the .25 level, and this in a contingency table whose largest expected cell frequency was 34.38. Of the 214 expected cell frequencies, 15 were below 30m- Table 6 Comparison of Eight Basic Three—Scale Attitude Patterns with Three Levels of Academic Achievement . Achievement levels (CGPA) Three-scale attitude patterns (IIS—SCS-AVS) Low Moderate High (below 2.15) (2.15-2.90 (above 2.51.) H-H-H t‘t‘r‘mhtfi I Hr‘mt'n: I I n: :‘NOHNU‘LUXN memI—‘L—NN HWV'LI-‘Ol-‘U'Lw A two-scale pattern, using 115 and SCS, also proved fruitless. Here a low‘Inoderateuhigh trichotomy of both the attitude and achievement con- tinue. produced a three-by-nine contingency table whose uncorrected chi square or 18.15 was found to be significant at only the .30 level and whose highest expected cell frequency was only 3.65. Not only did dichotomization of the achievement continuum for both three”Scale and two-scale pattern analyses fail to increase expected cell frequencies appreciably, obtained chi squares in all instances were found to be much less significant than those already cited. Thus, the present 68 study lends no support to the assertion that there is a relationship between students' attitude patterns per se (as measured by the three unidimensional scales of the Student Opinion Survey) and their academic achievement levels (as defined above for the homogeneous sample of male college freshmen). Each Attitude Scale Compared with Academic Achievement Attitude Levels and Achievement Levels. The simplest comparison, that of each of the attitude scales with achievement, is made in Table 7, Table 7 Composite Table for Comparisons of Each of the Attitude Scales with Academic Achievement Levels (Entries for Each Attitude Scale to be Considered Independently of the Others) Attitude levels Achievement Low (scale Moderate (scale High (scale levels (GGPA) types 0 8c 1) types 2 6c 3) types 14 8c 5) IIS SCS AVS IIS SCS AVS IIS SCS AVS M __ __ __ __ __ __ __ __ __ Hi h (abovi 2.51,) 6 3 7 5 lo 5 8 6 7 MCutlerate (2.154510 8 12 lo 9 6 7 6 5 6 I. (beloxwzols) 7 7 S u 6 lo 10 8 6 \ with the statistical analyses summarized in Table 8. The information °°ntained in Table 8 is based on each of the three-by-three contingency t ables of Table 7, and may be accepted with a fair degree of confidence be Gauge the sample size was sufficient to meet the criterion for minimum 69 expected cell frequency size, obviating the need for Yates correction. Table 8 Statistical Summary of Comparisons of Each of the Attitude Scales with Academic Achievement Levels Attitude Confidence Contingency scales Chi square level coefficient IIS 3.0? .55 .216 SCS 7 . ’46 . l2 . 325 ‘Avs 3 .19 . SS . 220 If any positive conclusion is to be drawn from the study, it is hinted at in Table 8, in which only SCS even approaches a significant relationship with academic achievement. The .325 contingency coefficient is certainly not large, and it goes without saying that the .12 level 5113151115 really very little confidence for rejecting the hypothesis that there is no relationship at all, even between SCS and academic achieve- ment, However, in view of the unreliability of the criterion measure (see pp. 13—17, above), the observed trend was thought significant enough to warrant further exploration of the pOSsible relationship. Attitudes and Achievement. Table 9 shows that, of the three atti- tude Scales, only SCS provides a potentially useful basis for prediction Of CGPA. Secondly, it shows that the relationship between these two vari"="~}i>les--and for prediction of CGPA from SCS scale types only--is not linear. However, in spite of the quite respectable levels of statistical confidence with which it appears the inferences of Table 9 may be enter- tained’ it must be pointed out that neither the product-moment correla- tion nor correlation ratio (.933) are really defensible statistical tech- 115. ques f or analyzing the data at hand. The reasons are that: (a) both 70 variables represent ordinal scales; (b) the attitude data has a very - narrow variance, being limited to only six scale types; (c) it appears doubtful that homoscedasticity obtains; and (d)3vthere are very few cases falling into several of the row and column totals in the scatter diagrams. The computations have been made and reported only because they may shed some possible further light on the nature of the still tentatively accepted relationship between SCS and academic achievement. Table 9 Relations Between Each of the Attitude Scales and Academic Achievement - Coefficient of Coefficient of Coefficient of Attitude linear regression curvilinear curvilinear (Pearsonian 3) regression (eta) regression (eta) scales between CGPA and of CGPA on each of each attime each attitude scale attitude scale scale on CGPA \— IIS .016 .252 .521; $05 .081* M3». .310 AVS -.1h9 .363 .hZS \ 1' *Chi square test for linearity of regression permitted rejection of _1nefilrity at the .01 level, and the eta coefficient was found to be s:gnlficantly greater than zero at the .01 level (with fiducial limits gthO212 to .73h). Linearity of regression could not be rejected for any er Variables, even at the .10 level. Summary. The present study furnishes no statistically significant evidence of relations between three areas of student attitude and the academic achievement of those students, either when the attitudes are considered together in various three-scale or two-scale patterns or when th . . . . ey are compared Singly With academic achievement. In fact, this study 1e nds no support to any of the positive assertions in the reviewed liter- atur . . e Concerning relations between attitudes toward authorities and aca— c achievement or between attitudes toward achievement and actual , , —; achievement. Self-confidence and Achievement Introduction. Only an attitude scale purportedly representing a dimension of self-confidence was found to manifest a suggestion of poten- tial gain in predicting academic achievement, and that with only a low curvilinear regression, with "high" achievement being most typical of students with only moderate levels of self-confidence. Self-Confidence Scale Levels and Achievement Levels. Returning to the SCS portion of Table 7, and pursuing the relationship in question a bit further, it is interesting to note that information concerning self- confidence level (on the low-moderate-high trichotomy) advances effi— ciency of forecasting achievement level (a comparable trichotomy) by 30.h per cent. That is, 23 of the men in the sample of 63 were found to have achieved at the "moderate" level, and since neither of the other achievement levels contained as many subjects, "moderate" achievement would have been the best estimate of achievement for the subjects of the study (a quite logical estimate for students of supposedly "moderate" potentiall). However, when the subjects are tallied in the three-by— three contingency table which introduces the variable of self—confidence to that of achievement, "low" achievement becomes the estimate for stu- dents with high self-confidence and eight of them are correctly predicted; "high" achievement is the estimate for students of moderate self-confi— dence and 10 of them are correctly predicted; and "moderate" achievement is predicted for students low in self—confidence and 12 of them are correctly assessed. The total correctly predicted with information con» cerning self-cenfidence level is thus 30, an increase of 30.h per cent Over the 23 predicted without such information. In \dew of the very careful sampling and the completeness of the 72 controls over intellective variables in the present study, and consider- ing that almOSt all of the selected subjects (63 out of the total of 66) responded to the Student Opinion Survey, the obtained forecasting effi- ciency with the Self-Confidence Scale alone appears promising, not so much for the dubious value of offering a means of assigning students to achievement categories, but for the possibility of providing additional leads into the dynamics of academic achievement. Self—Confidence Scale Levels, Three-Scale Attitude Patterns, and Achievement Levels. It will be recalled that a supplementary test was preposed (see pp. ZhPZS, above) to investigate the achievement levels of two of the three-scale attitude patterns, the pattern in which 115 and AYS are high, relative to SCS (H-L-H); and the pattern in which 115 and AYS are low, relative to SCS (L-H-L). It was stated that the eXpectation for students with the former pattern was for achievement beyond the level predicted by intellective indexes, while that for students with the lat- ter pattern was for lower-than-predicted achievement, with the further qualification that students with a really low SCS score would be no better than moderate-achievers regardless of their three-scale pattern. Although statistical tests for significance are vitiated by the extremely small cell frequencies, examination of Table 10 teases out some further implications of the present data for the predictions just reviewed. Contrary to prediction, students with IIS and AVS scores relatively higher than SCS scores (H—L—H) tended to achieve pglgw expectation. Five such students (half of the pattern total) were found among the 21 "low-achievers" and of those five, fOur were found to have moderate SCS scores! Four students with the pattern were "moderate- achievers," but three of those four had lgw SCS scores (scale types 0 or 1)! Only one student out of the 10 with the H-L—H pattern was found among the "high-achievers," and he had a low SCS score. , - l .' at 73 Conversely, students with low 115 