University of Southern California Dissertations and Theses

The Effect Of Roommates On The Scholastic Achievement Of College Students

(USC Thesis Other)

The Effect Of Roommates On The Scholastic Achievement Of College Students

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been microfilmed exactly as received Mic 60-6211 MURRAY, Mary Etta Brookes. THE EFFECT OF ROOMMATES ON THE SCHOLASTIC ACHIEVEMENT OF COLLEGE STUDENTS. University of Southern California, Ph.D., 1960 Education, psychology University Microfilms, Inc., Ann Arbor, Michigan THE EFFECT OF ROOMMATES ON THE SCHOLASTIC ACHIEVEMENT OF COLLEGE STUDENTS by Mary Etta Brookes Murray A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Education) June 1960 UNIVERSITY O F SOUTHERN CALIFORNIA GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES 7. CALIFORNIA This dissertation, written by ......MARY ET TA BROOIGSS MURRAY..... under the direction of h..&£..Dissertation Com mittee, and approved by all its members, has been presented to and accepted by the Graduate School, in partial fulfillment of requirements for the degree of D O C T O R OF P H I L O S O P H Y Dean Date.. . . . J u n < S L . . 1 9 . 6 . Q ............. DISSE^TATION^O^MITTEE ... ( f t I / <r~) Chairman TABLE OF CONTENTS Chapter P^ge I. INTRODUCTION ......................... 1 Purpose of Investigation Importance of the Investigation Delimitations of the Study Scope of the investigation Inherent weaknesses Definition of terms Organization of the Study II. RELATED LITERATURE Introduction Intellectual factors in academic achievement Nonintellectual factors in academic achievement Differential prediction Agriculture prediction Engineering prediction Chapter summary III. METHOD.................................. 73 Outline of procedures Description of variables Statistical procedures Chapter summary IV. FINDINGS................................ 91 V. SUMMARY, CONCLUSIONS, IMPLICATIONS AND RECOMMENDATIONS ............... 99 Summary Conclusions iv Chapter Page Implications Recommendations « APPENDIX......................................... 110 Table 1. Correlation of Matching Variables........................ Ill Table 2. Comparison of Sample Groups on Matching Variables ............. 112 Table 3. Comparison of Sample Groups on Entrance Clause.............. 113 BIBLIOGRAPHY ................................... 115 CHAPTER I INTRODUCTION The utilization of the talents of college ability students is a matter of critical concern to those respon sible for the guidance and training of our future profes sional and technical workers. In an age of rapid technological development, it is vitally important to society that young men and women who can achieve success in college be identified and encouraged to enter insti tutions of higher learning. It is important, too, that these students be assured space in a period of overcrowded facilities and increasing enrollments. To accomplish these goals, it would seem that more selective screening procedures for college admission as well as additional information about factors contributing to success are needed. In order that achievement expectancies may be determined with greater reliability, more accurate predic tive measures must be found. In earlier studies intellec tual screening alone was used to forecast school 1 performance. In recent years, however, wide interest has developed in various non-cognitive factors and their contributions to achievement. Many investigators have reported significant findings as a result of their research in the area. Since environment is one of the nonintellectual factors receiving the most attention, it would seem that if college environmental influences were investigated, they might be found closely related to scholastic achieve ment . This study was concerned with the area of college environment, and particularly with the affect of one aspect of campus residence hall living on scholarship. Purpose of Investigation The purpose of this investigation was to determine the extent to which roommates may influence the scholastic achievement of college students. The following hypotheses were advanced: 1. A student's grades are likely to deviate from expectancy, above or below, in the same direction as those of his roommate. 2. The grades of students who share rooms will show a greater deviation from expectancy than the grades of students who occupy single rooms. Importance of the Investigation Previous studies have shown that environment is one of the primary nonintellectual factors in the deter mination of success. If the effects of various aspects of the environment on scholastic achievement could be determined, more reliable predictions of academic success might result. Since residence halls will become more important centers of student environment as colleges continue to expand their programs and provide campus living facilities for greater numbers of students, a study of the effects of these facilities on achievement might be of great value In helping students attain success. Of the various intra-hall factors which might affect achievement, the influence exerted by a roommate could be a major one. If a relationship were found to exist between college grades and roommate influence, new methods of room assignments might be devised whereby students could be helped to reach or exceed grade expectancy. 4 Delimitations of the Study Scope of the Investigation The study was limited to the California State Polytechnic College in San Luis Obispo, California. The period of time involved consisted of the fall quarters of the 1955-56 and 1956-57 school years. During these years the college had a student population of slightly over 4,000. Approximately 800 students lived in campus residence halls, of which number about one third were freshman men. Since, the halls to which the freshman men were assigned had accommodations particularly suited to the present study, this group was selected as the sample for the investigation. Two quarters were used in order to gain a larger sample, and the first quarter of each year was chosen because the initial experience with residence hall living was desired. It seemed likely that a period of practice with the environmental influences might effect some change in behavior, thereby making the sample less similar. In September, 1955, 275 freshman men lived on the campus; and in September, 1956, 276. Of the total 551, 27 did not fully meet the residence criterion, so the final sample was reduced to 524. Women and upper class men students were excluded because residence halls for these groups had only double rooms. Students occupying rooms alone were needed for a control group, so the investigation was limited to freshman halls. In these units both single and double rooms were available. Also eliminated from the sample were eight freshmen who either withdrew from school or changed rooms during the quarter, and nineteen who shared rooms with upper classmen. Of the 524 in the total sample, 286 occupied double rooms; and 238 occupied single rooms. Inherent Weaknesses 1. The study was limited to one college campus. A broader base, such as a survey of all state colleges, was not possible because of the lack of comparable facili ties at other institutions. In the few state colleges which provided for resident students in 1955 and 1956, similar accommodations for single and double occupancy did not exist. 2. The sample consisted primarily of agriculture and engineering majors. Thus a typical cross-section of fields of specialization was not available. 3. The elimination of women from the study pre vented a determination of any sex differences which might exist. 4. Inter-class differences could not be deter mined since all subjects were freshmen. 5. From 5 to 10 per cent of all roommate assign ments were made in accordance with special requests of students. Thus the role of chance was not applicable to this group. 6. Time spent in paid employment or organized activities was not considered in matching related pairs. However, an effort was made to match the experimental and control subjects in number of units carried. Varia tions within a pair were seldom greater than three units and never greater than six units. Drive and motivation, important contributors to prediction, were not considered. Definition of Terms The following terms are defined as used in this study: Freshman - a student for whom the period under study was the first quarter or semester of college attendance. < . Cognitive factors - mental factors as measured by ability or achievement tests. Intellectual factors - synonymous with cognitive. Non-cognitive factors - factors having to do with areas of personality, environment, attitudes, age, use of time, high school variables, and physiological- data. Nonintellectual factors - synonymous with non- cognitive. Resident - a student who resides in a college residence hall on the school campus. Grade point average - the point on a scale of 3.0 that is derived by dividing the number of grade points earned by the number of units attempted. Entrance tests - placement tests administered before registration for the purpose of determining appropriate class level. Organization of the Study In Chapter II is presented a review of related literature. Because of the preponderance of material available on the subject of predicting academic success, certain selection procedures had to be employed. The review, therefore, is limited to the prediction of achievement at the college level. Within this framework nonintellectual factors are stressed. Chapter III details the locale of the study and the methods used in selecting the sample. Also described are the procedures for gathering the data and testing the hypotheses. Chapter IV reports the results of the investiga tion and lists the findings. Chapter V presents conclusions which were drawn from the findings, summarizes the study, and lists recommendations for further research. CHAPTER II RELATED LITERATURE Introduction During the last fifty years, the prediction of academic success has been extensively studied. The research falls largely into two categories: (1) the effect of intellectual factors on scholastic success, and (2) the effect of nonintellectual factors on achieve ment. The former has been under investigation for the longer period of time and seems to have more conclusive findings. However, the latter is rapidly gaining in importance since the previous, more objective data are limited in their predictive efficiency. Since the literature appears to include a volumi nous number of studies on academic prognosis, any review of the subject must be highly selective. The present study was concerned with higher education., and the litera ture surveyed was limited to prediction studies dealing primarily with students in undergraduate college and 10 university programs in the United States since the turn of the century. The intellectual factors considered were intel ligence and achievement tests, high school grades, high school rank, and college admission examinations. Non intellectual factors included the results of various personality inventories and attitude surveys, subjective ratings of teachers and principals, and data gathered from cumulative records and college application blanks. The literature cited in the present chapter is outlined in the order stated above; studies on intellectual factors are presented first, followed by those on non intellectual factors. Studies on differential prediction, and research in the prediction of success in the specific fields of agriculture and engineering are also surveyed. Intellectual Factors in Academic Achievement The earliest device used for the selection of students was an achievement of content type examination. Although this type was used during most of the last century, the validity of the results for screening purposes is not known. No studies on the subject are available. 11 In 1871, a system of high school accreditation began in Michigan. High schools on the list were expected to recommend for higher education only those students who were in the upper range scholastically. Thus was begun the custom of admitting students on the basis of high school credits. At the turn of the century the College Entrance Examination Board was established, and the use of intel ligence tests was beginning. Between 1905 and 1920 many additional intelligence tests appeared, and, in 1920, surveys began to appear about their predictive value. The United States Bureau of Education compiled one of the first summaries, which covered sixty-two studies at thirty-six institutions and reported a predictive validity in over two thirds of the studies of .30 to .50 (103). From 1920 to 1934, most of the predictions were attempted on the basis of single factors. These consisted of high school marks, achievement examinations, or intel ligence or aptitude tests. In a summary of the literature in 1931, Douglass reported a range of .29 to .77 and a median of .545 for correlation coefficients of high school marks and college 12 marks (113). In a review of sixty-seven studies, he found the correlation of achievement examinations with college marks to have a mean of .54 and a median of .55. A review of 250 studies showed a median correlation of .45 between general aptitude tests and college grados. Kinney, in 1932, made a survey of studies which compared measured intelligence with college attainment, and reported a median of .44 (160) . Segel summarized the literature in 1934, with similar results (208). He reported a mean of .54 for correlations between high school and college marks, and a median of .545 for thirteen studies comparing achieve ment examinations with college records. Surveying 100 studies correlating general aptitude tests with collage grades, he reported a median of .44. Segel concluded that for academic prediction, general achievement tests are best, general intelligence tests next, and specific tests third in value. Wagner, in a review of eighty-eight studies in 1934, reported a median of .56 when correlating achieve ment tests with college marks (232). She also obtained a median coefficient of .40 to .50 between general 13 aptitude tests and college records. Even after 1934, many studies were based on single factors. In a survey of correlation studies in 1943, Durflinger reported a central tendency of .50 to .60 and a median of .55 between high school and college grades (117). A median of .475 was reported for twenty studies comparing the results of achievement tests with college marks, while forty-seven studies completed between 1934 and 1942 showed a median correlation of .52 between general aptitude and college success. In 1952, Cain, Michaelis, and Eurich reviewed studies and summaries of studies concerning the relation ship of achievement tests and college scholarship (2). They reported a median correlation of .475 to .545, but stated that the difference in means was not significant because of the limited number of studies involved. Jackscn, investigating the effectiveness of several mental ability tests for prediction of scholastic success in 1955, concluded that they predict more effec tively for women than for men (153). Comparing probationary and non-probationary students at the University of Michigan in 1951, De Ridder 14 found a significant difference between the mean scores of American Council of Education Psychological Examinations (hereafter known as the ACE) for the two groups (30). However, the disparity between the mean scores of the two groups of men was greater than the disparity between the mean scores of the groups of women. Also studying the predictive value of the ACE, in 1949, Wallace concluded that ACE test is one of the best single predictors of academic success in higher education (234). In 1953, Fessenden determined that 75 per cent of a group of probationary students at Michigan State College had ACE scores in the lower half of the class (33). He also found with the same group that high school rank was not predictive of degree of difficulty encoun tered with college work. Bolton, in 1952, confirmed the findings that the ACE test is the best of a battery of intelligence and scholastic tests used to predict college success (91). He found the prediction higher for the first year, however, than for the second. In 1949, Wheeler reported conflicting findings from his study at the University of Miami. He concluded 15 that there is no substantial relationship between ACE scores and freshman grades (240). Considering the predictive value of the subscores of the ACE, Berdie, Dressel, and Kelso, in 1951, obtained correlations of .18 to .65 between ACE L and college grades; the correlations were .15 to .55 between ACE Q and college grades (85). In 1955 at the University of Michigan, Jackson reported that the percentage of an ability group failing to obtain a "C" average increases as the ability of the group decreases (154). For his group, also, the Michigan College Reading Test proved to be the best of four entrance tests used to predict achievement. Studying the freshman class at the University of Missouri in 1950, Ryan found high school rank to be the best predictor of achievement (60). In 1952, Frederiksen and Shrader confirmed this in an investigation of 4,030 freshmen from twelve colleges (126). Having obtained correlations of .57 between first year college grades and high school standing, and .47 between first year grades and ACE score, they concluded that high school marks are the best single predictor of academic success, with ACE score second. 16 Bergeron found similar results in a 1953 study of freshmen at the University of Arkansas. He obtained correlations of .64 to .67 between college grades and high school grades, and also found ACE scores to be sig nificant in predicting college success (20). High school grades were reported to be significant determinants of college achievement by Cooper in 1955, in a study at Texas Southern University (28), and by Weltman at Washington State College in 1957 (73). French, however, in a validity study of college entrance tests in 1958, reported high school record less effective for predicting the quality of major-field work than for predicting over-all freshman grades (129). In 1953, Williams and McQuary made a study of 2,000 high school students who over a period of years had been selected for various colleges (241). Rank in high school class was found to be the best predictor of col lege success for this group. Evaluating five factors for predicting college success, in 1948, Hertel and Di Vesta concluded that high school grades are best, and Ohio State Psychological Examination scores are second in predictive value (145). 