and AVS scores relative to their SCS scores (L~H—L) tended to achieve 32219 the level predicted from their academic aptitude scores. Once again the pattern was common to 10 out of the 63 students, and of the 10 ngng was found among the 21 "low- achievers." Five were "moderate-achievers," of whom three had high SCS scores and two had moderate SCS scores. Of the five in the "high-achieve— Table 10 Frequencies of Occurrence of the Eight Basic Three—Scale Attitude Patterns (IIS-SCS—AVS) When Levels on the Self-Confidence Scale Are Compared with Academic Achievement Levels (Table 6 Superimposed on the SCS Portion of Table 7) l ment" category, two had high SCS scores and three had moderate SCS scores. ‘ I I I Achievement levels (CGPA) 508 level (scale types) Low Moderate High (below 2.15) (2.15-2.5h) (above 2-5h) H-H-H 2 H—H—H 1 H-H-H 3 High H-H—L h H—H-L l H-H—L 1 (u, 5) L-H-H _g L—H—L _1 L—H-L _2 Totals 8 S 6 H-H—L l H-H-H l H-H-L h earn h HabL 1 baa, 3 Moderate L—L-H l H-L—H l L-L-H 2 (2,3) Laai 1 Irbi 1 , ___ lFH—L _g __ ’ Totals 6 6 10 H-L-H l H-L—H 3 H-L-H 1 LOW H-L—L l H-L-L 5 H-L-L l (O 1) L-L-H l L-L-H l L-L-H 1 ’ L—L—L __L_; L-L-L _; _ Totals 7 12 3 In other words, not only did this study not confirm the assertion . that "over-achieving" college students tend to be driven by their needs ‘ for dependence upon authority-figures, it tended toward exactly the Oppo- 7h site assertion: that students with attitude patterns commensurate with such needs achieve below a level predicted for them on the basis of intellective indexes. The trend toward ”under-achievement" is supported even when the achievement continuum is dichotomized. Eight of the 10 students with the H—L—H pattern were found among the 3b "low-achievers," while only two were tallied with the 29 "high-achievers." The most widely prevalent three—scale attitude pattern, H—H—L, was cammon to 12 of the 63 students. To some degree this pattern might reflect identification with authority-figures, at least to the extent that self-confidence and institutional identification are coequal in con- trast to relatively low achievement valuation; i. e., institution and self are both favorably appraised. Of the 12, five were found to be "low-achievers" and five, “high-achievers." Four of the five "lowe achievers" had high SCS scores, while four of the five "high-achievers” had moderate SCS scores. The two "moderate—achievers" with the H—H-L pattern split, one with high SCS and one with moderate SCS. Once again, the evidence suggests that the students most extreme in their "identifi- cation with anthority-figures" fail to achieve up to their potential, to say nothing of their being "over—achievers." Summary. Slightly over half the ”highrachievers" fell into two three-scale attitude patterns, neither of which seems to support the con- tention that such "over—achievers” tend toward exaggerated identifications with authority—figures as a defense against their own feelings of inade- quacy (or for any other reason), thereby achieving academic success as a function of their slavish efforts to emulate their academic superiors. In fact, the evidence suggests that most such "high-achieving" students hold no better than moderately favorable attitudes toward themselves and toward authorities, or even unfavorable attitudes toward the latter. It was also found that 12 of the 19 "high-achievers" had relatively low 75 achievement valuation scores.) Conversely, 10 of the 21 "lowaachievers" were found with the two three-scale patterns most suggestive of conformity or exaggerated identification with authority—figures. (And 11 of the 21 had relatively high achievement valuation scores.) No mention has been made yet of the "moderate—achievers." As already noted, it was something of a surprise to find students low in self—confi- dence tending toward "moderate" achievement (12 of the 22 students at that self—confidence level), and it was even more surprising to find stu- dents low in self-confidence going counter to the trends already reported. That is, of the 15 low self-confidence students who attained "moderate" or "high" achievement levels (see Table 10), 10 exhibited three-scale attitude patterns essentially compatible with the notion of identifica- tion with institutional authorities as a reaction against feelings of personal inadequacy. Six of the 10 had relatively high (moderate or high) 115 scores in contrast to their low SCS and AVS scores (H—L—L), while the remaining four displayed the H—L—H pattern which had been predicted for "high-achievers" but which was found typically among "low—achievers." Three of the latter subgroup of four students were, however, found in the lower half of the "moderate-achievement" category, the remaining student occupying the lone "high-achievement" position for subjects with the H-L—H pattern. Chapter IV DISCUSSION AND SUGGESTIONS FOR FURTHER RESEARCH Implications of Present Failure to Demonstrate Statistically Significant Attitude-Achievement Relations -: A number of factors combined in the present study to facilitate the appearance of any relations there might be between students' attitudes and their academic achievements. The participants in the study were drawn from a clearly—defined "average academic potential" pepulation, a group within which it was assumed that non-intellective factors such as attitudes would have maxi- mal influence on academic performance. The student sample was homogene- ous also with reference to academic experience, marital status, race, sex, ands-incidentally--age. The subjects had all completed nearly three full quarters of the freshman year at a large middle—western state univer— sity when the attitude data was collected from them, and their CGPAs have at least the stability attendant upon their being based on at least 39 academic credits, of which a minimum of no per cent were earned in a core curriculum common to all the subjects. The obtained attitude scores may be said to reflect a final, momenr tary cross—section of student reaction to experiences of the first year in college, experiences somewhat differently represented in the students' CGPAs. The importance of the fact that the measures of the two classes of variables were more or less concurrent is that all the "predictions" discussed in the last chapter represent temporally static statistical inferences rather than predictions forward in time. 76 ”7, 1 ‘ ‘ 77 ‘ith so many factors in the present study tending to maximize the rance of relations between students' attitudes and their academic vements, the fact that only meager hints of slight relations appeared plex ones at that—-raises very serious doubt that psychometric aches to the problem will be very productive, and leads to almost ete pessimism concerning the prospects for the more difficult task king accurate predictions of academic achievement forward in time attitude scores. 0n the other hand, there are the very real problems of the unreli- ty of the academic achievement criterion, CGPA. Conceivably, a r criterion would have permitted the appearance of some very signi- t relations between attitudes and achievement, even using the pre- psychometric approach. The problem in that event becomes the fonni- one of establishing a more reliable achievement criterion and ing large numbers of instructors (individualists all!) to adOpt it 0 apply it rigorously. At the present time, weekend excursions to ar galaxies appear more probable of accomplishment! At the very least the present study demonstrated the necessity for ully specifying the nature of the relationships under investigation he particular circumstances under which they are supposed to obtain, voiding loose or grandiose claims about such variables as "attitudes d authoritybfigures," "passive—dependent identification," "stereo— al thinking," "need for achievement," "need to strive," "drive d mastery," "emotional stability," "social adjustment," etc. Cer— y each of these labels, and many hundreds of others like them, 5 a multifarious range of specific behavioral referents and even e variety of classes of behavior. Behavior scientists are faced the task of getting down to more detailed descriptions of behavioral xts, rather than too hastily summarizing and categorizing behavior. ;, a ,, ———_—7. 78 For example, it appears quite probable that the Naval Aviation Cadets terviewed by Hollander and Bair (3h, 35) exhibited differences in atti- ies ("identifications") toward their instructors as a function of their lperfomances in flight training and/ or differences in flight training ‘formances as a function of their attitudes, but learning to fly an plane has little in common with the varied behaviors whose combined essments produce an academic CGPA. Learning to fly has perhaps a much 'ger component of imitative behavior, of deep reliance upon the specific lls—the smallest mannerisms—~0f the instructor, skills which the stu- t seeks to reproduce in detail, and always with the real threat to e or limb should he fail in his efforts. A respect for the instructor a person, perhaps even a kind of affection, might well be expected to 'Llitate learning. So too in the psychotherapeutic situation (19), rein the analysand or client is faced with but one "adversary," who 3imultaneously a most vital "protagonist." In college the picture is unquestionably different, if only in its laxity. The student—~each student-- has his own peculiar goals, and ay reach then because of, in spite of, or quite untouched by impres- numbers of his academic peers and superiors. As Bendig (1) and all and Bendig (57) have shown, students' ratings of instructors are itially independent of their course experiences. That is, they may ary fond of an instructor but have a profound dislike for his course, .ce versa; they may enjoy a course and yet do very poorly in it, or versa; they may earn a high grade and still dislike the instructor, ,ce verse; or any cornbination of any of the variables. Attempts to Rationalize the Trends of the Study: Feelings of Inadequacy, Institutional Identification, and Academic Achievement ['he collegiate experience—~especially the first year or two in a 79 beral arts college—brings the student into contact with a wide Variety Lf subject matter and perhaps an even wider spectrum of human variation. 