17 In 1954, Stone found high school grades first in predictive value among ACE scores, cooperative General Culture Test scores, and high school grades (220). About 1934, investigators began to study multiple prediction. The highest multiple R was found usually to come from a combination of an intelligence test, an achievement test, and high school grades. Durflinger reported a median of .60 to .70, with any correlation rarely over .80 (117). Other investigators reported multiple R's of .56 to .83. Harris, in 1940, summarized studies which used a combination of achievement and intelligence indices as predictors of college attainment and found a multiple R of .695 (144). In 1946, the Yale studies predicted college grades by using high school rank and entrance examination scores (3). Lins reported, in 1950, a system of success and failure probabilities as determined for a group of 1,789 University of Wisconsin freshmen by treating statistically ACE and high school rank percentiles (164). The use of previous scholastic records was inves tigated by Smith in 1945, who came to the following 18 conclusions regarding the prognosis of college success: (1) the high school record, when reduced to a single summary, is as useful as aptitude, reading, or English examination percentiles; (2) the previous high school record and entrance examinations seem to begin to lose value after a year or more; (3) the best single predictor in any given semester is the previous semester's record; and (4) errors of prediction are*still high, a reminder that prediction should depend on several factors (216). In his review of the literature up to 1949, Garrett concluded that high school scholarship has the greatest value for the prediction of college success (133) , followed in order by general achievement test scores, intelligence tests scores, and general college aptitude test scores. He listed combinations yielding the best prediction as being high school marks and intelligence test scores, high school marks and aptitude test scores, and intelligence and aptitude test scores. Reviewing the selection devices used at many Colleges, Cosand, in 1953, concluded that high school grades seem to be better predictors of achievement than any single test (105). He found single tests to yield 18 a maximum correlation of .50 with college grades, but obtained a multiple R of .80 when intelligence and language-skill tests wore correlated with grades. He therefore concluded that the latuer multiple measures yield better prediction than high school grades afford. Achievement was stuuied at the Cornell College of Agriculture by Chahbazi in 1.856 (25). He found for 813 students a multiple R of .536 between grade point average and five predictive variables. Cooper summarized studies dealing with college success by dividing them into four relationship areas (104). His fourth group will be discussed with the nonintellectual factors, but the other groups are as follows: (1) high school success and college success, (2) intelligence or aptitude for college work and college success, (3) achieve ment test scores and college success. He found the fol lowing ranges of relationships for the three groups: 1. Correlations ranged from .23 to .938. The most frequently reported r between the average of total high school grades and total college grades was .5. 2. Correlations from .2 to .5 were reported. Intelligence tests were considered the third 20 best predictor of college success. 3. Correlations ranged from .4 to .7. Writers seemed to feel that achievement test scores were second to high school grades and superior to other measures in predicting scholastic achievement. Spencer, also in 1955, included in his doctoral dissertation a very complete review of the literature on prediction (64). Regarding the power of intellectual factors to predict, he concluded that "the best predictor of college success has been found to be high school grades in combination with other measures, such as achievement, intelligence, and aptitude tests." Studying the academic achievement of agricultural students at the Agricultural and Mechanical College of Texas in 1955, Bertrand reported high school grades and ACE scores to be of substantial and about equal value to grade point average predictors (88). In a study to ascertain the effectiveness of certain factors in predicting the success of out-of-state students in the College of Arts and Sciences at the University of Colorado in 1956, Giles extracted from an 21 original group of eighteen variables three which yielded significant results (40). These were high school per centile rank, reading placement test, and academic placement test. In order to determine the value of high school rank in college grade prediction, Swenson studied 300 freshmen at the University of Pittsburgh in 1957. Dividing the group into fifths according to high school grades (223), and controlling ACE score, he found a significant difference between the first semester college grade point average of the upper two fifths and that of the lower three fifths. There was no significant difference, how ever, between the middle one fifth and lowest two fifths. He also found ACE scores to have lower predictive value for the upper two fifths than for the others. At the University of Minnesota, the Student Coun seling Bureau found that the ACE score, the high school percentile rank, and the Cooperative Reading Test all predict college grades, but a wide variation exists among colleges (9). It was concluded also that correlations determined for one college cannot be,used for another. Edmiston and Rhoades, in a 1959 study of ninety- four high school seniors, concluded that apparently results 22 from a general achievement measure can be used to predict college freshman marks (121). Chauncey, writing in the Journal of the National Education Association in 1958, stated that inasmuch as factors in heredity and environment which influence achievement cannot be measured separately, the present verbal and mathematics ability of a student offer the best prediction of his future academic success (98). In summary, it seems generally to be concluded that the best prediction of college success obtainable from intellectual factors probably can be found in a multiple correlation combining high school grades, achieve ment test scores, and intelligence test scores. As Traxler states, "The wider use of standard tests and cumulative records has given greater reliability to the quantitative data furnished the college" (15). If only one criterion measure can be used, it seems probable from the research that high school grades afford the highest single correlation with college grades. Nonintellectual Factors in Academic Achievement Although intelligence, achievement, and previous record have been found to be the best predictors of 23 college scholarship, many investigators feel that these criterion measures fall short of desired standards. Spencer, among others, states that the meager findings produced so far have perhaps been the result of influences of less tangible factors (64). Hutson states that as early as 1924 Feingold, in a summary of studies (7), showed that intelligence is not a good predictor of success at the college level. He found a correlation of .30 to .50 between intelligence and college grades, as opposed to one of .40 to .70 between intelligence and elementary school grades. His subsequent conclusion was that general intelligence is of less importance to success in college than to success in lower grades. In college, students avoid what they do not do well and choose subjects in line with aptitudes. One reason, of course, is that the range of ability is much more restricted in college. In 1931, Tyler reviewed the studies on academic prediction up to that time and found many investigators concluding that factors other than cognitive ones need to be considered (17). He reported that many studies already had been made on the prognostic value of interests and personality traits. 24 In other reviews of the literature concerning non- cognitive factors which contribute to achievement, the field has been subdivided into several components. These factors are presented separately in the following dis cussion. Vocational choice.--While definiteness of vocation al choice has not been conclusively proved to affect achievement, many studies have shown some degree of rela tionship between the two. Some of the earlier reviews have indicated that: (1) poor vocational choice may lead to poor academic work, (2) clear-cut vocational choice may lead to high grades in subjects closely related to that choice, (3) even a tentative choice of major may increase motivation and cause achievement to improve, (4) appropriateness of vocational goal is of greater importance than earliness or definiteness of choice, and (5) measured interest in a curriculum seems to be an index of probable success. In 1934, Van Tuyl and Eurich found that interest correlates with college success to the highest degree of any personality factor and is exceptionally stable (230). Studying probationary students at the University 25 of Michigan in 1954, De Ridder concluded that inappro priate major is the most frequent cause of probation (30). In his study at Michigan State College, Fessenden found that 70 per cent of the low aptitude students who were on probation were in the wrong major (33). Weigand, in 1953, compared a group of students who were successful in getting off probation with a group who remained on probation (236). He reported that goal orientation (definiteness of vocational choice) and goal involvement (personal interest in goal) are powerful motivating forces determining success or failure. Other recent research on vocational choice seems to corroborate the above findings. Correlating goal orientation and college achievement, in 1957, Sherwood found that liberal arts students who are educationally and vocationally oriented tend to achieve higher grades than those who are not so oriented (62). Weitz, Clark, and Jones found, in 1955, that the academic performance of college men is better if they have been motivated to choose a major before entering college (238). In 1957, Maier reported that interest test scores make sizable contributions to the accuracy of differential 26 prediction of an entrance battery (53). Investigating the OL key on the Strong Vocational Interest Blank in 1949, Ostrom reported its value in academic prediction to be significant (187). In a similar study in 1952, Gustad found sufficient correlation between college grades and the OL score to confirm the earlier findings (138). Several investigators, however, reported a lack of correlation between interest and achievement. In a 1949 study of the Kuder Preference Record, Hake and Ruedisili obtained low positive correlations between the inventory and freshman grades, the highest r being .25 (140). They concluded that interests as measured by the Kuder are a relatively minor factor in college achievement. In a study of 3,528 students at the University of Michigan in 1958, Lande found no significant relation ship between interest and achievement (49). This was confirmed in a study by Hewer, in 1957, at the University of Minnesota, though some relationship was indicated by Hewer (146). In 1954, Brooks and Weynand also reported low positive correlations between interest and academic success (92). 27 Studying the relationship of the breadth of academic interest to academic achievement in 1955, Collins found no significant correlation between the two (27). In 1956, Gaffney reported that no definite rela tionships seem to exist among age, ACE score, honor point ratio, and the likes and dislikes of college students (35). Studying undergraduate agricultural and engineering students in 1949, Stuit et al. found little correlation between interest inventories and scholastic records (14). The conclusions of the authors were that interest inven tories should be used only as bases for conferences, and not for predictive purposes. Iffert indicated the influence of interests on college success in his 1958 statement that high withdrawal rates frequently result when interests are subordinated to other considerations, such as accessibility and cost (152). Educational motives.--Although the evidence is not conclusive, seriousness of purpose, interest in academic work, and persistence seem to be related to academic achievement. In 1936, Crawford found a close relationship between seriousness of purpose and grades 28 when intelligence is controlled (106). A study at Colgate resulted in a correlation of .35 between grades and studiousness, the latter determined by a special Strong Vocational Interest Blank scale. In 1958, Teahan concluded that high academic achievers tend to look mostly toward the future, and also tend to be more optimistic (224). Reviewing the prediction research of the last twenty years in 1949, Moore attributed the still unac countable variance to motivational factors (179). Using a College Satisfaction Index in 1956, Almos found that "satisfied" students possess higher mean ability scores and make better grades than "dissatisfied" students (19). Isard, using a forced-choice inventory of attitudes toward school experience , reported in 1955 a substantial relationship between the attitudes measured and college achievement (44). Investigating academic achievement at Cornell in 1956, Chahbazi determined that the need achievement scores on two motivation tests--the Picture Stimuli Test, and the Sound Stimuli Test--add to the predictive value of an entrance battery (96). 29 Motivation was reported by Di Vesta and Woodruff, in 1949, to be significantly related to success. It was further reported by the authors that students who earn their own expenses are among the most highly motivated (112) . In 1955, a significant relationship was found by Hunger and Goeckerman between first semester grades and the degree of persistence in college (182), Aptitude scores, however, were not found to be related significantly to persistence. In a study conducted by Ryan in 1938, persistence or drive correlated .48 with scholarship (198). Ryan later developed a persistence test. Spencer's research confirmed the finding that per sistence, interest, and purpose tend to influence achievement (64). In 1948, French reported that the Educational Test Service had studied the contribution made by a per sistence test to the multiple R of an entrance battery, using grade point average as the criterion (128). The resulting R was .65 as opposed to an R of .58 without the persistence measure. Need for further research was concluded. 30 Edmiston and Jackson, in 1949, administered nine measures of persistence to 100 college students (120). Their conclusion was that little or no correlation exists between persistence and achievement. Little correlation was found between level of aspiration and college achievement in a study by Schultz and Ricciuti in 1954 (206). However, in 1959, Worell found strong support for his hypothesis that level of aspiration is related to academic success (244). Bendig reported some rather provocative findings in a study made in 1958 at the University of Pittsburgh (84). Testing the hypothesis that motivation has higher relation to future grades than to past grades, he found that need achievement is more highly related to past success than to future. This would indicate that achieve ment may be the result of past reinforcement, and that grades in college classes may be more a measure of fre quency of academic success than an index of motivation. Use of time.--It would seem that the amount of time spent in study is not as important as the manner in which the time is used. Spencer's review of literature cited studies which showed that superior students average 31 less time studying than those who make lower grades (64). One study, however, contradicts this. In 1931, Bell, controlling intelligence, obtained a correlation of .38 between time spent in study, and grades (83). Study practices.--From 1951 to 1958, many research ers found study methods a means of distinguishing high- achievers from low-achievers (23, 32, 47, 64, 70, 75, 108, 122, 147). Reading rate and reading ability also have been found to correlate substantially with college grades. Of the studies made on the results of how-to-study courses, Behrens (82) and Eckert (118), in 1935, showed low positive relationship; Edmiston (119), and Ross and Klise, in 1937 and 1927, showed high correlation between courses and grades (195). In a study of the relationship between diagnostic reading scores and college achievement in 1957, McQueen concluded that grade point average is influenced less by specific factors of reading achievement than by motivation and time devoted to study (176). In 1958, Jamuar tested the study habits of 200 college students and obtained a correlation of .51 between study habits and academic achievement (155). This finding was significant at the .01 level of confidence. Health and physical data.--Although the results of the studies are controversial, there seems to be a low positive correlation between health and scholarship. It seems logical to assume that health is a factor in achieve ment when it becomes bad enough to interfere with school work. Weber, in 1953, found a correlation of .41 between physical fitness test scores and grades of 246 freshmen at Iowa State (235). This was significant at the .01 level of confidence. Using a small sample in 1947, McCurdy found a sig nificant correlation between basal metabolism and grade point average (173). However, there has been little study undertaken in the area of the correlation of achievement and such physical data. In a 1933 study at Bucknell, no relationship was reported between grades and auditory acuity (142). In 1935, Lorenz and McClure reported that, of the students they studied, 9 per cent of those who were color blind were slightly above average in intelligence and slightly below average in scholarship (167). 33 Extracurricular activities.--Spencer concluded from his review of literature that superior students participate more than non-superiors in extracurricular activities (64). In a two part study by Siegel, in 1956, there emerged the significant finding that students who receive higher grades tend to participate in fewer activities (213, 214). In 1951, Somers found that participation in class team competition does not appreciably affect, either adversely or favorably, academic grades (217). Employment.--Although the evidence is not clear, there seems to be a tendency for the students who work part-time to be the more strongly motivated, therefore the more successful academically. In a study of 568 students at Indiana University in 1957, Trueblood found that part-time work has no significant effect on academic achievement (228). Sex differential.--Studies in this category report the correlation between intelligence and grades to be higher for women than for men. In 1935, Wagner and Strabel obtained a coefficient between intelligence and 34 grades of .72 for women and .58 for men (233), while Rundquist, in 1941, found such correlations to average .10 higher for girls than for boys (196). It was concluded by Abelson, in 1952, that high school grades are differently predictive of college grades for boys as compared with girls, but that aptitude test scores show no 3uch sex difference (76). Studying the non-cognitive influences on success at the University of Minnesota in 1955, Gerritz found a significant sex differential in contrasting successful and unsuccessful students (38). His data showed that women tend to be the more successful. Ryan, however, in his study at Missouri, found no sex differences when considering academic success (60). Spencer's summary of studies contains the con clusion that women tend to show greater achievement in college than men, although the author states that this may not hold true when aptitude and ability are con sidered (64). Iffert reported that statistics on the numbers of men and women graduating from college show too small a difference to speculate on which sex has the superior record (152). 35 High school factors.--Many studies have been made recently on the relationship of various high school factors to college achievement. One of these factors has been size of high school. In the 1953-1958 investigations of Bergeron (20), Boyd (24), Dickerson (31), Lathrop (50), Shafer (61), and Young (74), it was concluded that there is no significant relationship between high school size and college grades. One exception was reported by Young. He found that in the case of engineering majors, students from large high schools make better grades for the first semester of college than do students from small high schools. Ryan's findings differed somewhat from the above (60). He concluded that graduates from high schools with an enrollment of 200-400 achieve the greatest success in college. In 1956, Bertrand reported a relationship between ACE scores and size of high school, finding students from larger high schools making the higher scores (88). Bledsoe had reached similar conclusions in his 1954 study of Georgia high school graduates (90). He reported that students from large high schools make significantly higher 36 marks in the first year of college than students from small or medium-sized high schools. In a 1958 study at the University of Georgia, Miller compared overachieving and underachieving fresh men (56). His findings indicated many significant dif ferences independent of intelligence, but high school variables were found to be the most reliable. Over- achievers tend to come from smaller schools and towns and from rural areas. In the 1955 study at the University of California at Davis, it was concluded that being successful within a high school is twice as important as coming from a successful high school (81). Jackson, in 1955, found that students admitted to Michigan State College by special examination achieve as well as those who qualify for regular admission (154). At Harvard University in 1948, Seltzer concluded from an investigation of 1,871 freshmen that, with ability controlled, students from public schools make significantly higher grades than students from private schools. This conclusion was confirmed by Frederiksen in a study at Princeton in 1955 (124), and in a study by Davis in 1956 (109). 37 Regarding the influence of the pattern of high school courses on college grade point average, the evidence is divided. In 1958, Gephart (37) and Lathrop (50) each found mathematics and science course patterns to be predictive of college success. However, Leasman (52), Shafer (61), and Gilbert (39) found no significant relationship between high school course pattern and first year college achievement in their investigations of 1954, 1956, and 1957. Vaughan supports the latter finding by reporting that success in college depends more upon general ability than upon a set program of high school subjects (231). Age.--Studies in which intelligence has not been controlled have for the most part reported a tendency for younger students to obtain better grades. In 1934, Sarbaugh, however, controlling intelligence by using matched groups, found no age difference in achievement (201). Sarbaugh's findings were confirmed in a study by Pierson in 1948 (191). In recent years there have been many veteran versus non-veteran college achievement studies. Most of these conclude that the age and maturity of veterans 38 serve to make them more successful in college. Studies by Clark in 1947 (99), Frederiksen and Schrader in 1952 (127), and Weltman in 1957 (73), have produced evidence that veterans obtain significantly higher college grades in relation to ability than non veterans . Gideonse (134) and Gowan (136) support the con clusion that veterans excel non-veterans in grades. Their findings, reported in 1949, further indicate that veterans overachieve. The authors attribute this to age and maturity. Epler (123), and Garmezy and Crose (132) concur in their investigations of 1947 and 1948 that veterans do better in college than non-veterans; but their findings are not significant for the first semester of college. Epler's findings are significant, however, at the end of the first year. Studying veteran versus non-veteran performance at the University of California at Los Angeles, Atkinson, in 1949, concluded that veterans exceed non-veterans to an appreciable extent (80). This conclusion confirms the findings of Schrader and Frederiksen in 1948, on a 39 five-college sample of 3,756 students (204). The dif ference found between veterans and non-veterans was smaller, however, in the latter study. In 1948, Weintraub and Salley compared 507 veteran men with the same number of women students at Hunter College (237). The academic performance of the men was somewhat superior to that of the women. A further finding was a correlation of .50 between women's high school grades and college grades, and a correlation of .23 between the veterans' high school and college grades. Hansen and Paterson, in 1949, compared the pre war and post-war college academic records of the same group of veterans (141). The authors reported a sig nificant increase in post-war achievement. In contrast to the studies which show veterans to be superior, Shaffer, in 1948, studied veterans and non veterans while controlling age (211). He reported that in each age group the non-veterans surpassed the veterans; but that older students, both veterans and non-veterans, tended to make better grades. Personality factors.--In his review of academic prediction studies in 1931, Tyler reported that there had 40 been investigations on the influence of personality traits on scholastic achievement, and that some relationship had been found (17). The early studies were made by correlating with grade point average the scores on such personality tests as Bernreuter, Bell, and Pressey X-0. According to Durflinger, the correlations generally were quite low (117), with the highest reported being .26. In a 1935 study by Turrell, a relationship of .12 to .15 between the Bernreuter and scholarship was found (229). In an investigation using Borow's Inventory on College Adjustment (1), scores were correlated with first year grades of college men. The resulting r was sig nificant, but small. The Inventory yields scores on curricular adjustment, maturity of goals and level of aspiration, planning and use of time, study skills and practices, mental health, and personal relations with faculty and associates. Spencer administered the Inventory to a group of students at the Florida State University and obtained a significant relationship between composite scores on the Inventory and grade point average for the first year of college (64). 41 In other attempts to find variables which might be related to academic success, several investigators correlated results of various personality measures with college grades. The findings were divided as to outcome. Using factor analysis, Middleton, in 1958, dis covered five factors that formed a syndrome for a high- achieving group, and four factors that formed a syndrome for a low-achieving group (55). These personality syndromes were concluded to be predictive of academic success. Morgan found five variables related to academic success in his 1952 study (180): maturity and seriousness of interests, awareness of and concern for others, sense of responsibility, dominance and self-confidence, and motivation to achieve. Using Murray's need system, Pepper, in 1958, found a positive relationship between college grades and deference, order, and endurance needs (58). He found a negative relationship between achievement and hetero sexuality need. In 1947, Assum and Levy, comparing adjusted and non-adjusted college students, reported no significant difference in aptitude; however, the mean achievement of 42 the adjusted group was significantly higher than that of the non-adjusted group. Cooper attempted to determine differences between successful and non-successful college students (28). His findings showed that the non-successful tend to be more careless, be less self-critical, and have less ability; they also tend to be less self-analytic, have lower grade aspirations, and more often choose a major in line with their Kuder interest. In a similar study by Knaak, in 1955, successful students were found to miss fewer classes, participate more in class discussion, and be more per sistent in educational goals (48). In 1957, Kerns compared the characteristics of overachievers and underachievers, and found significant differences between the two groups in reasons for attending college, goal orientation, interest in academic matters, motivation, and determination (46). The above findings confirmed the results of a study made in 1954 by McQuary, who also had found background factors to be contributory to academic prediction (175). Several investigators reported on the relationship between attitudes and college achievement. Woodman, in 1952, using an instrument designed to measure attitudes 43 toward college (243),obtained a correlation of .30 with college grades. He concluded that for his particular sample the instrument was equally as predictive as the ACE. Holtzman, Brown, and Farquhar reported similar results in 1954, with the Survey of Study Habits and Attitudes (147). Testing 1,756 men and 1,118 women, they obtained a relationship with first semester college grades which was significant beyond the .01 level of confidence. They further reported that the validity of the instrument is higher for students who are sufficiently motivated to seek help with academic problems. At the State University of Iowa in 1955, Sie Georgians reported a significant correlation between academic success and results on the Survey of Study Habits and Attitudes (63). McGauvran found, in the same year, a small but stable influence of attitudes toward school in determining college success. This was confirmed by Brown, Abeles, and Iscoe in 1954, in a study which concluded that motiva tional factors are primary contributors toward scholar ship (93) . In a study of engineering students at the Univer sity of Minnesota in 1954, Wolfe determined that personal 44 motivation is an important factor in college success (9). On the other hand, in a study made in 1958, by Ahmann, Smith, and Glock, the Survey of Study Habits and Attitudes failed to make any appreciable contribution to academic prediction (78). Also Schultz and Green, in 1953, found no significant contribution to grade predic tion with an Attitude-Interest Questionnaire (205). In a study conducted by Hood at Cornell University in 1957, it was concluded that students who have a strong and satisfying family life, lack deep feelings of hostility and anxiety, and possess a feeling of security are likely to be happy and successful in college (43). Three investigators employed a biographical inventory to predict academic success. Dearborn in 1949 (4), Henley in 1955 (42), and Lathrop in 1957 (51), all reported a significant relationship between college grades and biographical factors. Weitz and Wilkinson, in 1957, determined that only-child status, a broken home, and graduation from a private military academy adversely affect college grades (239). In a study made in 1954 which compared the pre dictive efficiency of ACE score and high school grades, 45 Wolfe concluded that socioeconomic background helps determine which of the two is more important, ACE score probably becoming more important as socioeconomic back ground improves (9) . Stevens conducted a study in 1956 on the relation ship of self-concept to grades (65). He found that suc cessful students show better insight into intellectual abilities; show greater self-acceptance; and rank energy, productivity, and efficiency of greater importance than other personality traits. In a sociometric study of 225 college men, Carew, in 1957, found grades related to the degree to which a person is accepted (94). In the same year Chambers reported empathy to correlate positively and significantly with academic success (97). Also in 1957, Shaw and Brown concluded that hostility toward others and toward authority in general are negatively related to achievement (212). Schneider reported in 1957, that emotional maturity is the most important nonintellectual predictive criterion of college success (203). Frederiksen and Melville determined from their findings in 1954 that achievement can be predicted better for non-compulsives than for compulsives (125). Schmid, 46 on the other hand, in 1951, found no significant relation ship between grades and the degree of psychoneurosis (202). Three studies reported significant findings from an investigation of rating sheets. In a study at the University of Wisconsin in 1955, Sullivan found that high school principals' ratings differentiate significantly between successful and non-successful students on the following character traits (67): leadership, ability to think,study habits, industry, and initiative. At the University of Pittsburgh in 1958, Gardner obtained cor relations of .355 to .664 between first semester grades and ratings on reliability, industry, cooperation, initia tive, efficiency, and accuracy (36). Also in 1958, Hutson confirmed the above findings with a statement that character trait ratings can be highly predictive of college success. He further found first year college grades significantly indicative of later grades (7). First year marks were reported to cor relate .70 with marks in later years. The usefulness of the Rorschach in academic pre diction has been considered by several investigators. Munroe reported in 1945 that students showing neurotic tendencies do not make as high grades as students who are 47 free from such tendencies (183). She developed a Multiple Choice Rorschach yielding a neuroticism score, and showed that this score predicts college success better than an intelligence score. The combined use of both tests proved to be a better predictor than either test alone. Munroe's findings were partially substantiated in 1950 by Osborne, Sanders, and Greene, who also reported better prediction with the combined use of ACE scores and Rorschach responses (186). In the latter study, however, the ACE proved better as a single predictor than the Rorschach. Cooper concluded that earlier studies indicated only a small positive relationship between personality tests and academic success, but that studies after 1940 have tended to show that the Rorschach and certain other inventories are valid for distinguishing between the academically successful and non-successful (104). In one of these studies, McArthur and King, in 1954, reported evidence that the Rorschach can lead to correct oonclusions about college performance (170). Clark in 1958 (99), Cronbach in 1950 (107), McCandless in 1949 (171), and Sopchak in 1958 (218), investigated the usefulness of Rorschach responses in 48 academic prediction. All reported that the results failed to show significance. Inconclusive results regarding the relationship of Rorschach responses to grades were also obtained by Cooper in 1955 (104), Osborne and Landers in 1949 (185), Rust in 1955 (197), and Ryan in 1951 (59). A personality scale develdped in 1953 by Gough to predict college grades was reported by the author to have a mean correlation of .38 for a sample of 1,253 col lege students. Five researches using the Minnesota Multiphasic Inventory (MMPI) showed significant relationships between personality syndromes and grades. Drake and Oettling (115), Frick (130), Jensen (156), and Stone and Ganung (221), in studies made in 1955-1958, concluded that academic prediction can be improved with the use of the MMPI. Zeaman, in 1956, studied sex differential in relating the MMPI to achievement. He found both men and women achievers to have attitudes more conducive to achievement; but he concluded that women achievers are less tense and insecure, and better adjusted than men achievers (75). 49 Also in 1956 Frick and Keener, in a validation study, found that the MMPI adds little to a multiple R for predicting college grades (131). Quinn, in 1957, found that the inclusion of the MMPI in a freshman test battery is not warranted (57). The scales failed to differentiate characteristics of achievers and non achievers . Parrish and Rethlingshafer, in 1954, studied the relationship of the "need to achieve," as measured by the Thematic Appreception Test (TAT), and achievement in college. Their findings failed to show significance (188). In a study at the University of California at Santa Barbara, Thompson,in 1949, compared 100 students on results of six personality instruments (68). Intel ligence and achievement were controlled throughout the investigation. Findings indicated that achievers differ from non-achievers in good academic attitudes, social adjustment, masculinity of interests, and general good adjustment. Using five personality measures, a perceptual task, and an attitude scale, Field, in 1954, found high- achievers significantly higher on conformity, inquiring 50 intellect, and confident self-expression than low- achievers (34). The Rorschach, the Strong Vocational Interest Blank, family information, and a questionnaire were the bases of a study by Ryan in which he attempted, in 1951, to determine personality difference between achievers and non-achievers (59). He concluded that a positive relationship exists between grades and super-ego strength. In 1957, Horrall administered the TAT, Rorschach, and the Spencer Experience Appraisal to 188 college students at Purdue in order to compare adjustment with achievement (149). The findings supported the hypothesis that high-achievers tend to be better adjusted than low- achievers. Also studying the adjustment of college freshmen, Jensen, in 1958, concluded that non-achieving students of low ability are also handicapped in the nonintellectual areas of college life (45). There has been a number of studies on the contribu tion of combinations of nonintellectual factors to college grades. In 1958, Ward found motivation for college, background factors, personality, and work-study habits to be the most significant for academic prediction (70). 