'or the individual student, the only real stability in the situation is hat which he carries within himself. If, like Hollander and Bair's (3)4, S) cadets, he adheres to the ways and values of one instructor, he may and to excell in that man's field. Provided the student had sufficient redits for him to accept his model as an advisor and to concentrate his :ademic activities in the instructor's field of specialization, he might :11 build a quite respectable CGPA. However, in the first year of col- ge, a student (like those of the present research sample) who attempts emulate a few instructors exclusively is very likely to find himself odds with other instructors whose ways are counter to those of his osen models. The liberal arts college freshman, it is here maintained, needs a .r degree of independence, a breadth of interest and curiosity, an ,en-mindedness," if he is to produce an Optimal scholastic record. He forced (by academic requirements, if not by his own inclinations) to ociate himself with such a variety of peOple and ideas that he dare become- too dependent upon any of them too soon. The penalty for such nature adherence to a man or a point of view is, it is prOposed, a iced capacity to entertain the points of view or personal idiosyncra- ; of the many others he must respond to in his total educational :rience, and the mark of such a reduced capacity may well be a con- :rable number of less—valued performances and a lower CGPA. The foregoing appears to this writer to be an implication of the lts of the present study. There was a marked tendency for students only moderate self-confidence to achieve above predicted levels (10 of the 22 students with moderate self-confidence), especially when a level of self-confidence was combined with a comparable or lower 80 31 of institutional identification and a lower level of achievement lation. Seen in contrast to the overwhelming tendency toward achieve- Jw potential among students with moderate self-confidence but with 1er institutional identification and achievement valuation (four out 'ive students), the first of these facts suggests the possibility that "high-achievers" are better able to tolerate their feelings of inad- Loy without seeking premature and necessarily artificial identifica- 15 with institutions and peOple whose goals and means to goals they :rfectly understand. It is suggested that the "low-achievers" among students with only moderate self-confidence were reacting somewhat msively to similar feelings of inadequacy, a defensiveness reflected only in the high institutional identification of their relativisti- y defined attitude patterns, but also in the hypothesized prenature tallizing of behavior patterns in the college setting. It may be recalled also that high self-confidence, combined with . institutional identification and moderate or low achievanent valua- , was typical of "low-achievers," while moderate self-confidence, rate or high institutional identification, and low achievement valua- was a pattern common among "high-achievers." The H-H-L pattern g "low-achievers" implies a failure to recognize or to admit inade- Les either in the institution or in self, either out of defensiveness it of a more genuine satisfaction with things-as-they-are, with a aquent inflexibility in academic pursuits (if the pattern reflects :‘ensive reaction) and a subsequent "sour-grapes" rationalization of ralue of achievement (e. g., "working for recognition from others"), relative absence of a need for achievement (if the pattern reflects .isfaction with the status quo). 0n the other hand, the H-H-L pattern among "high-achievers" (whose :cores were really only moderate) is hypothesized to reflect an 81 ass of and a tolerance for some inadequacy in institution and self :orrespondingly flexible participation in the variety of student ance. It is hypothesized that for these "high-achievers" a rela- yvaluation of achievenent has a different significance than it has students who are failing or just barely " scraping by." Such rela- svaluation of achievement, coming in conjunction with actual high ament may reflect a bit of "false modesty," or an element of disap- ant with achievement once attained (assuming the students them- consider their level of achievement to be high), or even the rtance of the more formal signs of achievanent to students whose satisfactions come in the performance of daily activities rather a the fantasied anticipation of rewards from others; e. g., recog- and prestige. inally, there is the notable exception to the general trends, the enon of "moderate" achievenent among students with low self-confi- especially when that low self-confidence is combined with moderate n institutional identification and/or achievement valuation. That tudents are so inclined toward "moderate" achievement (12 out of low in self-confidence) rather than toward "low" achievement out of the 22), was itself something of a surprise; but to find these 12 "moderate-achievers" a preponderance (eight out of the 12) very patterns (H-L-H and H-L-L) which were not associated with ement at other levels of self-confidence is decidedly problematic. e trend is carried over into the "high-achievers" among students ow self—confidence, where each of the patterns in question was ed by one of the three students. allowing the "defensiveness" rationale, is it possible that stu- with really low self-confidence levels are only too aware of their _uacies; that is, recognize them (or even have an accentuated aware- _r- r1 —~ _~ ' o 82 ,' them), but are poorly able to tolerate them? If such is the case, ry find some strength through their fancied participation in the .ies and advantages of the institution and its representatives. rould be the students typically referred to by the psychoanaJysts :y Rust and Ryan (58), and by Ryan (59) as striving for a passive- :nt kind of identification with authority—figures and achieving .c success as a consequence. .udents who are low in self-confidence (absolutely, with scale t or 1 on the Self-Confidence Scale, rather than relative to l/or AVS scale types) might be expected to represent a considerably prOportion among the clients of counseling centers and psycholog- inics than their prOportionate representation in the total student .ion. In this connection, it has been most interesting to obtain ;ions from colleagues with reference to the achievement levels mying the two attitude patterns which received Special attention and L—H-L). In a quite informal survey of the Opinions of some tr ten colleagues, those with experience in college counseling or L1 psychology (roughly half) all predicted high achievement for mer pattern and low achievement for the latter (not confirmed in :sent study), while all of the men with only teaching and academic 1g experience with students predicted the reverse (which was the 'or the present study). At least there is here a kind of presmnp- idence that quite different student samples may be eXpected to very differently in response to attitude patterns and/or respond its different attitude patterns in the face of similar academic Suggestions for Further Research 1e present study certainly was not designed to cOpe with the many 83 gs of essentially an Egg interpretation suggested above. Indeed, not even apprOpriately address itself to a description of sequenr matterning of attitudes and achievement; for this, a longitudinal is needed. At best, the present study has hinted at some possibly icant concurrent patternings of attitudes and achievement. f the hypotheses of this chapter are to be given adequate tests, ly must future studies involve predictions across time, they will 0 go beyond a merely psychometric evaluation, into interviews of readth and subtlety, in order to derive more SOphisticated descrip~ of the life situations of the subjects and their typical reaction ns. inally, it should be emphasized again that the present study has sed itself entirely to an investigation of attitudes and achieve- n a single, homogeneous sample of single, white, male freshmen of e academic potential. In most other studies where academic poten- as been used to identify a relatively homogeneous sample of stur the emphasis has been on students of high potential (6. g., 16, , 60, 70). Further studies are needed to investigate the breadth ropriate application of the three attitude scales which have been ped from the Student Opinion Survey, or to derive new scales more riate to student samples having different academic potentials. tudies of the stability of the attitude scales within samples are , not that representations of attitudes should remain constant if titudes themselves vary across time. Most useful would be studies aviors other than those reflected in the attitude scale types, as ehaviors vary concurrent with and consequent to changes in subjects' types. as most ambitious research program might envision attacking