51 She further concluded that Man interest in academic matters" is a common element running through all these factors. Ward's findings confirmed those of Kim in 1957 (47), and Malloy in 1955 (168). Kim had reported that study habits, motivation, and personality play significant roles in determining college grades; Malloy had signifi cantly increased grade prediction by adding to a test battery a Life Experience Inventory covering past experience, attitudes, and background. Comparing superior and failing students in 1957, Weltman found the superior group to be better adjusted, possess more seriousness of purpose, and be more confident and secure (73). Preston and Botel concluded, in 1952, that per sistence, emotional adjustment, attitudes, interests, and level of aspiration are basic factors in college achievement (193). Several investigations included some of the above categories in their significant findings, and also added others. McQuary in 1953 (174), Dale in 1952 (108), and Hopkins, Mallison, and Sarnoff in 1958 (148), included type of financial support as predictive of grades. The 52 Sarnoff study, along with one by Boyce in 1956, added health as an important item (22); ffyers in 1952 (184), and Boyce included activity participation. Four researchers considered type of housing related to academic success. Gerritz, in 1955, made a study at the University of Minnesota (38), from which he concluded that campus residence is positively asso ciated with success and is a significant factor in con trasting successful and non-successful students. In a comparison of overachievers and under achievers at the University of Arkansas in 1957, Diener (32) found that campus residence halls have more over achievers and fraternities have more underachievers. Griffeth reported inconclusive findings from his study at the State University of Iowa in 1958, in which he investigated the effect of type of residence on academic achievement (41). His categories were: (1) residence halls, (2) fraternities, (3) rooming houses, (4) homes, (5) married, and (6) changed housing. In 1952, Dale concluded from his study of non- academic factors influencing success that unsatisfactory lodging, regardless of type, affects grades adversely (108). In several studies, results have indicated that 53 while nonintellectual factors are of some value, ability and achievement are of greater importance. Garrett found, in his 1949 review of studies, that no personality or character test has been devised which will predict college achievement to any significant extent (133). He did find, however, that certain rating scales from teachers and principals have possibilities— especially in areas of interest, leadership, citizenship, industry, and ability to do college work. Carter and McGinnis, in 1952, reported high school marks to be the best predictors of college success (95), Ohio State Psychological Examination scores next, and certain nonintellectual determinants third. In 1954, De Ridder ranked variables influencing grades in the order of their importance as follows: field of concentration, ACE scores, high school rank, and age (111). Travers stated, in 1949, that the best single predictor of college academic success is high school performance (226); he added that social adjustment, personality, and interests are contributing factors. Wellington, in an attempt to rank factors related to academic success, in 1955, concluded the following: intellectual measures are better than nonintellectual, 54 and observations of behavior are better than subjective evaluations by the student (72). In a study of science majors at Oregon State College, Crews, in 1957, reported that nonintellectual i factors generally lack significance when compared with college grades (29). He found the best predictor to be high school decile, although he stated that no single factor should be used for individual cases. Ward, in 1956, attempted to identify non-cognitive variables which would improve the predictive efficiency of the Selective Service College Qualification Test (71). He found no significant variables. Chapman investigated academic prediction at the University of Houston in 1955, and reported it reasonable to assume that personality factors bear some relationship to college success (26); he indicated that further research is needed. The following are some general remarks made by investigators regarding the value of nonintellectual factors in academic prediction. Clothia reported, in 1954, that multiple correla tions are higher when psychological examinations are used as part of an entrance battery designed to predict 55 college grades (101). In 1953, the Traxleis stated that an appraisal of the total personality Is of greater significance In selecting college students today than It was a decade ago (15). A study by Merrill, in 1954, concluded that factors other than intellectual ability alone enter into college success (178). Koelsche, in 1956, supported this conclusion with his report that low scholarship is not entirely dependent upon lack of ability, but rather is the result of many factors exerting an influence on the student (161). In 1949, Stuit et al. maintained that only a few factors associated with success have been identified and measured, and that the percentage of unknown factors is greater than the percentage of known (14). The statement is also made that more research is needed on the study of the nonintellectual correlates of success. In his survey of literature on academic predic tion, Travers concluded that the contribution made by one thousand studies in the last fifteen years was relatively small because the studies were repetitious (226), the criteria used were obscure, and all 56 inadequacies found were erroneously attributed to inade quacies in the tests. Travers further suggested that common variations in campus environment can be extremely influential on ultimate achievement, and that some good validation methods need to be developed. Differential Prediction Many investigators have suggested that differ ential grade prediction measures are needed rather than single predictive measures of over-all college success. Differential prediction is defined as the prediction of subjects in which a student will do his best work. The idea had its beginning with Thorndike, who, in 1921, postulated three types of intelligence: abstract, social, and mechanical (225). In 1927, Truman Kelley, following a similar pattern, reported his findings which consisted of three factors within intelligence: verbal, quantitative, and spatial (8). Hull and Limp, in 1925, did some of the first work in the prediction of differences in the aptitude of an individual (151). Hull concluded that the distribu tion of talent within an individual approximates a normal curve, that the range of trait differences within an 57 individual is about 80 per cent as great as that between individuals, and that differences in magnitude of trait variability suggest that guidance can be much more profitably applied with some individuals than with others. In 1932, Segel investigated the possibility of predicting differences in success between pairs of college students (207). He obtained the following correlations between predicted and actual differences: Subjects Correlation Language - economics .694 Economics - history .632 Language - physical science .555 History - physical science .543 Bio science - physical science .480 Economics - bio science .438 History - bio science .432 Physical science - English .427 Bio science - English .400 Language - bio science .333 At Yale in 1932, Crawford (106), and Wolf (242) attempted to predict success between two general types of courses: verbal, including English, history, and 58 languages; and quantitative, including mathematics, physics, chemistry, and engineering drawing. Using the Scholastic Aptitude Test (SAT) and the Mathematics Aptitude Test (MAT) of the College Entrance Examination Board, Crawford and Burnham (3) obtained the following correlations, in 1934, for 439 freshmen: average of English and history grade with high school rank and SAT ■ .55; average of mathematics and chemistry grades with high school rank and MAT ■ .66. Wolf, using three freshman classes, forecast the difference in direction of the two abilities to the extent of 4 points on a 100 point scale in 70 per cent of the cases. In the Minnesota studies in 1942, it was found that there is no predictive variable which is of much importance in predicting success in all schools and colleges of the university. Variables found especially useful in predicting success in one school or college are not necessarily prognostic for another (13). At the University of Minnesota, emphasis was placed on the use of tests which were found to be predictive of success in particular curricula. For engineering school, scores on a high school physics test, a chemistry test, a 59 mathematics test, and the high school grade average were used. For medical school, a knowledge of biological science, the ability to reason about scientific data, and reading as applied to scientific material were found the most prognostic. Other professional schools had similarly predictive programs. The conclusion reached in the Minnesota studies was that there exists no simple uniform threshold of ability to succeed in all programs of the university. Multiple R's found between grades and differential pre dictive measures used for the various schools were as follows: for nursing, dentistry, agriculture, engineering, and pharmacy, .70 to .80; for arts, medicine, education and business administration, .65 to .70; for law, .60. Douglass, in 1943, confirmed the finding that ability is not uniform for success in all fields (114), and further suggested that the prediction of college ability be replaced by the prediction of ability at a given college because of the wide differences in institu tions of higher learning. The Pennsylvania State studies were reported by Bernreuter in 1941 (87). Various batteries of tests had 60 been used experimentally since 1939, with Pennsylvania State freshmen serving as subjects. Multiple correlations with grades were computed for all major divisions, and the coefficients varied from .55 to .80. One of the instruments which Pennsylvania State studied carefully was the Thurstone Primary Mental Abilities Test. The highest correlation of a single factor with a subject grade was .44. The highest multiple R was .49. In the University of Iowa studies in 1942, Stuit and Hudson obtained somewhat higher correlations than did the Pennsylvania studies (222). In the School of Engi neering, the verbal factor correlated with grades .577, and the memory factor correlated .563. Induction and deduction correlated .400 and .385 respectively. Agriculture Prediction Since this investigation was concerned primarily with agriculture and engineering students, it seemed appropriate to sample the prediction studies in these areas. Some of the more pertinent ones are presented below. The American Council of Education investigated the prognosis of success in agricultural colleges and 61 reported, in 1949, the following findings: 1. Considerable reliance may be placed on high school record as an indicator of subsequent performance; 2. The secondary school agriculture curriculum appears as satisfactory a preparation for college agriculture as any high school curriculum. However, for predictive purposes, the quality of the high school record supersedes its composition; 3. A knowledge of science and algebra seems significantly related to performance in agriculture, particularly in the initial stages of training; 4. Students with actual farm experience have a definite advantage; 5. Aptitude and interest tests are of value when used in combination with other data; 6. None of the above is a sufficient predictor if used alone (14). These findings confirmed those in a Veterans' Administration Bulletin in 1948, which reported correla tions of .34 to .72 between high school grades and college achievement in agriculture (192), and correlations of .07 to .60 between scholastic aptitude and college achievement. 62 The article also reported a significant relationship between knowledge of science and mathematics, and success in agriculture. It further noted the advantage of actual work experience on a farm. Egbert and Hawkes used the Cornell Orientation Inquiry in 1950, in a study to predict success in an agriculture curriculum. The correlation with first grade point average was -.219. The investigators concluded that more research is needed since the items on which the Inquiry was built were felt to be highly predictive of success and should have yielded a higher correlation (122). Engineering Prediction Interest has been increasing in the prediction of engineering school success, and several investigators have reviewed present predictive measures in light of their value for producing the prospective successful engineer. Some of these measures correlated more highly than others with freshman engineering grades, a fact which suggests that differential prediction probably yields a more reliable basis for judgment in screening applicants than does general prediction. In 1954, Layton summarized the literature of the 63 past thirty years and concluded the following (9): (1) High school achievement is the best general predictor of engineering grades, yielding an r .55 to .60. (2) Scholastic aptitude tests are the next best general pre dictors. These yield an r of about .45. (3) Achieve ment tests are the third best general predictors, the r being .40 to .55. Mathematics achievement tests, however, correlate .55 with grades, and therefore are equal to high school grades in predictive power. (4) When com binations of the above measures are used, a multiple R of .40 to .85 results. Le Bold, in 1958, reviewed studies of intellectual and nonintellectual factors in engineering success. He concluded that we can predict engineering success by intellectual measures with a better than chance accuracy (162); but there is still an unknown area, probably in the nonintellectual realm, which needs additional research. At Oklahoma A and M in 1955, Stinson attempted to determine whether or not there might be an "engineering personality." He found significant differences between engineering graduates and engineering drop-outs in ability and interests, but not in personality (66). 64 In a 1949 study by Stuit et al., the quality of previous academic record was found equally as important for predicting engineering success as for predicting achieve ment in agriculture (14). The best single predictor, how ever, was reported to be demonstrated efficiency in mathe matics. Ability in English, aptitude for spatial relation ships, and comprehension of basic mechanical principles also were found to correlate significantly with success in engineering. Interest tests showed little or no correlation with grades. Similar findings were reported in 1953 by Coleman (102), and by McClanahan and Morgan in 1948 (172). The former listed tests of algebra, English, and mechanical comprehension as the most effective in freshmen engineering grade prediction; the latter found the use of English and chemistry tests best in multiple R's computed from com binations of tests. The Engineering and Physical Science Aptitude Test proved to be the best single predictor of freshman engi neering grades in a study conducted in 1949 by Treumann and Sullivan (227). Gregg, in 1951, found the same test to have reliability and validity as a predictive measure when it was administered to 352 freshmen in a study at the University of Colorado (138). 65 In 1950, Lord, Cowles, and Cynamon reported sig nificant correlations of .38 to .68 between scores on the Pre-Engineering Inventory and first semester engineering grades (166). The Iowa Aptitude Tests for Mathematics, Chemistry and Physical Science were found by Moore, in 1949, to be exceptionally good predictive instruments (179). However, he reported that mathematics ability is the best single predictor of engineering success. In reviewing studies of the last two decades, Moore stated that most correla tions between predictors and engineering grades have been in the . 50's with a few in the .60's, and a limited number in the .70*s. According to Jones and Case in 1955, the engi neering schools of the University of California at Berkeley and Los Angeles developed an aptitude test battery for lower division applicants which correlates .50 with grades (159). Several studies of prediction have been made using multiple variables. Although Sessions, in 1955, found the Pre-Engineering Ability Test to be of little selective value (210), Bowers, in 1956, combined it in a regression equation with the Cooperative Algebra Test, 66 ACE, and Minnesota Form Board Test to form a good predic tor of success at Oklahoma A and M (21). In 1953, Drake and Thomas also reported the Pre-Engineering Ability Test to be prognostic when used in combination with ACE score and high school centile rank (116). The ACE Q score and the Minnesota Paper Form Board Test were reported in 1955, by Malloy, Wysocki, and Graham to be comparatively effective as predictors of survival in first year engineering (169). In a study conducted in 1953 by Johnson at Purdue, the measures found most predictive for freshman engineering success were the Purdue Mathematics Placement and Physical Science Tests (158), with high school average being the next best. All yielded correlations of around .60 with grades. Remmers, Elliott, and Gage, in 1949, also reported achievement tests better predictors of success at Purdue than scholastic aptitude tests (194). At the University of Utah in 1951, Pierson and Jex studied 276 first year engineering students (190). They reported multiple R's in the high . 60's between first year grade point average and certain combinations of tests from the Pre-Engineering Inventory and high school grades. The best predictor of engineering grades 67 was a combination of high school grades, the total score on the Cooperative English Test (from the Cooperative General Achievement Tests), and the mathematics score on the Pre-Engineering Inventory. According to Jex, in 1957, no single item affords an adequate index of ability to do college work (157). When considering high school grades alone, in 1947, Pierson found them correlating with freshman engi neering grades to the extent of .58 (189). He further reported that marks earned in high school English are as closely related to achievement in engineering as marks in mathematics and science. In 1958, Hutson followed up Pierson's conclusion that high school success is predictive of first year col lege success with the finding that first year college grades correlate in excess of .70 with grades of later years (7). Thus, according to Hutson, a student's suc cess in his first year seems highly indicative of his success in the following years. In 1950, Berdie and Sutter found high school rank to be the most effective predictor of engineering grades, with the General Education Development Test III a close second (86). They point out, however, that the tests 68 used were not particularly geared to engineering majors, and that this fact may have had some effect on the outcome. In a five-year study, Layton used five predictors of success in a multiple regression equation (9). He reported that for his sample high school grades proved to be the best single predictor of first year engineering grades. The English Placement Test was the next best pre dictor, and the Ohio State Psychological Examination third best. In 1953, Long and Perry confirmed the findings of several previously mentioned studies by reporting a sig nificant contribution of high school grades to freshman engineering grades (165). The authors also reported that high school test scores contribute significantly to college grades. While Long and Perry concluded that interest tests make no particular contribution to prediction, two studies reported a conflicting finding. In 1948, Speer administered the Kuder Preference Record to 1,072 freshmen and obtained significantly different results from engi neering and not -engineering majors (219). Also, in 1952, Melville and Frederiksen correlated each scale of the Strong Vocational Interest Blank with engineering grades 69 and obtained significance at the .05 level of confidence for eight correlations (177). According to the authors, these results suggest that the interests of the academi cally successful are directly related to the interests of men in scientific occupations, and are inversely related to the interests of men in business. In a 1955 study by Boyce, an effort was made to improve selection in a cooperative engineering program by more adequate screening (23). Ten predictors of suc cess were finally selected for use from a group of forty- three which were taken from high school transcript, application blank, and freshman orientation test. Results were inconclusive. Two studies were concerned with upper division prediction in engineering. Ahmann, in 1955, constructed a predictive table by which graduation from Iowa State College could be predicted for transfer students (77). Variables used were previous academic achievement, high school grade average, and first grade point average at Iowa State. In 1942, Siemens studied the records of 1,400 engineering students at the University of California to determine how well upper division success could be predicted (215). Significant factors reported were grade 70 point average of first semester upper division work; grade point average of all lower division work; and grade point average of college mathematics, physics, and chemistry. Chapter Summary The review of related literature has been pre sented in the following sections: (1) intellectual factors contributing to college success, (2) nonintellec tual factors contributing to college success, (3) differ ential prediction, and (4) prediction for agriculture and engineering schools. Intellectual factors are concerned with ability, aptitude, and achievement. They were used in this study as measured by standardized tests; also considered were high school marks. Of these various factors, most investigators agree that the best single predictor is the high school grade average, which correlates about .55 with college grades. For more reliable prediction, a combination of high school grades, ability test score, and achievement test score seems to be recommended. The multiple correlation of these with college grades is about .75. 71 Nonintellectual factors are those which are more dependent on attitudes than on mental processes, and therefore seem more difficult to measure at the present time. Selected for consideration in this study were vocational choice, educational motives, use of time, study practices, health and physical data, extracurricular activities, employment, sex differential, high school variables, age, and personality traits. Most investi gators have concluded that, while these nonintellectual factors undoubtedly contribute to achievement, more research is needed to determine the significance of that contribution. Until additional research is available, ability and past achievement will no doubt continue to be considered preferable for predictive purposes. Differential prediction is concerned with deter mining the subject area in which there appears to be the most ability. Authors of the studies reviewed seem to agree that there exists no simple uniform ability for success in all colleges. Studies in the prognosis of success in agriculture and engineering schools were also reviewed. It was gener ally concluded that success in an agricultural curriculum can be predicted best by high school grades. The pattern 72 of courses was not found particularly important, but agricultural experience was reported helpful. In predicting success in engineering, high school grades correlated highest with engineering grades; how ever, the pattern of high school courses is also of significance. Specifically, mathematics and science courses seem to be the most predictive of all high school subjects. CHAPTER III METHOD The population sample used for the study was selected from new resident students at the California State Polytechnic College at San Luis Obispo, California. Students who shared rooms in college residence halls formed the experimental group, while the control group was composed of students occupying rooms alone. In order to maintain a controlled environment, only residents of halls providing both single and double room accommodations could be used; thus, the sample was limited to freshman men. Halls reserved for women and upper class men students had only double rooms, so these students were omitted from the study. The design of the investigation involved the use of related pairs. This technique seemed the most effective means for obtaining a comparison between two groups of students living under different conditions, and was par ticularly suited to the sample studied. At the time of the study, all residence halls for freshman men at the 73 74 California State Polytechnic College provided identical accommodations for all residents with the exception that half the rooms were for double occupancy, and half were for single occupancy. Thus the environment was similar for all students, except for a roommate factor. As was stated in the limitations in Chapter I, approximately 90 per cent of the room assignments were made on the basis of chance. A student desiring campus housing was placed on a waiting list in accordance with the date on which he was accepted for admission to the college. This date varied from a few days to several weeks after the filing of the application, depending upon the length of time involved in acquiring previous records and transcripts. For this reason it was extremely unlikely that two students applying together could be accepted on the same date, a requisite for simultaneous placement on the waiting list and subsequent assignment to the same room. All freshman men who resided throughout the period of the study in an originally assigned room, and with the same roommate if in a double room, were included in the investigation. Excluded were eight freshmen who 75 were involved in room changes, and nineteen who shared rooms with upperclassmen. The total number comprising the sample was 524. This included students entering college in the fall quarter of 1955 and in the fall quarter of 1956. The use of two quarters seemed advisable in order that a larger sample might be obtained. Fall quarters were chosen because they represented the students' first experience with residence hall environment. The experimental and control groups were matched by pairs on the factors of age, major field of study, and expected grade point average. The last of these matching variables was obtained from a statistical treatment of data gathered from college records. The data included various entrance test scores, and the admission status as determined from high school tran scripts. Each score and the admission status were cor related with a criterion variable, the actual grade point average for the first quarter in attendance. Subsequent to the matching, the achievement of the two groups was compared. The amount and direction of deviation of each actual grade point average from the expected grade point average was obtained, and the 76 variation in this deviation between members of pairs was used as a measure of the difference which might be caused by the influence of a roommate. Outline of Procedures The procedures used in conducting this investiga tion were as follows: 1. Nine factors generally considered influential on academic achievement including age, major field of study, four raw scores obtained from entrance tests, entrance clause under which admission to college was granted, the number of units attempted in the first quarter, and grade point average for the first quarter were selected as the bases upon which experimental and control groups might be compared. 2. The data for each subject were collected, and Pearson product-moment correlations were computed for all possible combinations of six of the factors. The remaining three factors, age, major field, and units attempted, were used in uncorrelated form. 3. The Pearson r's were combined in a Doolittle formula to obtain a multiple R. First quarter grade point average was used as the criterion, and the extent 77 to which five variables contributed to it was determined. 4. A multiple regression equation was formulated, and from this an achievement expectancy for each subject was derived. 5. Each subject's actual grade average was com pared with his expected average, and the deviation was noted. The difference in the amount, and the direction of, the deviations between members of matched pairs was determined; and appropriate tests of significance were applied to the results in order to test the hypotheses concerning roommate influence. Description of Variables The particular variables outlined were selected because they seemed to be the most objective of those factors which could be quantified for statistical pur poses, as well as being factors most commonly listed in the literature as contributing significantly to academic prediction. The exception to the above probably would be the entrance clause variable. However, in the par ticular situation in which it is used, it appears to be predictive enough to justify its inclusion here. 78 Age.--Age was considered important because the research has indicated that the interruption of an educa tion for employment or military service can effect a definite change in attitudes and motivation. An attempt was made in this investigation to control age to some extent by not pairing veterans with recent high school graduates. If any such pairs were included, the number did not exceed five. Ages of subjects in the sample ranged from seventeen to twenty-eight years. The mean age of students living alone was 18.63, with a standard deviation of 1.30 and a standard error of .08. The mean age of those sharing rooms was 18.57, with a standard deviation of 1.20 and a standard error of .07. A comparison of these figures yielded a critical ratio of .57, a figure below the level of significance. This would indicate that the difference in age between the two groups was no greater than chance expectancy and that both groups could come from the same population. Major field of study.— Since different fields of specialization employ different philosophies, courses, and grading criteria, it seemed appropriate to consider 79 major divisions separately in order to obtain a greater uniformity in predicting expectancy. The chief areas of study were agriculture and engineering. The college had been an all male school prior to 1956 and had specialized in these fields. Though coeducational status resulted in the expansion of arts and sciences, the number of men specializing in these subjects at the time of the study was small. The breakdown of the total sample into the three major fields was as follows: agriculture, 247; engineer ing, 252; arts and sciences, 25. Entrance tests.--It has been reasonably well established that entrance test scores are among the most reliable methods to date for predicting college achieve ment. The entrance tests used at the California State Polytechinic College in 1955 and 1956 consisted of the American Council of Education Test, 1953 Edition; the Diagnostic Reading Tests, Form B; and the Purdue Mathematics Placement Test. For statistical purposes, these tests will subsequently be referred to as ACE, DRT, and PMT. Raw scores on the tests were collected for all subjects in the sample. In order to determine whether 80 or not a significant difference existed between the two sample groups, the mean, standard deviation and standard error were calculated for the test scores of the roommate and singles groups. The following is a summary of the results: Roommates Singles M <r ^m M <r < p - m ACE L 58.7 15.64 .93 57.5 15.93 1.03 ACE Q 42.8 9.97 .59 41.4 11.23 .73 DRT 67.7 16.07 .95 67.8 16.08 1.04 PMT 31.7 14.54 .86 33.0 14.64 .95 Critical ratios were obtained for the above findings. The results were: ACE L ACE Q DRT PMT CR .867 1.492 .071 1.016 Since all CR's were below the level of significance, the two sample groups could be assumed 'to come from the same population in terms of the tested variables. 81 Entrance clause.--In California state colleges a series of entrance clauses has been employed as a basis of admission. In 1955 and 1956, this series consisted of three steps, each of which had certain qualifications that had to be met. The qualifications varied in degree, but all tended to measure past performance. Since past performance is considered a valid predictor of future behavior, the entrance clause under which a student was admitted seemed pertinent to a predictive study. Also an indication of motivation and drive can often be pro cured from a study of habits and behavior patterns. Although the entrance clause is not actually a quantitative measure, in its practical application at the California State Polytechnic College it has proved to be of predic- contributor in the multiple regression equation. This rather unique situation exists probably because of the curriculum at California State Polytechnic. Habit would seem more important than more verbal predictive measures. Entrance clause data are listed in the following table rather than being treated statistically. tive value. It also held 82 Roommates Singles Clause 1 178 146 Clause 2 49 43 Clause 3 57 51 When the Chi Square Test was applied to the table, the X obtained was .20. In terms of the null hypothesis, therefore, both groups could have come from the same population. The entrance clauses which composed the standard admission requirements for all California state colleges in 1955 and 1956 were as follows: 1. An applicant must have completed the equivalent of 70 semester periods (seven Carnegie units) of course work, in subjects other than physical education and military science, with grades of A or B on a five-point scale during the last three years of high school. 2. An applicant must have completed the equivalent of 50 semester periods (five Carnegie units) of course work, in subjects other than physical education and military science, with grades of A or B on a five-point scale during the last three years of high school and 83 attained the twentieth percentile on the national norm of a standard college aptitude test. 3. An applicant who fails to meet either of the above may be admitted if, in the judgment of the appro priate college authorities, he gives promise of being able to succeed in college. Unit load.--The number of units attempted was included in the data because equality of academic load would probably give some uniformity to the amount of time not spent in class. Co-curricular and extracur ricular activities were not controlled, however. The experimental and control groups were not actually matched on unit load, but no pair had more than a five unit differential; most differences were fewer than three units. In determining the unit load, only courses granting credit toward a degree were considered. Remedial and preparatory courses were excluded. Since units attempted was not one of the matching variables, the mean, standard deviation, and standard error for the two groups were not computed. Grade point average.--The grade point average for the first quarter of attendance was used as the criterion 84 of success, since grades still appear to be the most objective method of measuring academic achievement. The first quarter was chosen because that was the period of attendance studied. In computing the grade averages, a three point scale was used. This was the scale in operation at California State Polytechnic at the time. Statistical Procedures After the data had been collected, Pearson product- moment coefficients were found for all paired combinations of the following: ACE L score, ACE Q score, DRT score, PMT score, entrance clause number, and grade point average (GPA). The resulting correlations are listed below: ACE L and ACE Q .57 ACE L and DRT .78 ACE L and PMT .36 ACE L and entrance clause .21 ACE L and GPA .26 ACE Q and DRT .54 ACE Q and PMT .59 ACE Q and entrance clause .25 85 ACE Q and GPA .22 DRT and PMT 77 DRT and entrance clause 16 DRT and GPA .14 PMT and entrance clause 34 PMT and GPA .18 Entrance clause and GPA 36 A multiple R was then computed by means of the Doolittle formula, with GPA serving as the criterion variable. The R obtained was .45. This is not a high correlation, but is substantial and allows a prediction considerably above chance. An E of 10.7 was computed. In order to determine the grade expectancy for each subject, a multiple regression equation was formulated. Expected grade point average was represented by x^, and the other parts of the equation were as follows: 1.111 * constant x2 - ACE L score x3 * ACE Q score x4 - DRT score x5 = PMT score x 6 ■ entrance clause 86 The beta weights as computed were: 1,2 - '6522 1>3 - --0908 1 (4 - -.7134 1.5 “ ‘4776 1.6 " --1963 The complete equation with its relative weights was: x x - 1.111 + .0219*2 -.0052x3 -.0219x4 + .0183x5-.1405xg When the expected grade averages had been obtained, a few cases were found to be extreme. This was unavoidable, considering the multiple R of .45. However, most cases were in agreement with what could be reasonably anticipated from a study of the raw data. Pearson product-moment coefficients and grade average expectancies were computed for all 524 subjects. The expected grade point average, which incorporated ACE L score, ACE Q score, Reading test score, Mathematics test score, and entrance clause, was then used as one of the variables on which pairs were matched. The other variables were age and major field of study. 87 The matching was carried out by equating on the basis of the three factors mentioned above, one student from each pair of roommates being matched with a student occupying a room alone. The maximum variation allowed in the matching was: .10 in grade point expectancy, one year in age (although most pairs fell within a differential of six months), and no variation in major field. Pairs were not actually equated on academic load, but the intra-pair variation in number of units attempted was in no case more than six units, and seldom more than three units. The matching procedure resulted in a loss of 133 cases from the original sample, thus reducing the number of subjects from 524 to 411. Of this number, 274 composed 137 matched pairs. One member of each pair lived with a roommate, while the other member lived alone. Roommates of pair members accounted for the remaining 137 of the total 411. Matched pair members living with roommates formed the experimental group, and matched pair members living in single rooms formed the control group. The experimental and control groups composed 65 pairs majoring in agriculture, 4 pairs majoring in arts and sciences, 88 and 68 pairs studying engineering. After the matching had been completed, the actual grade point averages of all 411 subjects were compared with expected averages. The amount by which the actual grades deviated from expectancy, and the direction in which the deviation occurred, were noted. Comparisons were then made between the experimentals and their room mates as to the direction of deviation, and between the experimentals and controls as to the amount of deviation. Chapter Summary Two hypotheses were advanced concerning the possible influence of roommates on the scholastic achieve ment of college students. Freshman male students living on the campus of the California State Polytechnic College during the fall quarters of 1955 and 1956 comprised the population sample. Data collected for the sample included college entrance test scores, entrance clause under which admission was granted, age, major field of study, number of units attempted, and grade point average for the quarter. The entrance tests used were the American Council of Education Test, 1953 Edition; the Diagnostic Reading Tests, Form B; 89 and the Purdue Mathematics Placement Test. Raw scores were employed in all statistical procedures. Pearson product-moment correlations were computed for the total sample for all possible combinations of the following: ACE L score, ACE Q score, DRT score, PMT score, entrance clause, and grade point average. The resulting r's are presented in tabular form in the Appendix. With grade point average used as the criterion variable, a multiple R was computed. The R obtained was .45. While not high, this yielded an E of 10.7 and is considered a substantial correlation. A multiple regression equation was then formulated, by means of which the expected grade point average was determined for each subject. The difference between expected and actual grade point averages, and the direc tion of the difference, were noted. The total sample of 524 was divided into students sharing rooms and students living alone. Matched pairs were formed on the bases of age, major field of study, and grade point average expectancy. Maximum variations of one year in age and .10 in grade expectancy were allowed. 