all problems (and more!) in a large, heterogeneous student pepulation, 8h ‘ the several years of the members' participation in collegiate .ties; the least ambitious program might concentrate on so limited :tion as perfoming a more adequate crOSS-validation of the three Lde scales of the present study, using a second homogeneous sample Lgle, white, male freshmen of moderate academic potential. Chapter V SUMMARY AND CONCLUSIONS ares five—item attitude scales were empirically derived from the ses made to 135 rating-scales by 63 unmarried, white, male college my students who had displayed moderate academic aptitude on the HIState University Freshman Orientation Test Battery (the MSU 1 Placement Test, the MSU Arithmetic Proficiency Test, the "Quanti— ' and "Linguistic" scores of the ACE Psychological Examinations- ution, and the "Vocabulary" and "Comprehension" sections of the 1ding Test). All the rating-scales had been selected by the writer “as other counseling psychologists as legically representative of :ally-experienced aSpects of college life. Each of the scales met a for unidimensionality, and scalability was largely confirmed in alidation. e attitude scales were labelled as representing: (a) Institur Identification, the extent to which a student considers the acar nterprise to be compatible with his own needs and aspirations; f-Confidence, the degree to which the student feels comfortable—- hmulated--in his performance of the student role, particularly of 3 public aSpects; and (c) Achievement Valuation, the extent to nrking for recognition and enjoying positions of prestige are con- worthwhile by the student. aparisons were made between students' scores (scale types) on the : scales and their cumulative grade point averages (CGPA) for arms of the freshman year (with a minimum of h2 course credits 85 86 and a minimum of 39 credits earned). No statistically significant 1 ons (coefficient of contingency) were found between the attitudes hievement, either when the attitude scores were combined in various scale or two-scale profiles or when they were individually compared as achievement criterion. The small size of the sample and the mility of the achievement criterion probably contributed to the tally negative results. 1e strongest relationship was that between low (scale types 0 and ierate (scale types 2 and 3), and high (scale types h and 5)‘§§l£- age m (SCS) levels and low (below 2.15), moderate (2.15—2.94), ;h (above 2.5h) CGPA levels. The obtained .325 contingency coef- . was significant at the .12 level of confidence. High achievement ;ociated with moderate self—confidence; moderate achievement, with f-confidence; and low achievement, with high self—confidence. .ge of SCS levels increased efficiency of forecasting CGPA levels . per cent beyond that obtained without such knowledge. Although tistic may not have been fully justified, a curvilinear regression on SCS scale types was computed and reported (§t§=.h73, signifi- greater than zero at the .01 level). en the distribution of three-scale attitude patterns was super- on the low, moderate, high three-by—three contingency table comp SCS level and achievement level, a number of interesting trends d. The trends were not worth testing statistically because of anple size, but may be worth pursuing with further research. The 1 pattern trends failed to support the widely—held contention that hr-achieving students (high-achievers) are inclined toward exag- identifications with authority-figures. Most students with atti- ;terns commensurate with such identification were low-achievers. iky, among the high-achieving students the majority exhibited pat- 87 7 essentially contradictory to such‘ identification. The only students patterns in keeping with the “identification" thesis who were found wing above the lowest level were those with SCS scale types 0 or 1. is suggested that this last group of students is perhaps more typical sunseling center and psychological clinic clientele than of students eneral, thus possibly accounting for the inclination of many clinically- .ned personnel to predict high achievement for most students with strong lencies toward identification with authority-figures. An attempt was made to explain the trends in terms of the concepts "feelings of personal inadequacy," "defensiveness" v. "tolerance," and 'emature identification" v. "independence and flexibility." Optimal Lievement for the students of the sample was viewed as symptomatic of essentially realistic awareness of and tolerance for self and surround- gs, with no particular concern for the conventional signs and symbols prestige per se. Failure to achieve was seen as a concomitant of: a) premature narrowing of interests (relative to the demands of first ear liberal arts curricular requirements); (b) a straining after rela- ively superficial appearances of academic success as compensation for ‘elt but poorly tolerated personal inadequacies; and/ or (c) genuine satis- faction with the personal status quo such that no need to strive is felt. ?urther research was reCOmmended to test the very tentatively-held and essentially 3d h_o_g inferences of the present study. SELECTED REFERENCES Bendig,A . w. The relation of level of course achievement to stu- dents' instructor and course ratings in introductory psychology. Educ. pgxchol. Measmt, 1953, 13, u37- -uu8 Bendig, A. W. The reliability of self—ratings as a function of the amount of verbal anchoring and of the number of categories on the scaleo i. 8.221. Pgiycholc’ 1953’ 37, 38-h)». Bendig, A. W. Reliability and the number of rating scale categories. Bendig, A. w. Reliability of short rating scales and the hetero- geneity of the rated stimuli. J. appl. Psychol., 19Sh, 38, 167- 170. Bendig, A. W. a Sprague, J. L. The Guilford—Zimmerman Temperament Survey as a predictor of achievement level and achievement fluc- tuatipn in introductory psychology. g. appl. Psychol., l95h, 38, h09- 3- Borow, H. A psychometric study of non—intellectual factors in col- lege achievement. Unpublished doctoral dissertation, Penn. State College, l9h5. Brooks, M. S. a weynand, R. 8. Interest preferences andt effect upon academic success. Social Forces, l95h, 32, h281—285. Brown, W. F., Abeles, N., & Iscoe, I. Motivational differences between high and low scholarship college students. 1. educ. P5201101 , 19514, 145, 215—223. Brown, W. F., & Holtzman, W. H. A study-attitudes questionnaire for predicting academic success. J. educ. ngchol., 1955, ho, 75-8h. . Burgess, Elva. Personality factors in over- and under-achievers in engineering. Penn. State Univer. Abstr. of Doctoral Dissertar tions, 1953, 15, S36—5E3.‘ . Clark, J. H. Grade achievement of female college students in rela- tion to non-intellective factors: MNPI items. g. Egg. ngchol., 1953, 37, 275-281. . Clark, K. E. & Kriedt, P. H. An application of Guttman's new scaling techniques to an attitude questionnaire. Educ. psychol. Measmt, 19148, 8, 215-223 0 i. Crmmach, L. J. Essentials of psychological testing. New York: j Harper & Brothers, 1959. ‘ 88 L _____ .1 89 ronbach, L. J. Studies of the group Rorschach in relation to suc- cess in the College of The University of Chicago. g. educ. PEZChOIc, 1950, ’41, 65-82- LVesta, F. J.: Woodruff, A. D., & Hertel, J. P. Motivation as a predictor of college success. Educ. psychol. Measmt, l9h9, 9, 339-3h8. “ "‘”" nub R. J. Underachieving students of high capacity. J. higher Educ., 1952, 23, 327-330. tax A. L. 0n Guttman' 5 scale analysis. Educ. pgychol. Measmt, l9h8: 8: 313-315 herds, A. L. & Kilpatrick, F. P. A technique for the construction of attitude scales. J. ppl. Psychol., 19h8, 32, 37ur38h. :nichel, 0. The gfigchoanalytic theory 2; neurosis. New York: W. W. Norton, 1 . astinger, L. The treatment of qualitative data by "scale analysis." P§2 ch01 . Bull 3 , 19147 , LL14, lh9‘161 o mgh, H. G. The construction of a personality scale to predict scholastic achievement. J. app1.Psychol., 1953, 37, 361-366. dlford, J. P. Fundamental statistics in psychology_ and education. (2nd ed.) New York: EcGraw-i SU— dlford, J. P. Psychometric cMethods. (2nd ed.) New York: McGraw- Hill: 1951;. .tm ,L. Gn Festinger's evaluation of scale analysis. Psychol. B_:n_ull:, 19,47: M4: hSI-hés. ttman, L. The Cornell technique for scale and intensity analysis. Educ. psychol. Measmt, l9h7, 7, 2h7-280. ttman, L. The basis for scalogran analysis. In Stouffer, S. A., et a1. Measurement and prediction. Princeton: Princeton Univer. Press, 1m. ttman, L. The scalogram board technique for scale analysis. In Stouffer, S. A., tal. Measurement and rediction. Princeton: Princeton Univer .gFress, 19EU. Pp. 91:12E. ttman, L. The utility of scalogran analysis. In Stouffer, S. A., 33 31. Measurement and prediction. Princeton: Princeton.Univer. Press, 1 . p. - ttman, L. 0n Smith's paper on "'Randomness of error‘ in repro- ducible scales." Educ. psychol. Measmt, 1953, 13, 505—511. dley, J. M. & Kennedy, V. E. A comparison between performance on a sentence completion test and academic success. Educ. pgychol. Measmt, l9h9, 9: 6h9-670 it; 9O ;arris, D. The relation to college grades of some factors other than intelligence. Arch. Psychol., 1931, 20, No. 131. larris, D. Factors affecting college grades: A review of the lit- eramue, 1930—1937. Psychol. Bull., l9h0, 37, 125—166. Lerriott, h. E. Attitudes as factors of scholastic success. Univer. 93 Illinois Bureau 9; Educ. 1193». Bull., 1929, No. 147. ollander, E. P. & Bair, J. T. Pre—trainin attitudes toward authorit -fi res as redictors of inade uate motivation amon Neva Anation Cad-55$. Proaect WmBer mil GUI W U. S. Naval School of Aviation Medicine, Pensacola, Florida, 10 Nov. 1952. ollander, E. P. 6;: Hair, J. T. Attitudes toward authority-figures as correlates of motivation among Naval Aviation Cadets. J. 3.221. P5201101.) 