90 Matching reduced the sample from 524 to 411 sub jects. This resulted in 137 matched pairs plus 137 room mates. The 137 matched pairs comprised an experimental group (subjects sharing rooms) and a control group (sub jects living alone). Roommate pairs were then studied in order to determine the intra-pair difference in the directional deviation of their actual grades from their expected grades. Matched pairs were studied to determine the intra-pair difference in the quantitative deviation of actual grades from expected grades. CHAPTER IV FINDINGS In order to establish that the two groups under study were samples from the same population, it was necessary to determine for each group the mean, the standard deviation, and the standard error of the mean for four correlated variables. Age was also included in the statistical treatment. After the calculations had been completed, appropriate tests of significance were applied. The results of these statistics follow: M <T CR Age 1. Experimental 18.57 1.20 .07 2. Control 18.63 1.30 .08 .57 ACE L 1. Experimental 58.70 15.64 .93 2. Control 57.50 15.93 1.03 .87 ACE Q 1. Experimental 42.80 9.97 .59 2. Control 41.40 11.23 .73 1.49 91 92 M CR DRT 1. Experimental 67.70 16.07 .95 2. Control 67.80 16.08 1.04 .07 PMT 1. Experimental 31.70 14.54 .86 2. Control 33.00 14.64 .95 1.02 The remaining three variables used in comparing the groups--major field, number of units attempted, and grade point average--were used as matching variables or as the criterion variable. Entrance clause was presented in tabular form. All tests of significance resulted in critical ratios of less than 1.96, which is the .05 level of con fidence. Thus it could be assumed that there was no sig nificant difference between the experimental and control groups; they could have come from the same population sample. All variables used to match related pairs were correlated with grade point average, the criterion variable. Pearson product-moment correlations of coefficients found were as follows: 93 ACE Q DRT PMT Entrance clause GPA ACE L .57 .78 .36 .21 .26 ACE Q .54 .59 .25 .22 DRT .77 .16 .14 PMT .34 .18 Entrance clause .36 When these Pearson r's were combined in a Doolittle formula to obtain a multiple R, the result was .45. This is considered a substantial correlation, yielding an E of 10.7, although it does allow for some extreme cases. In testing hypothesis number 1, which was con cerned with direction of deviation of students’ grades from expectancy, students sharing rooms were compared with their roommates. Of the entire sample, which consisted of 411 college freshmen, 274 shared rooms. Thus, 137 pairs were involved in the comparison. After expected and actual grade point averages had been obtained for the 137 pairs, and the direction in which actual grades of each student deviated from expect ancy was noted, roommates were compared as to the direction of the deviation. It was found that 82 pairs varied in 94 the same direction and 53 pairs varied in the opposite direction. For the remaining 2 pairs the grade point average of the experimental member of the pair equalled expectancy, thus indicating no extraneous influence. To determine the significance of the finding, the non-parametric sign test was applied. This test was selected because its data are composed of plus and minus signs rather than quantitative measurement. Also it is particularly appropriate for a study involving directional measurement. The one-tailed version of the test was used, since a plus or minus direction frequency had been predicted in advance. As is customary in sta tistical tests involving differences, the null hypothesis is tested by the sign test. The N was 135, since this was the number of pairs of which both members showed some deviation and therefore had a plus or minus prefixed. The number of the fewer signs, 53, was represented by x. An N of 135 requires a correction for continuity, so the z formula was used. The formula and its results for this study are given below: 95 z = (x ± .5) - 1/2(N) 1/2 N z * (53 + .5) - 1/2(135) 1/2 135 z = (53.5) ~ 67.5 . -14.0 5.81 5.81 z = 2.41 P = .0080 Since P is less than .01, the null hypothesis was rejected at the .01 level of confidence. Hypothesis number 1 was therefore found tenable, and it may be assumed that for the sample studied, a student's grades were likely to differ from expectancy in the same direction as those of his roommate. In hypothesis number 2, which was concerned with the amount rather than the direction of differences in grade deviation, students sharing rooms were compared with students living alone. For this part of the study the formula for the standard error of difference between correlated means was used. This is a parametric statisti cal test used exclusively with matched groups, and is 96 interpreted in terms of the null hypothesis. In applying the formula, it was necessary to cal culate a mean, a standard deviation, and a standard error. The first step was to find between members of related pairs the difference in the deviations of their actual grade point averages from their expected averages. A sign of plus or minus was prefixed in accordance with the direction representing the greater distance from expectancy. These differences were then summed algebrai cally, and the mean was obtained by the following equation: M - ~ C) - -19.52 m .1445 N 135 The standard deviation was calculated as follows: ■ f £ ■ " 2 cr -f Xi - M* - I 70.0904 - (.1445)2 - .7059 135 Next was found the standard error of the mean. (T-^ = <r - .7059 - .0610 ^ N - T ^ 134 Because of the large sample, the critical ratio was used rather than the t. The result was a CR of 2.37. Since 2.37 is larger than 1.96, but smaller than 2.58, the null hypothesis was rejected at the .05 level of confidence 97 and found tenable at the .01 level. Hypothesis number 2 of the study therefore was found tenable at the .05, but was rejected at the .01 level of confidence. This may be interpreted to mean that, in the sample studied, roommates significantly influenced the academic achievement of college students. In summarizing this chapter, there have been presented the statistical findings of the study. Tests were applied to the experimental and control groups, and the two samples were found to be from the same population on every variable used for matching. Pearson product-moment correlations were computed for all possible combinations of six of the variables, including the criterion. These correlations are listed in Table 1 in the Appendix. A multiple R was computed with the criterion, grade point average. The R of .45 which was obtained was substantial, but it did cause some extreme cases to occur in the findings of the subsequent multiple regression equation. The latter equation was used to calculate the expected grade point average for each subject. Variations found in the deviations between the 98 actual and expected grades of roommate pairs and of related pairs were subjected to tests of significance in order to check the hypotheses of the study. The sign test found hypothesis number 1, which stated that a student's grades are likely to deviate from expectancy, above or below, in the same direction as those of his roommate, tenable at the .01 level of confidence. The standard error of the difference between correlated means determined hypothesis number 2, which stated that the grades of students living with roommates will show a greater deviation from expectancy than the grades of students living alone, tenable at the .05 level of con fidence . CHAPTER V SUMMARY, CONCLUSIONS, IMPLICATIONS AND RECOMMENDATIONS Summary The problem with which this study was concerned was the extent of influence a roommate might exert on the scholastic achievement of a college student. The hypotheses advanced were: 1. A student’s grades are likely to deviate from expectancy, above or below, in the same direction as those of his roommate. 2. The grades of students who share rooms will show a greater deviation from expectancy than the grades of students who live alone. The design of the research involved the study of related pairs. Students who shared rooms were known as the experimental group and were matched individually on certain variables with students living alone, che latter comprising the control group. The population sample was 99 100 composed of resident freshman men students at the Cali fornia State Polytechnic College at San Luis Obispo, California. Women and upper class men students were excluded from the study because of the lack of comparable living accommodations in their residence halls. The fall quarters of 1955 and 1956 were the periods studied. Two quarters were used in order to increase the size of the sample, which totaled 524 for the two periods. For the 524 subjects the following data were collected from college records: age, major field of study, first quarter grade point average, number of units attempted, four entrance test scores, and the number of the entrance clause under which admission to the college was granted. Pearson product-moment coefficients were computed for all possible combinations of the four test scores, entrance clause number, and grade point average. The correlations obtained are listed in Table 1 in the Appendix. In order to determine the combined contributions of the test scores and the entrance clause to grade point average, a multiple R was computed. The result was a substantial, though not a large .45. A multiple regression 101 equation was formulated, by which the expected grade point average of each subject was obtained. After the expectancies had been computed, the direction and amount of variance between actual and expected grade point averages were noted. Related pairs were then formed by equating members of the experimental and control groups on the factors of age, major, and expected grade point average. This matching procedure reduced the sample from 524 to 411. Of the 411, 274 consisted of 137 matched pairs, and the remaining 137 were roommates of the experimentals. Following the matching, two intra-pair studies were made. Roommate pairs were compared on the directional deviations of their actual grade point averages from expectancy, and matched pairs were compared on the quantitative deviations of their actual grade point averages from expectancy. These results were then sub jected to tests of significance. The non-parametric sign test was selected for the first hypothesis, which was concerned with direction of deviation. The one-tailed version of the test was used because a plus or minus direction had been predicted in 102 advance. It was found that grades of 82 students deviated from expectancy in the same direction as those of their roommates, grades of 53 deviated in the opposite direc tion, and grades of 2 equalled expectancy. The null hypothesis was rejected at the .01 level of confidence, and hypothesis number 1 was found tenable. The parametric standard error of the difference between correlated means was used to test the second hypothesis, as quantitative differences were involved. A mean, a standard deviation, and a standard error of the mean were computed. Following this, a critical ratio was obtained. The finding was significant in terms of the null hypothesis at the .01 level of confidence, but not at the .05 level. Therefore, hypothesis number 2 of the study was found tenable at the .05 level of confidence. Conclusions The findings reported in the preceding chapter appear to support the general conclusion that roommates undoubtedly exert some influence on the academic achieve ment of college students. On the particular sample studied, the influence extended more often in a negative than in a positive direction. This would seem to 103 indicate that students are not adequately prepared to meet the challenge that college provides. They appear to lack independence, and thus are subject to distracting or negative roommate influences. The results of the study further suggest that residence hall environment can contribute to achievement. A third conclusion supported by the findings indicates that many times nonintellectual factors can be almost as predictive as intellectual determinants of academic success. This confirms the findings of previous studies, which have shown that many factors which have no particular relationship with ability have made sig nificant contributions to achievement. The conclusion reported in earlier researches that greater reliability is needed in the prediction of college success has been substantiated by the results of this investigation. Also the finding that many students drop either to probationary status or out of college confirms the need for continuing predictive studies and for a continuing re-evaluation of admission procedures. 104 ImplieatIons Implications suggested by the present study seem to confirm those reported in earlier investigations on nonintellectual determinants of academic success. As previously mentioned, it is important that greater reli ability be attained in the prediction of achievement. With only 20 per cent of the total school population entering college, and, according to Iffert, only 60 per cent of those who enter ever graduating, an alarming waste of manpower is represented. Though the following extend beyond the findings of this investigation, on the basis of the literature reviewed they seem to be measures which might be con sidered by educators for purposes of more adequate screening of college applicants: 1. Colleges might re-evaluate their admission procedures, and in the screening of applicants incorporate more nonintellectual contributors to academic success. 2. More consideration might be given to the area of differential screening. More and more evidence is pointing to the need for predicting 105 success in a given field and at a given school rather than predicting college success in general. 3. Students planning to enter college might be given a more thorough understanding of what will be expected of them, and of what they might expect in return. Well-organized orientation programs have been found to contribute significantly to achievement. 4. Since environment undoubtedly is a very important influence, particularly in the case of the less secure student, an effort should be made to control the environment to some extent until self-confidence is gained. The literature summarized in Chapter II related academic success and a good attitude toward college to the following more specific factors: self-confidence, personal-social-emotional adjustment, socioeconomic back ground, purpose, insight into self, level of aspiration, early clarification of vocational goals, ability, and previous scholastic achievement. The inverse 106 relationship might be assumed to exist for lack of success. Previous studies have shown that college students who proceed with confidence and achieve at expectancy or above usually have well-clarified attitudes toward academic life. It has further been found that students who underachieve and appear unconfident usually lack well-defined attitudes toward school. It seems reasonable to assume that environmental factors would be more likely to have an effect on the latter group. This is substan tiated by the results of the present study, which indicate that the roommate influence on grades is more often in an adverse than in a favorable direction. In light of the importance of attitudes on academic success, education seems to be faced with the challenge of helping students clarify their value-ends. Many of the young men and women who are presently enrolled in institutions of higher learning need to gain a deeper understanding of their own feelings and attitudes, and a greater awareness of how these attitudes relate to future goals. When a thorough understanding of self is acquired, goals can be more clearly and realistically defined. 107 Again, extending beyond the findings of this study the trend toward "drifting" which seemed evident in many of the sample suggests that realistic goals may be blocked by one of three rather commonly occurring experiences in the lives of college students: 1. College could be a parental goal, a bid for prestige, or a means of postponing adult responsibilities. 2. Decision-making could be the responsibility of parents or other authorities. 3. Tension could be created through a fear of changing social class, either moving up or moving down. Since dealing with such problems is within the realm of counseling services, it would seem that high school and college counselors could do a great deal to help students who are in these various conflict situations If feelings and attitudes could be clarified, students would then be in a position to establish more realistic goals. If this clarification of self could be achieved early enough, colleges would tend to attract only students who could profit from attendance and who would be suc cessful academically. 108 Recommendations In view of the limitations imposed by the facilities available for this study, further investigation at other institutions would seem advisable. Specifically, the following recommendations are suggested: 1. A similar study might be made using the first year grade point average as a criterion rather than first quarter average. Students seem to develop more confidence and stability after the first quarter or semester. 2. A study at the junior or senior level would be useful in determining whether or not room mate influence extended to upperclassmen. 