195).}, 38, 21‘25. ayt, D. P. ‘a Norman, W. T. Adjustment and academic predictability. J. counsel. Psychol., 19514, 1, 96-99. 1ghes, W. H. Personality traits and the college success of high sclgmolégraduates. California Quart. secondfl Educ. , 1925, l, 22 -23 a andall, M. G. 6:. Smith, B. B. Tables 3: random sampling numbers. London: Cambridge Univer. Press, 1939. .ugh, H. E. The prediction of academic achievement from measures of personality. Unpublished masters dissertation, Univer. of Pittsburgh, 1952. athwohl, w. G. Specificity of over— and under-achievement in college courses. J. appl. Psycho ., 1952, 36, 103-106. iedt, P. H. & Clark, K. E. "Item-analysis" versus "scale anal- ysis." J. app . ngchol., 19119, 33, 1114-121. rd, F. N. Prediction of scholastic achievement from non-cognitive factors. ETS Res. Bull., Princeton: Educational Testing Service, 1950 . :andless, B. R. The Rorschach as a differential predictor of academic success for matched groups of highl superior men. Amer. szcholu 19117, 2, hlh—h15. (Abstract landless, B. R. The Rorschach as a predictor of academic success. :10 8.221. P§ZChOlog 19149, 33, )43‘50- lelland, 1). Co, Atkinson, J. w-o, Claz‘k, no A., &. Lowell) E. L. The achievement motive. New York: Appleton-Century Crofts, 1953. emgg‘, Q. Psychological statistics. (2nd ed.) New York: Wiley, 19 . . . . «3.44- 91 haary, J. P. Some relationships between non-intellectual charac- teEistgcs and academic achievement. J. educ. Psychol., 1953, uh, 21 -22 . iroe, Ruth L. Prediction of the adjustment and academic perform? ance of college students by a modification of the Rorschach method. Appl. P§ychol. Monogr., l9h5, No. 7. 1talt0, Fannie D. An application of the group Rorschach technique to the problem of achievement in college. J. clin. P§ychol., 19146, 2, 2511-260. rgan, H. H. A psychometric camparison of achieving and non— achieving college students of high ability. J. consult. Psychol., 1952, 16, 292—298. ers, R. C. The academic overachiever: stereotyped aSpects. J. 6X2. Educ.) 1950, 18, 229-2380 aorne, R. T. & Sanders, Wilma B. Multiple-choice Rorschach reSpon- ses of college achievers and non-achievers. Educ. pgychol. Measmt, 19119 , 9 , 085-691 0 good, C. E. The nature and measurement of meaning. ngchol. Bull., 1952, h9, 197-237. rrish, J. a Rethlingshafer, Dorothy. A study of the need to achieve in college achievers and non-achievers. J. gen. ngchol., 195k, 50, 209—226. illips, W. S. & Osborne, R. T. A note on the relationship of the Kuder Preference Record scales to college marks, scholastic apti- tude and other variables. Educ. pgychol. Measmt, 19h9, 9, 331-337. ley, Matilda N., Riley, J. W. Jr., & Toby, J. Sociological studies ig scale analysi . New Brunswick: Rutgers Univer. Press, . ssell, H. E. & Bendig, A. W. Student ratings of instructors and course achievement with academic aptitude controlled. Educ. pgychol. Measmt, 1953, 13, 626-635. st, R. M. a Ryan, F. J. The relationship of some Rorschach vari- ables to academic behavior. J. Pers., 1953, 21, hhl—h56. an, F. J. Personality differences between under- and over— achievers in college. Ph. D. thesis, 1951, Columbia Univer. Ann Arbor: Univer. hicrofilms, No. 2857. hofield, W. A study of medical students with the MNPI: III. iergonality and academic success. J. appl. Psychol., 1953, 37, 7" 2' oemaker, H. A. & Rohrer, J. H. Relationship between success in the study of medicine and certain psychological and personal data. J. Assoc. Amer. Egg. Colle es, 19h8, 23, 1-12. 3. 92 Schultz, D. G. & Green, F. F. Jr. Predicting academic achievement with a new attitude-interest questionnaire-II. Educ. pgychol. Measnlt, 1953, 13, ELL-but Smith, R. G. Jr. Reproducible scales and the assumption of nonnal- ity. Educ. psychol. Neasmt, 1950, 10, 395-399 Smith, R. G. Jr. "Randomness of error" in reproducible scales. Educ. psychol. heasmt, 1951, 11, 587—596. Stagner, R. The relation of personality to academic aptitude and achievement. J. educ. Hes., 1933, 26, 6&8-60 Stouffer, S. A. An overview of the contributions to scaling and scale theory. In Stouffer, S. A., 23 El. Measurement and re- diction. Princeton: Princeton Univer. Press, I955. Pp. 3- Suchman, E. A. The logic of scale construction. Educ. psychol. Ivieasmt, 1950, 10, 79-93- Thompson, Grace H. Non-intellective factors and grades: The group Rorschach. Amer. ngchol., l9h7, 2, hlr. (Abstract) Thompson, Grace N. College grades and the group Rorschach. J. 8.221. P520h01., 19,48, 32) 398-)4070 Tiebout, H. M. The misnamed lazy student. Educ. 322., l9h3: 2h, Weigand, G. Goal aspiration and academic success. Personnel Guid. J-: 1953: 31, h58-h61. Wrenn, C. G. a Number, W. J. Study habits associated with high and low scholarship. J. educ. ngchol., l9hl, 32, 611—616. Young, C. W. & Estabrooks, G. R. Non—intellective factors related to scholastic achievement. Psychol. Bull., 193A. 31, 735-736. APPENDIX A Fifty Polar Adjective and Present-participle Combinations l. active-passive 2. autocratic-democratic 3. bold-shy u. calm—excited(ing) 7. careful-careless 5. clear-vague Z. confident—timid 3. consistent-inconsistent 9. eager-reticent J. easy-difficult L. encouraged(ing)-discouraged(ing) Z. enthusiastic-apathetic 3. exhilarated(ing)-depressed(ing) 4. fair-unfair >. flexible-rigid S. generous-demanding F. goodrbad I. happy-sad ?. healthy-sick '. helped(ful)-hindered(ing) .. honest-dishonest .. honorable-dishonorable 1. intelligent-stupid .. interested(ing)—bored(ing) . kindpcruel ; , 33. 3h. 35. 36. 37. 38. 39. hO. bl. h2. h3- uh. h5. h6. h7. h8. h9. so. liberated(ing)-restricted(ing) lively-dull lovable-hateful loyal-disloyal mature-immature necessary-unnecessary objective-subjective Optimistic—pessimistic orderly-chaotic organized-disorganized patient—impatient pleasant (pleased) - unpleasant (diSpleased) realistic-unrealistic reasonable—unreasonable relaxedrtense rewarded(ing)-punished(ing) sensible-senseless sensitive-insensitive stimulated(ing)-suppressed(ing) strongdweak tolerant-intolerant trusting-suspicious useful-useless vaanble-worthless well-adjusted - neurotic APPENDIX B The Student Opinion Survey On the pages that follow you will find incomplete sentences at the 50p, and sets of Opposing attitudes or traits be10w. Each pair of Oppo- sites is separated by six spaces. Each page will look something like his: Zthink most children are . . . . Some— Some- , , Always Often times times Often Always 1 2 3 h S 6 96. weak ====== ====== ====== ====== ====== ====== strong 1 2 3 h 5 6 97. hateful ====== ====== ====== ====== ====== ====== lovable l 2 3 h 5 6 98. passive ====== ====== ====== ====== ====== ====== active 1 2 3 h 5 6 99‘ realiStj-c ====== ====== ====== ====== ====== ====== unrealistic l 2 3 h S 6 00. loyal ====== ====== ====== ====== ====== ====== disloyfl on are being asked to do two things: you want to com the sentence at e tOp of the page,m reference to each set of Opposites. 2. As you make each of these decisions, also make a judgment as to whether you think the direction you have chosen applies “Some- times," "Often," or "Always." (Try to think of the step from "Sometimes" to "Often" as being equal to the step from "Often" to"Mwwsfl) On the answer sheets, darken the space apprOpriate to your two deci- Lona—for each of the sets of Opposites as they apply to the incompleted antence at the tOp of the page. Darken only 229 space £23 each ppp_p£ >posites Egg d9 p23 skip gpy 3: Egg sets. Turn your "Background - Answer Sheets" packet over and practice on umbers 96—100 for the incomplete sentence and Opposites as they appear 1 the above sample. For example, suppose you decide that, "I think most Lildren are . . . weak," (rather than "strong"). Then you decide that Lay are "Often" weak (rather than "Always" or "Sometimes"). You would eren space number 2 Opposite 96 on the answer sheet. Do not mark in the 1. On the basis of yfi rfirst impressions, decide in which direction P9 iEE booklet! Do ratings 97-100 as you wish. Are there any questions? Please be particularly careful to read and follow the instructions the bottom and tOp of each page of the test booklet. 22 not continue : frwmn one page 22 the next without changing your position pp the answer eets pp instructed! ; if .4 l think most instructors are . . . Always Often l 2 1o immature ====== ====== l 2 2. unreasonable ====== ====== l 2 3o relaxed ====== ====== l 2 be rigid. ====== ====== 1 2 50 hindering ====== ====== 1 2 5. autocratic ====== ====== l 2 ?. organized ====== ====== l 2 3. unrealiStiC ====== ====== l 2 ). stimulating ====== ====== 1 2 ). Stupid ====== ====== 1 2 o boring 3332:: ====== l 2 T. inCOnSiStent ====== ====== l 2 o SUbjeCtive ====== =a==== l 2 -o fair =====3 ====== l 2 . encouraging =====a ====== rn.this answer sheet over and begin with Some- times 3 -’-'-——— -—‘—--. -—-—-.. —————’ -’---— -—O-_—— ——.—-—I— number ——¢—-—‘ —— -—— --=-_— -——_-- —_——-- —----- -‘c-o--- -—.'-_¢-- 9—...- ——-—.—---0 .—_¢-- .