3. A validation study made at a traditional liberal arts college would be beneficial. Because of the unique "upside-down" education at the California State Polytechnic College, liberal arts courses are spread throughout four years. This results in more technical courses being taken the first year, a policy which tends to make the verbal entrance tests less predictive for this period. If motivation could in some way be controlled, a study incorporating such a control would greatly increase the reliability of the findings. A P P E N D I X Ill TABLE 1 CORRELATION OF MATCHING VARIABLES Entrance ACE Q DRT______PMT clause GPA ACE L 57 78 36 21 26 ACE Q 54 59 25 22 DRT 77 16 14 PMT 34 18 Entrance clause 36 TABLE 2 COMPARISON OF SAMPLE GROUPS ON MATCHING VARIABLES Variable Roommates Singles CR M cr ^m M cr <r Age 18.5 1.20 .07 18.6 1 .08 .570 ACE L 58.7 15.64 .93 57.5 15.93 1.03 .867 ACE Q 42.8 9.97 .59 41.4 11.23 .73 1.492 DRT 67.7 16.07 .95 67.8 16.08 1.04 .071 PMT 31.7 14.54 .86 33.0 14.64 .95 1.016 112 113 TABLE 3 COMPARISON OF SAMPLE GROUPS ON ENTRANCE CLAUSE Entrance clause Roommates Singles Clause number 1 178 146 Clause number 2 49 43 Clause number 3 57 51 BIBLIOGRAPHY BIBLIOGRAPHY Books 1. Borow, H. Manual for College Inventory of Academic Adjustment. Stanford, California: Stanford University Press, 1949. 2. Cain, L. F., et al. Prognosis. Encyclopedia of Educational Research. New York: Macmillan and Co., 1952. 3. Crawford, Albert B., and Burnham, Paul S. Forecasting College Achievement. New Haven: Yale University Press, 1946. 4. Donahue, Wilma T., Coombs, Clyde H., and Travers, Robert M. W. (eds.). The Measurement of Student Adjustment and Achievement. Ann Arbor: Uni versity of Michigan Press, 1949. 5. Frederiksen, N., and Schrader, W. B. Adjustment to College. Princeton, New Jersey: Education Testing Service, 1951. 6. Guilford, J. P. Fundamental Statistics in Psychology and Education. New York: McGraw-Hill Book Co., 1950. 7. Hutson, Percival. The Guidance Function in Education. New York: Appleton-Century-Crofts, 1958. 8. Kelley, Truman L. Crossroads in Mind of Man. Stanford, California: University Press, 1928. 115 116 9. Layton, Wilbur L. Predicting Engineering Grades. Selection and Counseling of Students in Engineering. Edited by Wilbur L. Layton. Minneapolis: University of Minnesota Press, 1954. 10. MacPhail, Andrew H. The Intelligence of College Students. Baltimore: Warwick and York, Inc., 1924. 11. Schmitz, Roy M., and Holmes, John L. Relationship of Certain Measured Abilities to Freshman Engineering Achievement. Selection and Counsel ing of Students in Engineering. Edited by Wilbur L. Layton. Minneapolis: University of Minnesota Press, 1954. 12. Siegel, Sidney. Nonparametric Statistics for the Behavioral Sciences. New York: McGraw-Hill Book Co., 1956. 13. Studies in Predicting Scholastic Achievement. Minneapolis: University of Minnesota Press, 1942. 14. Stuit, Dewey, et al. Predicting Success in Profes sional Schools. American Council on Education. Washington, D, C.: Government Printing Office, 1949. 15. Traxler, Arthur, and Townsend, Agatha. Improving Transition from School to College. New York: Harper and Bros., 1953. 16. Turabian, Kate L. A Manual for Writers of Term Papers, Theses and Dissertations. Revised edition. Chicago: University of Chicago Press, 1955. 17. Tyler, Henry T. The Bearing of Certain Personality Factors other than Intelligence on Academic Success. Contributions to Education, No. 468. New York: Teachers' College, Columbia University 1931. 117 18. Wolfe, Dael. Factors Determining Who Goes to College and Who Succeeds in College. Selection and Coun seling of Students in Engineering. Edited by Wilbur L. Layton. Minneapolis: University of Minnesota Press, 1954. Unpublished Material 19. Almos, Kermit Odel. "A Study of Interrelationships between Measured Satisfaction with College and Certain Academic and Personality Variables." Unpublished doctoral dissertation, University of Minnesota, 1956. 20. Bergeron, Wilbur Lee. "An Analysis of the Relation ship between Selected Characteristics and Academic Success of Freshmen at the University of Arkansas." Unpublished doctoral dissertation, University of Arkansas, 1953. 21. Bowers, Royal Hulch. "The Selection of the Optimal Predictors of Success in the First Semester of the Engineering Program of the Oklahoma Institute of Technology." Unpublished doctoral dissertation, Oklahoma A. and M. College, 1956. 22. Boyce, Ernest Marshall. "A Comparative Study of Overachieving and Underachieving College Students on Factors Other Than Scholastic Aptitude." Unpublished doctoral dissertation, University of Wisconsin, 1956. 23. Boyce, James Edward. "Comparisons of Methods of Combining Scores to Predict Academic Success in a Cooperative Engineering Program." Unpublished doctoral dissertation, Purdue University, 1955. 24. Boyd, Joseph Don. "The Relative Prognostic Value of Selected Criteria in Predicting Beginning Academic Success at Northwestern University." Unpublished doctoral dissertation, Northwestern University, 1955. 118 25. Chahbazi, Parviz. "Prediction of Achievement in New York College of Agriculture at Cornell University." Unpublished doctoral dissertation, Cornell University, 1956. 26. Chapman, Harold Martin. "The Prediction of Freshman Scholarship from a Combination of Standardized Test Scores and High School Grades." Unpublished doctoral dissertation, University of Houston, 1955. 27. Collins, Charles Cornelius. "The Relationship of Breadth of Academic Interest to Academic Achievement and Academic Aptitude." Unpublished doctoral dissertation, Stanford University, 1955. 28. Cooper, Mathew Nathaniel. J I To Determine the Nature and Significance, if any, of Certain Differences in the Social and Personal Adjustment of Fifty- one Successful and Fifty-one Non-Successful College Students at Texas Southern University." Unpublish ed doctoral dissertation, New York University, 1955. 29. Crews, Graydon Talmadge. "Selected Factors in Relation to College Success for Science Majors at Oregon State College." Unpublished doctoral dissertation, Oregon State College, 1957. 30. De Ridder, Lawrence Melsen. "Selected Factors Related to the Academic Achievement of Probationary Students graduated in 1948 from the College of Literature, Science, and the Arts of the Uni versity of Michigan." Unpublished doctoral dissertation, University of Michigan, 1951. 31. Dickerson, Elbert Lee. "An Analysis of the Rela tionship of Size of Arkansas High Schools to Academic Success of Graduates in the First Year at the University of Arkansas." Unpublished doctoral dissertation, University of Arkansas, 1958. 119 32. Diener, Charles L. "A Comparison of Over-Achieving and Under-Achieving Students at the University of Arkansas." Unpublished doctoral dissertation, University of Arkansas, 1957. 33. Fessenden, Beatrice Iva. "A Case Study of Factors Contributing to the Academic Deficiency of Selected Michigan State College Students on Probation." Unpublished doctoral dissertation, Michigan State College, 1953. 34. Field, Lewis William. "Personality Correlates of College Achievement and Major Areas of Study." Unpublished doctoral dissertation, University of Houston, 1954. 35. Gaffney, Louis B. "Likes and Dislikes of College Students in Relation to Adjustment and Achieve ment." Unpublished doctoral dissertation, University of Minnesota, 1956. 36. Gardner, Marion Edmund. "Prediction of Academic Success in College from Personality Trait Ratings Obtained for High School Graduates." Unpublished doctoral dissertation, University of Pittsburgh, 1958. 37. Gephart, Woodrow Wilson. "The Relationship of High School Relative Subject Preference and Relative Achievement with First Year College Relative Accomplishment fcr4 Academic Subject Areas." Unpublished doctoral dissertation, University of Pittsburgh, 1958. 38. Gerritz, Harold G. J. "The Relationship of Certain Personal and Socio-Economic Data to the Success of Resident Freshmen Enrolled in the College of Science, Literature, and the Arts at the University of Minnesota." Unpublished doctoral dissertation, University of Minnesota, 1955. 39. Gilbert, Arthur C. F. "High-School Curricular Pat terns as Related to Academic Success at the University of Nebraska." Unpublished doctoral dissertation, University of Nebraska, 1957. 120 40. Giles, LeRoy Herbert. "The Effectiveness of Certain Factors in Predicting Collegiate Success in the Selection of Out-of-State Freshmen at the University of Colorado College of Arts and Sciences." Unpublished doctoral dissertation University of Colorado, 1956. 41. Griffeth, Paul Lyman. "Type of Residence as a Factor in Academic Achievement at the State University of Iowa." Unpublished doctoral dissertation, State University of Iowa, 1958. 42. Henley, Gordon Henry. "The Use of a Forced-Choice Self-Rating Method in the Prediction of College Achievement." Unpublished doctoral dissertation, Ohio State University, 1955. 43. Hood, Albert Bullard. "Certain Non-Intellectual Factors Related to Student Attrition At Cornell University." Unpublished doctoral dissertation, Cornell University, 1957. 44. Isard, Eleanore S. "The Development of a Forced- Choice Inventory of Attitudes Toward School Experience for Predicting Scholastic Achievement in College." Unpublished doctoral dissertation, Temple University, 1955. 45. Jensen, Vern Harmon. "An Analysis and Comparison of the Adjustment Problems of Nonachieving College students of Low Scholastic Ability and Other Groups of Achieving and Nonachieving Students." Unpublished doctoral dissertation, University of Colorado, 1957. 46. Kerns, Byron Lyle. "A Study of Under-Achieving and Over-Achieving First-Semester College Freshmen as Revealed by the Way in Which They View the College Situation and Themselves as College Students." Unpublished doctoral disserta tion, University of Illinois, 1957. 121 47. Kim, Ki Suk. "The Use of Certain Measurements of Academic Aptitude, Study Habits, Motivation, and Personality in the Prediction of Academic Achievement." Unpublished doctoral dissertation, Louisiana State University, 1957. 48. Knaak, Nancy Katherine. "A Study of the Character istics of Academically Successful and Unsuccess ful Freshman Women Who Entered Northwestern University in the Fall of 1954." Unpublished doctoral dissertation, Northwestern University, 1956. 49. Lande, Leon Alvin. "The Relationship of Selected Interests of Male College Freshmen to Three Academic Levels of Achievement." Unpublished doctoral dissertation, University of Michigan, 1958. 50. Lathrop, Irvin Tunis. "Scholastic Achievement at Iowa State College Associated with High School Size and Course Pattern." Unpublished doctoral dissertation, Iowa State College, 1958. 51. Lathrop, Robert Lee. "Relation of Biographical Inventory Responses to Curricular Choice and Academic Success." Unpublished doctoral dis sertation, Iowa State College, 1957. 52. Leasman, Rudolph Eugene. "The Relation of the Pat tern of High School Courses to College Success." Unpublished doctoral dissertation, University of Illinois, 1954. 53. Maier, Glenn Emil. "The Contribution of Interest Test Scores to Differential Academic Prediction." Unpublished doctoral dissertation, University of Washington, 1957. 54. McGauvran, Mary Elizabeth. "A Study of the Rela tionships between Attitude Toward School and Scholastic Success at the High School and College Level." Unpublished doctoral disserta tion, Boston University School of Education, 1955. 122 55. Middleton, George Jr. "Personality Syndromes and Academic Achievement." Unpublished doctoral dissertation, Pennsylvania State University, 1958. 56. Miller, James Furman. "A Comparison of Overachievers and Underachievers at the University of Georgia." Unpublished doctoral dissertation, University of Georgia, 1958. 57. Quinn, Stanley Brittain. "Relationships of Certain Personality Characteristics to College Achieve ment." Unpublished doctoral dissertation, University of Wisconsin, 1957. 58. Pepper, Roger S. "The Relationship between Certain Dimensions of Psychological Need and the Achieve ment of 1st Semester College Freshmen." Unpublished doctoral dissertation, University of Michigan, 1958. 59. Ryan, Francis Joseph. "Personality differences between Under- and Over-Achievers in College." Unpublished doctoral dissertation, Columbia University, 1951. 60. Ryan, Suler Eldon. "Some Characteristics of the 1948 Freshman Class at the University of Missouri and the Relation of These Characteristics to Academic Success." Unpublished doctoral dis sertation, University of Missouri, 1950. 61. Shafer, Dwight Thomas. "Analysis of Certain Factors in the High School Preparation of Iowa High School Graduates Entering Selected Iowa Colleges." Unpublished doctoral dissertation, State Univer sity of Iowa, 1956. 62. Sherwood, Emily J. "An Investigation of the Rela tionship between the Academic Achievement and Goal-Orientation of College Students." Unpub lished doctoral dissertation, Temple University, 1957. 123 63. Sie Georgians, Djen-dzi Wei. "The Relation of two Experimental Measures of Student Motivation to Academic Success in College." Unpublished doctoral dissertation, State University of Iowa, 1955. 64. Spencer, George Minard. "An Investigation of Some Non-Intellectual Factors Presumably Affecting the Academic Adjustment of College Students at Florida State University." Unpublished doctoral dissertation, Florida State University, 1955. 65. Stevens, Peter H. "An Investigation of the Rela tionship between Certain Aspects of Self- Concept Behavior and Students1 Academic Achievement." Unpublished doctoral dissertation, New York University, 1956. 66. Stinson, Pairlee J. "Relationship of Certain Measures of Ability, Interest, and Personality to Achievement in the Engineering Program at Oklahoma Agricultural and Mechanical College." Unpublished doctoral dissertation, Oklahoma A and M, 1955. 67. Sullivan, Ben A. "A Study to Determine the Char acteristics of the Potentially Successful and Unsuccessful Freshman Students at the University of Wisconsin, Milwaukee Extension Division, in the School Year 1953-1954." Unpublished doctoral dissertation, Northwestern University, 1955. 68. Thompson, Grace M. "Non-Intellective Personality Factors Related to Academic Achievement in College." Unpublished doctoral dissertation, University of California, 1949. 69. Trueblood, Dennis Lee. "Selected Characteristics, Including Academic Achievement, of Employed and Non-Employed Students in the Indiana University School of Business." Unpublished doctoral dissertation, Indiana University, 1954. 124 70. Ward, Annie W. "Development of an Inventory of Noncognitive Predictors of Academic Success." Unpublished doctoral dissertation, University of Tennessee, 1958. 71. Ward, William Hamlin. "Validation of the Selective Service College Qualification Test and Relation ship with Certain Non-Cognitive Variables." Unpublished doctoral dissertation, Purdue University, 1956. 72. Wellington, John Adam. "Factors Related to the Academic Success of Resident Freshman Men at a Midwestern Liberal Arts College During the Academic Year 1952-1953." Unpublished doctoral dissertation, Northwestern University, 1955. 73. Weltman, Morris. "A Comparison of Failing and Superior Students in a State College." Unpub lished doctoral dissertation, University of California, Los Angeles, 1957. 74. Young, John Frederick. "A Comparison of the Academic Preparation and Achievement of College Students from Various Size High Schools in Indiana." Unpublished doctoral dissertation, Purdue University, 1958. 75. Zeaman, Jean Burgdorf. "Some of the Personality Attributes Related to Achievement in College: a Comparison of Men and Women Students." Unpublished doctoral dissertation, Michigan State University, 1956. Articles and Periodicals 76. Abelson, Robert P. "Sex Differences in Predictabil ity of College Grades," Educational and Psycho logical Measurement. 12:638-44, 1952. 77. Ahmann, J. Stanley. "Prediction of the Probability of Graduation of Engineering Transfer Students," Journal of Experimental Education, 23:281-88, 1955. 125 78. 79. 80. 81. 82. 83. 84.. 85. 86. 87. Ahmann, J. Stanley, Smith, William L., and Glock, Marvin D. "Predicting Academic Success in College by Means of a Study Habits and Attitude Inventory," Educational and Psychological Measurement, 18:853-57, 1958. Assum, Arthur L., and Levy, Sidney J. "Comparative Study of Academic Ability and Achievement of Two Groups of College Students," Journal of Educational Psychology, 38:307-10, 1947. Atkinson, Byron H. "Veteran vs. Non-Veteran Perform ance at U.C.L.A." Journal of Educational Research, 43:299-302, 1949. Baker, Elizabeth C., Baker, G. A., Roessler, E. B., and Shontz, H. B. "Factor Analysis of High School Scholastic Experience and Success in the First Semester at the University of California at Davis," College and University, 30:351-58, 1955. Behrens, H. D. "Effects of a 'How to Study' Course," Journal of Higher Education, 6:195-202, 1935. Bell, H. M. "Study Habits of Teachers College Students," Journal of Educational Psychology, 22:533-43, 1931. Bendig,. A. W. ."Predictive and Postdictive Validity of Need Achievement Measures," Journal of Educa tional Research, 52:119-120, .1958. ’ ’ ’ Berdie, R. F., Dressel, P., and Kelso, P. "Relative Validity of the Q and L Scores of the ACE Psycho logical Examination," Educational and Psycho logical Measurement, 11:803-12, 1951. Berdie, R. F., and Sutter, Nancy A. "Predicting Success of Engineering Students," Journal of Educational Psychology, 41:184-190, 1950. Bernreuter, Robert G., and Goodman, Charles H. "A Study of the Thurstone Primary Mental Abilities Tests Applied to Freshmen Engineering Students," 126 88. 89. 90. 91. 92. 93. 94. 95. Journal of Educational Psychology. 32:55-60, 1941. Bertrand, J. R., "Relation between High School Average Grade and Academic Achievement of Agriculture Students, Agricultural and Mechanical College of Texas," College and University. 30:166-181, 1955. _______ . "Relation between Enrollment of High Schools from Which Students Graduated and Academic Achievement of Agriculture Students, A. and M. College of Texas," Journal of Experi mental Education, 25:59-69, 1956. Bledsoe, Joseph C. "An Analysis of the Relationship of Size of High School to Marks Received by Graduates in First Year of College," Journal of Educational Sociology, 27:414-18, 1954. Bolton, E. B. "Predictive Value for Academic Achievement of the ACE Psychological Examination Scores," Peabody Journal of Education, 29:345-60, 1952. Brooks, Melvin S., and Weynand, Robert S. "Interest Preferences and Their Effect Upon Academic Success," Social Forces, 32:281-85, 1954. Brown, William F., Abeles, Norman, and Iscoe, Ira. "Motivational Differences between High and Low Scholarship College Students," Journal of Educational Psychology, 45:215-23, 1954. Carew, Donald K. "A Comparison of Activities, Social Acceptance, and Scholastic Achievements of Men Students," Personnel and Guidance Journal, 36:121-24, 1957. Carter, Homer L. J., and McGinnis, Dorothy J. "Some Factors Which Differentiate College Freshmen Having Lowest and Highest Point-Hour-Ratios," Journal of Educational Research. 46:219-26, 1952. 127 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. Chahbazi, Parviz. "Use of Projective Tests in Pre dicting College Achievement," Educational and Psychological Measurement. 15:484-86, 1956. Chambers, Frank M. "Empathy and Scholastic Success, Personnel and Guidance Journal, 36:282-84, 1957. Chauncey, Henry. "How Tests Help Us Identify the Academically Talented," Journal of the National Education Association, 47:230-31, 1958. Clark, Edward L. "The Veteran as a College Fresh man," School and Society. 