--—-o—- -- _ — -- w 95 disorganized realistic suppressive intelligent interesting consistent Objective unfair discouraging 96 w éegin with number 51. ihen I am studying, I feel . . . :1. suppressed ====== ====== ====== ====== ====== ====== stimulated == ====== excited 3. liberated ====== ====== ====== ====== === == ====== restricted 1 2 3 h 5 6 h. enthusiastic ====== ====== ====== ====== ====== ====== apaxhetic 2 3 h S 6 5. chaotic ====== ====== ====== ====== ====== ====== orderly == ====== sad (0 cal‘eflll ====== ====== ====== ====== ====== :===== careless 1 2 3 h 5 6 3. clear ====== ====== ====== ====== ====== ====== Vague l 2 3 h 5 6 3. interested ====== ====== =a=a== ====== ====== ====== bored 1 2 3 h 5 6 l. tense ====== =a==== ==s=== =a:=== =uu=== ====== relaxed 5 6 . useful == == ====== =====z useless . stupid =====g == === == an. n===== ==n=== =3: == intelligent l 2 3 it 5 6 . discouraged ====== ====== ====== ====== ====== =3==== encouraged . rewarded ====== ====== ====== ====== === == ====== punished == reluctant . eager == 3p! Turn to answer sheet 2, and begin with number 1. egin with number 1 on answer sheet 2. aving a college uegree would be . . . S .- Always Often t3; 2 3 1. realistic ====== ====== ====== 1 2 3 20 good ====== ====== ====== 1 2 3 3. unnecessary ====== ====== ==:::: 1 2 3 h. suppressive ====== ====== ====== 1 2 3 S. discouraging ====== ====== ====== 1 2 3 5. hlndering ====== ====== ===:=== 1 2 3 7. elating ====== ====== ====== 1 2 3 3. useless ====== ====== ====== 1 2 3 9o punishing ====== ====== ====== l 3 )o reasonable ====== ====== ====== 1 2 3 . senseless ====== ====== ====== 1 2 3 ). boring ====== ====== ====== 2 3 ’- stupid ====== ====== ===us== 1 2 3 In Valuable ====== ====== ====== 1 2 3 :0 pleasant ====== ====== ====== ,op! Go back to answer sheet 1 and begin Some times Often h 5 “E _____ 5-- 121,15, _-f:- 15-- h S with number 16. 977 unrealistic bad necessary stimulating encouraging helpful depressing useful rewarding unreasonable sensible interesting intelligent worthless unpleasant Begin with number 16 on answer sheet 1. I think most textbooks are . . . Always Of t en 33:: Egg; Often Always 1 2 3 h S 6 16. stimulating ====== ====== ====== ====== ====== ====== suppressive 1 2 3 h S 6 L7. Vague ====== ====== ====== ====== ====== ====== clear 1 2 3 h S 6 L8 . helpful ====== ====== ====== ====== ====== ====== hindering 1 2 3 , u 5 6 L9. easy ====== ====== ====== ====== ====== ====== diffic‘ult 1 2 3 h S 6 :0. disorganized ====== ====== ====== ====== ====== ====== organized 1 2 3 L; 5 6 :1. worthless ====== ====== ====== ====== ====== ====== valuable ' 1 2 3 h S 6 :2. consistent ====== ====== ====== ====== ====== ====== inconsistent 1 2 3 I; 5 6 3. unrealistic ====== ====== ====== ====== ====== ====== realistic 1 2 3 S 6 b. pleasant ====== ====== ====== ====== ====== ====== unpleasant 1 2 3 h 5 6 5, senseless ====== :===== ====== ==:=== ====== ====== sensible . 1 2 3 h 5 6 be lively ====== ====== :===== ====== ====== ====== dull 1 2 3 h S 6 7. useless ====== ====== ====== ====== =:==== ====== useful 1 2 3 h S 6 3. unnecessary ====== ====== ====z= ===z== ====== ====== necessary 1 2 3 h 5 6 ?. boring ====== ====== ====== ====== ====== ====== interesting 1 2 3 h 5 6 ). subjective ====== ====== ====== ====== ====== ====== obj active ,Op! Turn this answer sheet over and begin with number 06. aegin with number 66. lust before a test, I feel . . . Always Often fig; 153%; Often Always , 1 2 3 l1 6 6. pleasant ====== ====== ====== ====== ====== ====== unpleasant 1 2 3 h S 6 q. stimulated ====== ====== ====== ====:= ====== ====== suppressed 1 2 3 h S 6 r8 . encouraged ====== ====== ====== ====== ====== ====== discouraged 1 2 3 h 5 6 9. active ====== ====== ====== ====== z===== ====== passive 1 2 3 h S 6 O. rigid ====== ====== ====== ====== ====== ====== flexible 1 2 3 h 5 6 1. pessimistic ====== ====== ====== ====== ====== ====== Optimistic 1 2 3 u S 6 2. confident ====== ====== ====== ====== ==:=== ====== timid 1 2 3 h S 6 3. bad ====== ====== ====== ====== ======= ====== good 1 2 3 5 6 i. disorganized ====== ====== ====== ====== ====== ====== organized l 2 3 h S 6 - ;. senSitive ====== ====== ====== ====== ====== ====== insensitive l 2 3 b S 6 ). calm ====== ====== ====== ====== ====== ====== excited I 3 h 5 6 '. intelligent ====== ====== ====== ====== ====== ====== stupid 1 2 3 h 5 . Shy ====== ====== ====== ====== ====== ====== bold 1 2 3 h S 6 . relaxed ====== ====== ====== ====== ====== ====== tense l 2 3 h S 6 . Vague ====== ====== ====== ====== ====== ====== clear op! Turn to answer sheet 2 , and begin with number lb. 3egh1with number 16 on answer sheet 2. iofldng for recognition from others is . . . >pl unrealistic senseless immature stimulating valuable elating encouraging rewarding unpleasant stupid necessary unreasonable confident unfair Go back to Always Often 2 answer sheet 1, Some- times ——..._=.—= = "— f. -— L1 = Some- times h ====== Often Always and begin with number 31. 100 , realistic difficult sensible mature suppressing worthless depressing discouraging punishing pleasant intelligent unnecessary reasonable Begin with number 31 on answer sheet 1. I think mOSt academic requirements are . . . 31. 32. 33- 3h. 35. 36- 37. 38. ’43 o ht. LS. StOp! Always l useful ====== l intelligent ====== l inconsistent ====== l punishing ====== sensible ====== l worthless ====== l orderly ====== 1 helpful ====== l unnecessary ====== 1 reasonable ====== 1 fair ====== 1 easy ====== l realistic ====== 1 Clear ====== l encouraging ====== Turn this answer sheet over ome— Some- times times 3__ "it- .3___ “it and begin with ——_....—_ ==—"‘=‘——"‘ number 81. useless stupid consistent rewarding senseless valuable chaotic hindering necessary unreasonable unfair difficult unrealistic vague discouraging Begin with number 81. when called on to Always 1 81. eager ====== 1 82. sensitive ====== l 83. interested ====== 1 81]». Sad ====== l 85; intelligent ====== 1 86. timid ====== l 8?- apathetic ====== l 88. bOld ====== l 89. orderly ====== l 90. Clear ====== l 91- eXCited ====== l 92. tense ====== 93o elated ====== 1 9h. organized ====== 1 9S. immature ====== Stop! L__ contribute to a class discussion, I feel . . . Often 3%; __§_-_ __§__- -3 _____ 3__- __E _____ 3__- m..- Ema - 3__- __3___ __E _____ §___ __3 _____ 3__- 2 __§__ __§__- -3- _ __§___ _-§- 2 3 _ -_2 __ _~__ _ 3--- _-3 _ __2__ _ 3__ 102 Some— , times Often Always h S 6 ====== ====== ====== reluctant h 5 6 ====== ====== ====== insensitive h S 6 ====== ====== ====== bored h S 6 ====== ====== ====== happy h 5 6 ====== ====== ====== stupid h 5 6 ====== ====== ====== confident h S 6 ====== ====== ====== enthusiastic h S 6 ====== ====== ====== shy h 5 6 ====== ====== ====== chaotic h S 6 ====== ====== ====== vague S 6 n===== ====== ====== calm h S 6 ====== ====== ====== relaxed h 5 6 ====== ====== ====== depressed h 5 6 ====== ====== ====== disorganized h 5 6 ====== ====== ====== mature Turn to answer sheet 2, and begin with number 31. 103 Begin with number 31 on answer sheet 2. Being an outstanding success would be . . . Some— Some— Always Often times times Often Always 1 2 3 h S 6 31. depressing ====== ====== ====== ====== ====== ====== elating 1 2 3 h 5 6 32 . diSl-lonorab'le ====== ====== ====== ==r==::= ====== ====== honorable 1 2 3 h 5 6 33. calm ====== ====== ====== ==1=== ====== ====== eJch-ting 1 2 3 h. 5 6 3h. mature ====== ====== ====== ====== ====== ====== mature 2 3 h 5 6 35. Stupid ====== ====== ====== ====== ====== ====== intelligent 1 2 3 b. 5 6 36. sensible ====== ====== ====== ====== ====== ====== senseless 1 2 3 h S 6 37 . unpleasant ====== ====== ====== ====== ====== ======= pleasant 1 2 3 h 5 6 38. difficult ====== ====== ====== ====== ====== =====g easy 1 2 3 S 6 39. punishing ====== ====== ====== ====== ====== ====== rewarding ’ l 2 h S 6 ho. lively ====== ======= ====== ====== =====I ====== dull 1 2 3 h 5 6 m. gOOd =8==== ====== ====== ====== ====== ====== bad 1 2 3 5 6 142 . restricting ====== ====== ==a=== ====== ====== ====== liberating I 1 2 3 S 6 ’43. interesting ====== ====== ====== ====== ===:== 3===== boxing _ 1 2 3 h 5 6 M. suppressive ====== =====u ====== n===== ===n== ====== stimulating l 2 3 h 5 6 I45 . Valuable ====== ====== =B===B ======= ====== =====- worthless This is the end. Thank you! GRADE 0R FLA‘S " 2 4 8 NAME or TEST “RT 1 Prlnted by the IMernatlonal Buslness Machines Corporatlon, Endlcon. N. Y., U. S. A. 1 15 1 3 4 5 6 8 9 10 11 12 13 14 15 I 21:12 2:12: 2:222 222:: :22: 26 222:: 2:22: 1:22: 22:22 212:: 2:222 222:: 222:: 22222 2:22: 221:: :22: 22:2: , \ 11 12 13 14 15 1 2 3 5 6 7 8 9 10 11 12 13 14 15 I 22:2: 222:: 22:22 2:21: 2:22: 27 122:: 1:21: 22:22 1:22: 1:122 121:: 221:: 222:: 2:12: 2122: 222:: 2211: 22:2: 112:: 11 12 13 14 15 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 '1 12 13 14 15 22:2: 12:2: :2... 32:12 22:22 29 3212: 1:222 222:: 2:12: 212:: 11222 21122 222:: 2:12: 2:21: 1:21: 1:21: 22:22 22:2: 22:2: 11 12 13 14 15 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 HEAVY AND BLACK. ERASE COMPLETELY ANY ANSWER YOU WISH TO CHANGE. 1 1 1 13 14 15 10 11 12 13 14 15 ‘ I: I Z I: :Z Z I: Z 31 12:2: 12:2: 12:2: 212:: 2:22: 2:11: ; MIDDLE IBM FORM |.T.S. 1100 A l51-2 .2. .2. 5..5 ..2..._5.1._5.. ..2. .5.. ‘ 1. ...2. .2. 2.5.. 2 3 80 - ERASE COMPL 105 APPENDIX D may 29, 1956 Dear Mr. You are one of sixty-eight Michigan State men who are receiving let- ters like this today. Out of my personal account, I will give a ¢20 prize to one of these sixty-eight men, a $10 prize to another, on the basis of a drawing to be held before interested members of the group at 9:00 PM, Monday, June 11, in the Counseling Center waiting room, on the second floor of the Basic College Building. Offering these prizes is the best way I have of making it worth your while at this busy time of the year to participate in a research study which I am conducting. Participation in the study will take only one hour of your time and involves a paper and pencil method requiring only slight concentration or thought. In short it is painless, and even interesting. Because your name, and those of the sixty-seven others, has been very carefully selected out of the total group of men who first came to State last fall, it is necessagy that ygu bring this letter and your 12 card EE proof g£_your 1 en ity when you come 33 participate. Sessions will be held at 7:00 PM and 8:30 PM in the basement of the Basic College Building each of these evenings: June 1, h, 5: 6: and 7. Saturday, June 2, sessions will be at 9:00 AM and 10:30 AM. If none of these times is convenient for you, please call me at the University, Extension 2567, and I will be glad to arrange a time suitable for you. The drawing for the $20 and $10 prizes will be from among signed slips which you will be able to submit when you have completed your part in the study. Only those men 322 actually artici ate will be eli ible for the prizes, ut you mega—not be present a e raw1ng tE—Win. you win ut are not present, a check will be mailed to you at your home address. Not onLy do you have a good chance of winning one of the prizes, you will be assisting in research which may benefit you and other students in the years to come. I will very greatly appreciate your c00peration in this study, and look forward to working with you. Sincerely yours, Walter R. Stevens Counselor APPENDIX E Background - Answer Sheets Name: BirthDAIE: Where did you graduate from high school? When? karital status (underline one): Never married, harried, Separated, Divorced, Other If you have worked while here in college, indicate 0pp05ite the prOper terms below the approximate average number of hours per week worked: Fall '55 Winter '56 Spring '56 Had you earned any college credits before coming to State?______ How many? What is your major field? Have you changed your major since coming to State? From what! Have you ever taken Psychology 101, "Methods of Effective Study"? When? Why? If you are married, for how long? Are you a veteran? If you have taken any of the Improvement Service courses here, underline the prOper term below, Opposite the course(s) you took, and briefly give the reason for your taking it: Reason: Arithmetic Improvement: Fall, Winter, Spring Reading Improvement : Fall, Winter, Spring Speech Improvement : Fall, Winter, Spring Writing Improvement : Fall, Winter, Spring Concerning your parents: Father mother are they living? Extent of their education years years Educated mostly in U. 5.? Are they still living together? If not, how old were you when the family was broken! , and with whom did you live subsequently? Indicate your position among your brothers and/or sisters. (lndicate brothers with a "B,” sisters with an "S," and yourself with an "h.") youngest oldest ) APPENDIX F Background — Answer Sheets Note: All of the information asked for below will be held completely confiden— tial. It is for my ersonal use in eValuating the results of the Opinion survey which follows. In no instance will it be given to any member of the Psychology Department or to any other hichigan State University staff member or student. It will not be a part of any file. Name: Your high school: Are you a veteran? when did you first come to State? rupted at any time? What is your major field? since coming to State? From what? When? Sex: BirthDATE: Town: Grad. when? If you are married, since what term in college? When? For how long? Has your college education been inter- Have you changed your major Why? How did you first hear of Psychology 101? when? What (or who) actually led you to enroll in this course? If someone else urged you to take it, do you think you were yourself generally in favor of it or Opposed to it? What do you now think of the course? If you have worked while here in college, indicate below the appropriate average number of hours per week worked in each tenn you have been here: Concerning your parents: Father Mother Are they living? Extent of their education? years Educated mostly in U. S. (or other English-Speaking country ? Are they still living together? was broken? , and with whom did you live subsequently? years If not, how old were you when the family Indicate your position among your brothers and/or sisters. (Indicate brothers with a "B," sisters with an "S," and yourself with an "E.") youngest ..-. .me oldest APPENDIX G Background — Answer Sheets Note: All of the information asked for below, and that you will be giving in reply— ing to the Opinion survey which follows, is for my own personal use. In no instance will your replies be available to any other member of the Psychology Department or to any other SDSC staff member or student, except in the form of the group results for the total sample of students being asked to participate. Several weeks after the Spring recess I hepe to be able to return the results to you and discuss with you the purpose of the study. Name: Sex: BirthDATE: Where did you graduate from high school? When? Are you a veteran? Are you married?______ when were you married? When did you first come to SDSC? Had you been in college before?______ Has your college education been interrupted at any time?______ When?______ For how long? Reason: What is your major field? Have you changed your major since coming to SDSC? From what? When? Reason: If you have worked while here in college, indicate below the approximate average number of hours per week worked in each term you have been here: lst semester 2nd semester 3rd semester hth semester 5th semester 6th semester Concerning your parents: Father Mother Are they living? Extent of their education years years Educated mostly in U. S. (or other English—Speaking country)? I N Y N If both are living, are they still living together? If not living together, or if one or both are deceased, how old were you when the famiry was broken? , and with whom did you live subsequently? Indicate your position among your brothers and/or sisters. (Indicate brothers with a "B," sisters with an "S," and yourself with an ”M.") youngest oldest APPENDIX H 109 L--... Terms Which Defined the a Priori "Positive" Extreme for Each Item of the Student Opinion Survey, as Determined by the Judges C D O ODKIO‘U‘E‘U’NH O O HHi-‘H wml—‘OW O O O 0 O Attitude Area X mature reasonable relaxed flexible helpful democratic organized realistic stimulating intelligent interesting consistent 'objective fair encouraging stimulating clear helpful easy organized valuable consistent realistic pleasant sensible liveLy useful necessary interesting Objective useful intelligent consistent rewarding sensible valuable orderly helpful necessary reasonable fair easy realistic clear encouraging Attitude Area Y stimulated c liberated enthusiastic orderly happy careful clear interested relaxed useful intelligent encouraged rewarded eager pleasant stimulated encouraged active flexible Optimistic confident good organized sensitive caLn intelligent bold relaxed clear eager sensitive interested happy intelligent confident enthusiastic bold orderly clear calm relaxed elated organized mature Attitude Area Z H ECWCDNOU'IC'WNH O O :55 O! realistic good necessary stimulating encouraging helpful elating useful rewarding reasonable sensible interesting intelligent valuable pleasant realistic easy sensible mature stimulating valuable elating encouraging rewarding pleasant intelligent necessary reasonable confident fair elating honorable exciting mature intelligent sensible pleasant easy rewarding lively good liberating interesting stimulating valuable 110 APPENDIX I Items Approximating Desired Scale Distributions in Original Forty-five- item Eatrices; Also Data on Which Final Item Selection Was Based (A) (B) Item "Positive" 5 of $5 No. of 35 No. of Ratio No. categories "positive" beyond cut* errors* B/A h, 5, 6 65 25 13 .52 , 5h 19 18 .95 : 6 70 2h 17 .71 29 20 .70 h, 5: 6 2h 18 .75 , 56 28 19 .68 u, 5, 6 52 27 20 .7h 5, 6 no 27 18 .67 a 6 59 27 15 .56 )4: 5: 6 29 18 .62 ‘ 5, 6 65 26 12 .h6 5, 6 51 26 17 .65 5, 6 bk 31 12 .39 5, 6 63 25 12 .h8 5, 6 56 22 18 .82 S, 6 Sh 26 15 .58 5, 6 35 13 13 1.00 L, S, 6 h? 21 19 .91 5, 6 67 1h 11 .79 5, 6 28 18 .6h 5, 6 8 23 17 .7h h, 5, 6 3 22 15 .68 h, 5, 6 no 22 15 .68 5: 6 67 30 19 .63 h. 5. 6 no 26 15 .58 5, 6 62 31 17 .55 5, 6 70 18 13 .72 h. S. 6 28 13 . h, S. 6 S9 19 15 .79 ’4: 5: 6 ‘46 28 1’4 .50 , 6 68 20 12 .60 5. 6 A9 29 15 .52 h, 5, 56 20 10 .50 . hl 23 15 .65 (continued on next page) 111 APPENDIX I (continued) No. categories "positive" beyond cut* errors* B/A (A) (B) Item "Positive" % of Ss No. of $5 No. of Ratio Area I (continued): 88 h, 5, 6 56 29 17 .59 89 3 ’41 29 013-8 90 5, 6 38 20 12 .60 9 h: 5: 6 57 23 16 .70 95 LL: 5: 76 15 12 .80 Area Z: h 6 51 28 20 .72 ‘ 8 6 56 29 12 . 12 6 25 13 13 1 .00 '15 6 ul 1h 16 1.1h 18 6 21. 18 9 .50 19 5, 6 56 25 .28 21 5: 6 71 15 13 .87 26 5, 6 5h 23 1h .61 29 5: 6 h0 20 15 .75 31 6 51 2h 16 .67 32 6 52 27 17 .63 \ 3h 6 37 21 10 .h8 36 6 hh 17 15 .88 37 6 29 1b, 11 . 79 ii 6 33 16 13 .81 6 2h 15 8 .53 ‘ * These figures are derived from the original, fortyhfive—item, item-score matrices. The "cut" referred to in Column A is the cutting point in the distribution of respondent rankings in those matrices at which the investigator would alter predictions as to item responses: "positive" being the prediction for all subjects ranked above the cut- ting point, and "negative" the prediction for all subjects ranked below the cutting point. For a fuller discussion, see pages 32 and 33. in . . . — .... . . - ._.~—...... . .... .3. .i . .anaaunoal—f . . _ z: . 112 APPENDIX J Final Item-Score Matrix (5 Dichotomized Items); Attitude Area I Items and scoring weights Subject Unique Scale Total number» Item No.: 15 33 113 35 3).; score type errors Weight : 16 8 1 9 x x x x x 31 5 0 15 x x x x x 31 5 0 16 x x x x x 31 5 O 18 x x x x x 31 5 0 23 x x x x x 31 5 0 146 x x x x x 31 5 0 57 x x x x x 31 5 0 58 x x x x x 31 5 0 63 x x x x x 31 5 0 27 x x x x ( ) 30 5 1 M4 x x x ( ) x 29 5 1 h x x ( ) x x 27 5 l 28 x x E g x x 27 5 l 118 x x x x 27 5 1 5 x x x x 15 h 0 25 x x x x 15 h 0 35 x x x x 15 h 0 37 x x x x 15 )1 O 50 x x x. x 15 h 0 52 x x x x 15 h 0 55 x x x x 15 h 0 2 x x x ( ) 1h h 1 119 x x x ( g 11; h l 7 x x ( ) ( 12 h 2 2h x x x 7 3 0 32 x x x 7 3 0 62 x x x 7 3 0 3 (x x x x 23 3 l 38 (x x x x 23 3 l 39 x x x x 23 3 1 51 (x x x x 23 3 1 11; x x ( ) 6 3 l 22 x x ( 3 6 3 1 115 x x ( 6 3 1 26 x x 5 3 1 31 x x 5 3 1 61 x x 5 3 1 (continued on next page) APPENDIX J (continued) 113 Items and scoring weights Subject Unique Scale Total number* Itan No.: h5 33 h3 35 3h score type errors Weight : 16 8 h 2 l 21 x x 2 0 h2 x x 2 0 l (x) x x 19 2 1 17 (x; x x 11 2 l 19 (x x x 11 2 l 12 x l l 0 29 x l 1 0 A7 x 1 1 0 56 x 1 1 0 6 ix; x 17 1 1 11 x (x) x 25 l 2 : I 8 0 O 0 10 0 O O 13 0 0 o i 20 O O O 33 o o 0 31, o o o 6 o o o , 0 0 O O bl O O 0 A3 0 0 0 53 o o 0 5b. 0 o o 59 o o o ! 60 O 0 0 30 (x) (x) 2h 0 2 l Total errors: 8 h 3 5 7 27 5 error : 12.7 6.h h.8 7.9 11.1 8.6 * Subjects are numbered according to their CGPA rank, with 1 being assigned to the student with the highest CGPA and 63 assigned to the Where subjects were tied for a particular CGPA, they were numbered in keeping with the alphabetical order of their last initials. In Appendixes M and N subjects will be found ranked in student with the lowest CGPA. descending order. as: 1. . 0...! 1 I... APPENDIX K Final Item—Score Matrix (5 Dichotomized Items); Attitude Area'I Items and scoring weights Subject Unique Scale Total number* Item NO.: 93 89 8h 86 SD score type errors Weightzlé 8 2 l I x x. x x x 31 5 0 x x x x x 31_ 5 0 x x x x x 31 5 0 x x x x x 31 5 0 x x x x x 31 5 0 x x x x x 31 5 0 x x x x x 31 5 0 x x. x x x 31 5 0 x x x x x 31 5 0 x .x x x x 31 5 O x x x x ( 30 5 1 x x x x ( 30 5 1 x x x x ( 30 5 l x x ( ) x x 27 5 1 ‘ l x x. x. x 15 h 0 x x x x 15 h 0 x x x x 15 h 0 x x ( a x, 13 h 1 ' x x( () 12 h 2 X X X 7 3 0 I x x x 7 3 0 x x x 7 3 0 l (X x x x 23 3 1 x x x x 23 3 1 i (x x x x 23 3 l ' x x. ( ) 6 3 1 5 " x E 3 i ; x x x H x S 3 l (x) x x ( ) 22 3 2 (x) x ( ) x 21 3 2 I (continued on next page) APPENDIX K (continued) Items and scoring weights * Subjects are numbered according to their CGPA rank, with 1 being student with the lowest CGPA. Where subjects were tied for a particular CGPA, they were numbered in keeping with the alphabetical order of their last initials. In Appendixes M and N subjects will be found ranked in descending order. Subject Unique Scale Total number* Item No.: 93 89 8h 86 5h score type errors Weight : l6 8 h 2 l 10 x x 3 2 0 ‘ 33 x x 3 2 o l h3 x x 3 2 0 1 as x x 3 2 0 T 50 x x 3 2 0 1 12 (x) x x 19 2 1 1 8 Ex) x x 11 2 l 28 x x x 11 2 1 h6 (x) x x 11 2 1 ho (x) x ( ) 18 2 2 7 x 1 l 0 13 x 1 1 o 22 x l l 0 27 x 1 l 0 31 x l l 0 3h x l l O 7 x l 1 0 5 (x) x 9 1 1 20 0 0 0 29 O O 0 39 0 o o 112 o 0 o 5 O 0 0 S 0 0 0 56 0 0 0 59 o o o 60 0 O 0 21 (x) 16 O 1 0 (x) 8 O 1 9 (x) 8 0 l 9 (x) h 0 1 35 (X) (X) 21; o 2 Total errors: 9 7 2 6 7 31 % error : 1h.3 11.1 3.2 9.5 11.1 9.8 assigned to the student with the highest CGPA and 63 assigned to the ___—L APPENDIX L Final Item-Score Matrix (5 Dichotomized Items); Attitude Area Z Items and scoring weights Subject Unique Scale Total numberk Item No.: hl 3h 36 19 21 score type errors Weight : 16 8 2 1 x x x x x 31 5 0 x x x x x 31 5 0 x x x x x 31 5 O x x x x x 31 5 O x x x x x 31 5 O x x x x x 31 5 0 x x x x ( ) 30 5 1 x x x ( ) 29 5 l x x ( ) x x 27 5 l x x ( ) x x 27 5 1 x x x ( ) ( ) 28 5 2 x x x x 15 h 0 x x x x 15 b, O x x x x 15 h 0 x x x x 15 b, 0 x x x x 15 h 0 x x ( g x 13 h 1 x x ( x 13 )4 l X x ( ) ( ) 12 h 2 x x x 7 3 0 x x x 7 3 0 x x x 7 3 0 x x x 7 3 0 x x x 7 3 0 (x) x x x 23 3 1 (x) x x x 23 3 l x ( ) x 5 3 1 x x 3 2 O x x 3 2 0 x x 3 2 0 x x 3 2 O x x 3 2 0 x x 3 2 O x x 3 2 0 ‘ (continued on next page) ! . APPENDIX L (continued) I Items and scoring weights Subject Unique Scale Total number* Item No.: bl 3h 36 19 21 score type errors Weight : l6 8 h 2 l i 55 x x 3 2 0 56 x x 3 2 O 57 x x 3 2 0 6 (x x x 11 2 l ‘ 7 2x x x 11 2 l 35 x x x 11 2 1 ‘ 61 x < ) 2 2 1 1 \ l x 1 l 0 3 x 1 l 0 L 11 x 1 1 0 20 x l l O 29 x l 1 0 37 x 1 l O 19 x 1 1 o 1 2h (x) x 9 l O 2 0 0 0 8 O O 0 12 O 0 0 1h 0 0 O 23 0 0 0 27 0 0 0 36 0 0 0 51 0 O O 5 o o o 5 o o o 60 O 0 0 21 (x) h 0 1 h (x; (x) 20 0 2 O (x 20 O 2 Total errors: h h 5 6 h 23 % error : 6.h 6.h 7.9 9.5 6.h 7.3 * Subjects are numbered according to their CGPA rank, with 1 being assigned to the student with the highest CGPA and 63 assigned to the student with the lowest CGPA. Where subjects were tied for a particular CGPA, they were numbered in keeping with the alphabetical order of their last initials. In Appendixes M and N subjects will be found ranked in descending order. APPENDIX M Data Summary: Achievement, Intellective, and Attitude Measures Orientation test* Attitude scores r o 5 al t e Subject CGPA aw sc res ( c e yp 5) number E A q L v c 115 505 AVS 1 3.31 18 36 18 62 2o 23 2 3 1 2 3.31 22 39 10 63 22 1 3 o g 3.26 16 35 10 67 21 20 3 3 1 3.21 19 37 10 58 19 18 S 5 S 5 3.00 16 37 15 63 12 19 1 3 2 6 2.98 12 38 1o 56 20 23 1 5 2 7 2.91 19 38 13 62 12 18 1 1 2 8 2.90 13 35 10 62 17 21 o 2 0 9 2.79 17 37 39 56 17 21 5 0 1 10 2.71 18 38 16 61 11 22 o 2 5 11 2.70 17 36 39 61 22 1 3 1 12 2.67 16 26 19 67 17 2o 1 2 0 13 2.65 16 35 11 56 11 23 0 1 1 11 2.60 19 37 61 16 17 3 3 0 15 2.58 18 31 11 51 17 20 5 S 1 16 2.58 17 30 17 63 13 l9 5 5 5 17 2.57 20 33 18 55 20 22 2 5 3 18 2.56 23 3 12 55 16 19 5 1 2 19 2.55 15 O 16 63 17 l7 2 3 5 20 2.51 11 39 15 67 18 18 0 o 1 21 2.53 20 28 11 63 11 18 2 0 0 22 2.52 16 33 38 52 12 20 3 1 2 23 2.51 16 29 11 62 13 17 5 5 o 21 2.51 16 31 39 61 13 17 3 3 1 25 2.17 12 36 17 65 21 18 1 3 5 26 2.16 18 36 13 67 16 22 3 5 1 27 2.13 21 35 15 65 15 17 5 1 o 28 2.10 19 36 11 65 18 19 5 2 1 29 2.38 17 31 38 51 11 18 1 0 1 30 2.33 12 38 17 61 1 17 0 0 1 31 2.31 15 38 17 51 1 22 3 1 5 32 2.31 15 38 18 57 13 18 3 5 2 33 2.29 18 10 17 63 15 17 o 2 2 31 2.29 21 38 16 61 16 17 0 1 0 (continued on next page) 7 .fi, I __. __ 1‘ AFPENDIX M (continued) Orientation test* Attitude scores al t Subject CGPA raw scores (so e ypes) number E A Q_ L v c IIS scs Avs 35 2.21 11 39' 10 55 16 21 1 0 2 36 2.22 11 31 11 57 20 20 o 3 o 37 2.22 21 38 10 63 18 19 1 1 1 38 2.20 15 35 13 62 11 17 3 5 2 9 2.20 17 39 15 60 11 18 3 0 1 0 2.18 16 31 38 52 16 22 o 2 o 11 2.18 17 31 11 53 20 19 o 1 2 12 2.15 11 37 39 20 19 2 0 3 13 2.11 11 27 11 67 17 17 0 2 5 11 2.13 19 35 15 59 19 21 5 5 3 15 2.10 20 35 18 61 13 20 3 1 3 16 2.09 22 1o 11 52 12 19 5 2 1 17 2.08 13 35 15 60 18 21 1 5 3 18 2.07 16 27 11 67 15 17 5 2 5 19 2.01 15 36 17 63 15 23 1 0 1 so 2.01 20 36 16 62 19 2o 1 2 3 51 2.00 16 39 11 58 12 23 3 3 0 52 1.98 12 3 17 62 13 23 1 5 3 5 1.98 16 1 13 61 19 18 o o o 5 1.98 11 30 17 53 13 23 0 0 0 55 1.92 21 39 17 51 12 2o 1 2 56 1.88 15 31 18 63 13 17 1 g 2 57 1.87 20 38 15 53 22 5 1 2 58 1.83 11 3 18 53 1 18 5 3 5 59 1.71 21 36 10 63 1 22 o 0 3 60 1.68 17 31 39 66 15 18 0 0 0 61 1.68 16 32 16 58 18 18 3 1 2 62 1.62 11 37 11 59 21 20 3 5 5 63 1.08 19 31 16 66 20 23 5 1 5 * See page 18 for names of Orientation Tests. 120 APPENDIX N Data Summary: Time Spent in Outside Work, Field of Specialization, and Credits Carried . Aver. outside . Previous Credits 823:2: work load 1:22:21 major carried (hrs. per wk.) field? (1 year) 1 o Pre—M Pre—V 17 2 11 E E N c 18 3 O For N G 51 1 29 Econ N P 19 5 0 Civ E N c 52 6 0 Engin N C 52 ' 7 17 P01 A G B 50 8 o N P N c 12 9 13 D Mfg N c 13 10 0 Pre-V N c 50 11 0 Ind D M E 50 12 O Journ Pol A 19 13 o Hist N c 13 o M E N c 50 15 10 Art N c 18 16 o M E N c 55 17 o E E N c 56 18 0 Pre-V N c 18 19 10 Engin N c 55 20 0 M E N c 50 21 0 N P N c 19 22 o Engin N P 50 1 23 0 Art N c 15 21 ll Civ E. For 19 25 o E E N c 55 26 o G B N c 11 27 21 Pre-V N c 19 28 20 Ag Ed N c 15 29 o Sp Co N P 15 3o 0 Res B N c 19 31 0 G B N c 19 32 10 G B N P 18 33 15 Pre-D N c 15 31 0 G B N c 18 35 0 D Mfg N c 16 36 o M E N c 15 37 0 F Dis N c 19 (continued on next page) p ,1 7 fl 6 APPENDIX N (continued) 121 . Aver. outside . Previous Credits Subject work load M93°r major carried number (hrs. per wk.) fleldfi field* (1 year) 38 0 N P N C 16 39 0 Geol N P 15 10 11 G B N C 50 11 0 Hotel N C 51 12 0 Int R Best 18 13 8 Pre-V N C 13 11 12 G Agr N c 17 15 1 G B N c 19 16 O For Pol A 53 17 O Markt F Dis 19 18 0 P01 S N C 16 19 O M E N C 51 50 7 G B Ph Ed 19 51 12 N P N C 18 52 o M E N c 13 5 3 G B Hotel 19 5 10 Adv N C 19 55 1 P01 A N c 53 56 8 Draft M E 19 57 o N P Engin 51 58 8 Ph Ed G B 18 59 22 N P Gem E 18 60 o Pre-M N c 17 61 0 Hotel N c 50 62 3 N P Pre—M 15 63 O G B Hotel 19 * Key to abbreviations used: Adv = Advertizing F Dis: Food distrib. M E = Mech. engin. Ag ed: Agric. educ. For : Forestry N C : No change Art : Art G Agr: Gen. agric. N P : No preference Bact z Bacteriology G B : Gen. business Ph Ed. Phys. educ. Cem E: Chem. engin. Geol : Geology Pol A: Police admin. Giv E: Civil engin. Hist : History Pol S: Polit. sci. D Mfg: Dairy manuf. Hotel: Hotel admin. Pre-D: Pre—dental Draft: Drafting Ind D: Indust. design Pre-N: Pre—medical Econ : Economics Int R: Internat. relat. Pre-V: Pre-vet. med. E E : Elect. engin. Journ: Journalism Res B: Resid. bldg. Engin: Engineering Narkt: Marketing Sp Co: Speech corr. ”Ni. 1.4 $881 . '. . “7"? [/- 1 Room USE ONLY \ mm USE out}: tIHIWUIWHHNWWIJMHMHQIMl|iH 3 1293