66:205-07, 1947. Clark, Selby G. "The Rorschach and Academic Achieve ment," Personnel and Guidance Journal, 36:339- 41, 1958. Clothia, F. S. "Predicting Success in Colleges," Indian Journal of Psychology. 29:185-89, 1954. Coleman, William. "An Economical Test Battery for Predicting Freshman Engineering Course Grades," Journal of Applied Psychology, 37:464-67, 1953. Colvin, S. S., and MacPhail, H. "Intelligence of Seniors in the High Schools of Massachusetts," U. S. Office of Education Bulletin, No. 9, 1924. Cooper, James G. "The Inspection Rorschach in the Prediction of College Success," Journal of Educational Research, 49:275-82, 1955. Cosand, J. P. "Review of Selection Devices Used at Many Colleges," California Journal of Secondary Education, 28:12-21, 1953. Crawford, A. B. "Individual Differences in Educa tional Guidance," Psychological Monographs, 47: 148-72. Cronbach, Lee J. "Studies of the Group Rorschach in Relation to Success in the College of the University of Chicago," Journal of Educational Psychology, 41:65-82, 1950. 128 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. Dale, R. R. "Some Non-Academic Factors Influencing University Studies," British Journal of Sociology, 3:14-19, 1952. Davis, Junius A. "Differential College Achievement of Public vs. Private School Graduates," Journal of Counseling Psychology, 3:72-73, 1956. De Ridder, Lawrence M. "Relation between Gross Scores on the ACE and Academic Success, Journal of Educational Research, 46:353-58, 1953. _______ . "Factors Influencing Probationary Students," Journal of Higher Education, 25:95- 97, 1954. Di Vesta, Francis J., Woodruff, Aschel D., and Hertel, John P. "Motivation as a Predictor of College Success," Educational and Psychological Measurement, 9:339-48, 1949. Douglass, Harl. "The Relation of High School Prepa ration and Certain Other Factors to Academic Success at the University of Oregon," University of Oregon Education Service, 3:No. 1, 1931. _______ . "Different Levels and Patterns of Ability Necessary for Success in College," Occupations, 22:182-86, 1943. Drake, L. E., and Oettling, Eugene R. "An MMPI Pattern and a Suppressor Variable Predictive of Academic Achievement," Journal of Counseling Psychology, 4:245-47, 1957. Drake, L. E., and Thomas, W. F. "Forecasting Academic Achievement in the College of Engineering," Journal of Engineering Education, 44:275-76, 1953. Durflinger, Glenn. "Prediction of College Success A Summary of Recent Findings," Journal of the American Association of College Registrars, 19:No. 1, 1943. 129 118. 119. 120. 121. 122. 123. \ 124. 125. 126. Eckert, R. E., and Jones, E. S. "Value of 'How to Study’ Course for College Students," University of Buffalo Studies. 10:No. 2, 1-46, 1935. Edmiston, R. W. "The Effects of Emphasizing 'How to Learn' Upon Knowledge of Course Content and School Marks," Journal of Educational Psychology. 28:371-81, 1937. Edmiston, R. W., and Jackson, Lewis A. "The Relationship of Persistence of Achievement," Journal of Educational Psychology, 40:47-51, 1949. Edmiston, R. W., and Rhoades, Betty Jane. "Pre dicting Achievement," Journal of Educational Research, 52:177-80, 1959. Egbert, Robert L., and Hawkes, Glenn R. "The Use of an Orientation Inquiry as an Aid in Predicting Success in College Agriculture Curricula ' Journal of Educational Research, 44:295-302, 1950. Epler, Stephen E. "Do Veterans Make Better Grades Than Non-Veterans?" School and Society, 66:270, 1947. Frederiksen, Norman. "The Comparison of Public and Private School Graduates at Princeton University," Journal of Educational Research, 49, 1955 (Abstract). Frederiksen, N., and Melville, S. D. "Differential Predictability in the Use of Test Scores," Educational and Psychological Measurement, 14: 647-56, 1954. Frederiksen, N., and Schrader, W. B. "The ACE Psychological Examination and High School Standing As Predictors of College Success," Journal of Applied Psychology. 36:261-65, 1952. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 130 Frederiksen, N., and Schrader, W. B. "The Academic Achievement of Veteran and Non-Veteran Students," Psychological Monographs. 66, No. 345, 1952. French, John W. "The Validity of a Persistence Test," Psychometrika. 13:271-77, 1948. _______ . "Validation of New Item Types Against Four-Year Academic Criteria," Journal of Educa tional Psychology. 49:67-76, 1958. Frick, J. W. "Improving the Prediction of Academic Achievement by Use of the MMPI," Journal of Applied Psychology, 39:49-52, 1955. Frick, J. W., and Keener, Helen E. "A Validation Study of the Prediction of College Achievement," Journal of Applied Psychology, 40:251-52. Garmezy, Norman, and Crose, Jean M. "A Comparison of the Academic Achievement of Matched Groups of Veteran and Non-Veteran Freshmen at the University of Iowa," Journal of Educational Research. 41:547-50, 1948. Garrett, Harley F. " A Review and Interpretation of Investigations of Factors Related to Scholastic Success in Colleges of Arts and Sciences and Teachers Colleges," Journal of Experimental Education. 18:91-138, 1949. Gideonse, Harry D. "Educational Achievement of Veterans at Brooklyn College: A Study of the Performance of Some 2,400 Veterans in the Period from 1946 to 1949," Educational Record, 31:453- 68, 1950. Gough, Harrison G. "The Construction of a Personal ity Scale to Predict Scholastic Achievement," Journal of Applied Psychology. 37:361-66, 1953. Gowan, Arthur M. "Characteristics of Freshman Veter ans," Journal of Higher Education, 20:205-06, 1949. 131 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. Grater, Harry, and Thalman, W. A. "A Statistical Analysis of the Relationship between ACE and Grade-Point Average," Journal of Educational Research, 49:307-10, 1955. Gregg, George W. "An Investigation of the Reli ability and Validity of the Engineering and the Physical Science Aptitude Test," Journal of Educational Research, 45:299-305, 1951. Gustad, John W. "Academic Achievement and Strong Occupational Level Scores," Journal of Applied Psychology, 36:75-78, 1952. Hake, Dorothy T., and Ruedisili, C. H. "Predicting Subject Grades of Liberal Arts Freshmen with the Kuder Preference Record," Journal of Applied Psychology, 33:553-58, 1949. Hansen, Louis M., and Paterson, Donald G. "Scholastic Achievement of Veterans," School and Society, 69:195-97, 1949. Harriman, P. L. "Auditory Acuity and Success in College," Peabody Journal of Education, 11:25-31, 1933. Harris, D. "The Relationship of College Grades to Some Factors Other Than Intelligence," Archives of Psychology, 20:No. 131, 1931. Harris, Daniel. "Factors Affecting College Grades: A Review of the Literature, 1930-1937," Psychological Bulletin, 37:125-66, 1940. Hertel, J. P., and Di Vesta, Francis J. "An Evalua tion of Five Factors for Predicting the Success of Students Entering the New York State College of Agriculture," Educational and Psychological Measurement, 8:389-95, 1948. Hewer, Vivian H. "Vocational Interest-Achievement- Ability Interrelationships at the College Level," Journal of Counseling Psychology, 4:234-38, 1957. 132 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. Holtzman, Wayne H. , Brown, William F., and Farquhar, W. G. "The Survey of Study Habits and Attitudes: A New Instrument for the Prediction of Academic Success," Educational and Psychological Measure ment. 14:726-32, 1954. Hopkins, J., Mallison, N., and Sarnoff, I. "Some Non-Intellectual Correlates of Success and Failure Among University Students," British Journal of Educational Psychology. 28:25-36, 1958. Horrall, Bernice M. "Academic Performance and Personality Adjustment of Highly Intelligent College Students," Genetic Psychology Monographs, 55:3-83, 1957. Horst, Paul. "Differential Prediction in College Admissions," College Board Review, 33:19-23, 1957. Hull, Clark L., and Limp, Charles E. "The Differ entiation of the Aptitude of an Individual by Means of Test Batteries," Journal of Educational Psychology, 16:73-78, 1925. Iffert, Robert E. "Retention and Withdrawal of College Students," U. S. Office of Education Bulletin, No. 1, 1958. Jackson, Robert A. "Prediction of the Academic Success of College Freshmen," Journal of Educational Psychology, 46:296-301, 1955. _______. "Admission by Examination," College and University, 30:182-85, 1955. Jamaur, K. K. "Study Habits and Achievement," Psychology Studies of Mysore, 3:37-42, 1958. Jensen, Vern H. "Influence of Personality Traits on Academic Success," Personnel and Guidance Journal. 36:497-500, 1958. 133 157. 158. 159. 160. 161. 162. 163. 164. 165. Jex, Frank B. "University of Utah Studies in the Prediction of Academic Success," Research Monographs in Education. University of Utah, Salt Lake City, 1957. Johnson, A. Pemberton. "Counseling Engineering Freshmen," Educational and Psychological Measurement, 13:133-44, 1953. Jones, Margaret H., and Case, Harry W. "The Validation of a New Aptitude Examination for Engineering Students," Educational and Psycho logical Measurement, 15:502-08, 1955. Kinney, L. B. "A Summary of the Literature on the Use of Intelligence Tests in Colleges and Universities,” University of Minnesota Committee on Educational Research, University of Minnesota Press, Minneapolis, 1932. Koelsche, Charles L. "A Study of the Student Drop-Out Problem at Indiana University," Journal of Educational Research, 49:357-64, 1956. Le Bold, William K. "Intellectual and Non-Intel lectual Factors Involved in Predicting Engineering Success," Journal of Engineering Education, 48: 514-19, 1958. Limp, Charles E. "The Use of the Regression Equation in Determining the Aptitudes of an Individual," Journal of Educational Psychology, 16:414-18, 1925. Lins, L. J. "Probability Approach to Forecasting University Success with Measured Grades as the Criterion," Educational and Psychological Measurement, 10:386-91, 1950. Long, L., and Perry, J. D. "Academic Achievement in Engineering Related to Selection Procedures and Interests," Journal of Applied Psychology, 37:468-71, 1953. 166. 167. 168. 169. 170. 171. 172. 173. 174. 134 Lord, Frederic, Cowles, John T., and Cynamon, Manual. "The Pre-Engineering Inventory as a Predictor of Success in Engineering Colleges," Journal of Applied Psychology, 34:30-39, 1950. Lorenz, A. B., and McClure, W. E. "The Influence of Color Blindness on Intelligence and Achievement of College Men," Journal of Applied Psychology. 19:320-30, 1935. Malloy, John P. "The Prediction of College Achieve ment with the Life Experience Inventory," Educational and Psychological Measurement, 15: 170-80, 1955. Malloy, John P., Wysocki, Boeslaw, and Graham, Leo F., "Predicting Attrition-Survival in First Year Engineering," Journal of Educational Psychol ogy. 46:217-21, 1955. McArthur, Charles C., and King, Stanley. "Rorschach Configurations Associated with College Achieve ment ," Journal of Educational Psychology, 45: 492-98, 1954. McCandless, Boyd Rowden. "The Rorschach as a Predictor of Academic Success," Journal of Applied Psychology, 33:43-50, 1949. McClanahan, Walter R., and Morgan, David H. "Use of Standard Tests in Counseling Engineering Students in College," Journal of Educational Psychology, 39:491-501, 1948. McCurdy, Harold G. "Basal Metabolism and Academic Performance in a Sample of College Women," Journal of Educational Psychology, 38:363-71, 1947. McQuary, John P. "Some Relationships between Non- Intellectual Characteristics and Academic Achievement," Journal of Educational Psychology. 44:215-28, 1953. 135 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. McQuary, John P. "Some Differences between Under- and Over-Achievers Iv Coll " Educational Administration and Supervi^ j 40:117-20, 1954. McQueen, R. "Diagnostic Reading Scores and College Achievement," Psychological Reports, 3:627-29, 1957. Melville, S. D. and Frederiksen, Norman. "Achieve ment of Freshman Engineering Students and the Strong Vocational Interest Blank," Journal of Applied Psychology, 36:169-73, 1952. Merrill, Reed M. "An Evaluative Study of Probation Students' Academic Performance in a University," Journal of Educational Research. 48:37-45, 1954. Moore, Joseph E. "A Decade of Attempts to Predict Scholastic Success in Engineering Schools," Occupations, 28:92-96, 1949. Morgan, Henry H. "A Psychometric Comparison of Achieving and Non-Achieving Students of High Ability," Journal of Consulting Psychology, 16:292-98, 1952. Hunger, Paul. "Student Persistence in College," Personnel and Guidance Journal, 35:241-43, 1956. Hunger, Paul, and Goeckerman, Robert W. "Collegiate Persistence of Upper- and Lower-Third High School Graduates," Journal of Counseling Psychology, 2:142-45, 1955. Munroe, Ruth. "Prediction of the Adjustment and Academic Performance of College Students by a Modification of the Rorschach Method," Applied Psychology Monographs, No. 7, 1945. ftyers, Robert Cobb. "Biographical Factors and Academic Achievement: An Experimental Investiga tion," Educational and Psychological Measurement, 12:415-26, 1952. 136 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. Osborne, R. Travis, and Sanders, Wilma B. "Multiple- Choice Rorschach Responses of College Achievers and Non-Achievers,1 1 Educational and Psychological Measurement, 9:685-91, 1949. Osborne, R. Travis, Sanders, Wilma B., and Greene, James E. "The Prediction of Academic Success by Means of 'Weighted' Harrower-Rorschach Responses," Journal of Clinical Psychology. 3:253-58, 1950. Ostrom, Stanley R. "The OL Key of the Strong Vocational Interest Blank for Men and Scholastic Success at the College Freshman Level," Journal of Applied Psychology. 33:51-54, 1949. Parrish, J., and Rethlingshafer, Dorothy. "A Study of the Need to Achieve in College Achievers and Non-Achievers," Journal of General Psychology, 50:209-26, 1954. Pierson, George A. "School Marks and Success in Engineering," Educational and Psychological Measurement, 7:612-17, 1947. Pierson, George A., and Jex, Frank B. "Using the Cooperative General Achievement Tests to Predict Success in Engineering," Educational and Psychological Measurement, 11:397-402, 1951. Pierson, Rowland R. "Age Versus Academic Success in College Students," School and Society, 68: 94-95, 1948. "Predicting Success in Training for Agriculture," Veteran Administration Technical Bulletin, Nos. 7-90, 1948. Preston, Ralph C., and Botel, Morton. "The Relation of Reading Skill and Other Factors to the Academic Achievement of 2,048 College Students," Journal of Experimental Education, 20:363-71, 1952. Remmers, Hermann H., Elliott, Donald, and Gage, Nathaniel. "Curricular Differences in Predicting 137 195. 196. 197. 198. 199. 200. 201. 202. 203. Scholastic Achievement: Applications to Counseling," Journal of Educational Psychology. 40:385-94, 1949. Ross, C. C., and Klise, N. M. "Study Methods of College Students in Relation to Intelligence and Achievement," Educational Administration and Supervision, 13:551-62, 1927. Rundquist, E. A. "Sex, Intelligence, and School Marks," School and Society, 53:452-56, 1941. Rust, Ralph M. , and Ryan, F. J. "The Relationship of Some Rorschach Variables to Academic Behavior," Journal of Personality, 21:441-56, 1953. Ryan, D. G. "A Study of the Observed Relationship between Persistence Test Results, Intelligence Indices, and Academic Success," Journal of Educa tional Psychology, 29:573-80, 1930. Ryan, F. J., Davie, James S. "Social Acceptance, Academic Achievement, and Academic Aptitude Among High School Students," Journal of Educa tional Research, 52:101-06, 1958. Sanders, Wilma B., Osborne, R. Travis, and Greene, J. E. "Intelligence and Academic Performance of College Students of Urban, Rural, and Mixed Backgrounds," Journal of Educational Research, 49:185-93, 1955. Sarbaugh, M. E. "The Young College Student," University of Buffalo Studies, 9:58-73, 1934. Schmid, John Jr. "Psychoneurosis and Scholastic Achievement," Journal of Educational Research, 44:387-89, 1951. Schneider, Alexander A. "Emotional Problems and Academic Performance in College Students," Catholic Counselor, 1:4-7, 1957. 138 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. Schrader, William B., and Frederiksen, Norman. "The Comparative Achievement of Veteran and Non-Veteran Students in College," American Psychologist, 3:259, 1948 (Abstract). Schultz, D. G., and Green, B. F. Jr. "Predicting Academic Achievement with a New Attitude- Interest Questionnaire - II," Educational and Psychological Measurement, 13:54-64, 1953. Schultz, D. G., and Ricciuti, H. N. "Level of Aspiration Measures and College Achievement," Journal of General Psychology, 51:267-75, 1954. Segel, David. "Differential Prediction of Scholastic Success," School and Society, 39:91-96, 1934. Segel, David. "Prediction on Success in College," U. S. Office of Education Bulletin, No. 15, 1934. Seltzer, Carl. "Academic Success in College of Public and Private School Students: Freshman Year at Harvard," Journal of Psychology, 25: 419-31, 1948. Sessions, Frank Q. "An Analysis of the Predictive Value of the Pre-Engineering Ability Test," Journal of Applied Psychology, 39:119-22, 1955. Shaffer, Robert H. "A Note on the Alleged Scholastic Superiority of Veterans," School and Society, 67:205, 1948. Shaw, Merville C., and Brown, Donald J. "Scholastic Underachievement of Bright College Students," Personnel and Guidance Journal. 36:195-99, 1957. Siegel, L. "A Biographical Inventory for Students: I Construction and Standardization of the Instrument," Journal of Applied Psychology, 40:5.-10, 1956. 214. . "A Biographical Inventory for Students: II Validation of the Instrument," Journal of Applied Psychology, 40:122-26, 1956. 139 215. 216. 217. 218. 219. 220. 221. 222. 223. Siemens, Cornelius H. "Forecasting the Academic Achievement of Engineering Students," Journal of Engineering Education. 32:617-21, 1942. Smith, Francis F. "The Use of Previous Record in Estimating College Success," Journal of Educational Psychology, 36:167-76, 1945. Somers, Madeline R. "A Comparative Study of Par ticipation in Extracurricular Sports and Academic Grades," Research Quarterly of the American Association for Health, Physical Educa tion and Recreation. 22:84-91, 1951. Sopchak, Andrew L, "Predic;ion of College Performance by Commonly Used Tests," Journal of Clinical Psychology, 14:194-97, 1958. Speer, George S. "The Vocational Interests of Engineering and Non-Engineering Students," Journal of Psychology, 25:357-63, 1948. Stone, J. B. "Differential Prediction of Academic Success at Brigham Young University," Journal of Applied Psychology, 38:109-10, 1954. Stone, David R., and Ganung, George R. "A Study of Scholastic Achievement Related to Personality as Measured by the Minnesota Multiphasic Personality Inventory," Journal of Educational Research, 50:155-56, 1956. Stuit, Dewey B., and Hudson, Harry H. "The Relation of Primary Mental Abilities to Success in Professional Schools," Journal of Experimental Education, 10:179-82, 1942. Swenson, Clifford H., Jr. "College Performance of Students with High and Low High School Grades When Academic Aptitude Is Controlled," Journal of Educational Research, 50:597-603, 1957. 140 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. Teahan, John E. "Future Time Perspective, Optimism, and Academic Achievement," Journal of Abnormal and Social Psychology, 57:379-80, 1958. Thorndike, E. L. "Intelligence and Its Uses," Harper*s Magazine, 140:227-35, 1920. Travers, Robert M. W. "The Prediction of Achieve ment," School and Society, 70:293-94, 1949. Treumann, Mildred Jenkins, and Sullivan, Ben A. "Use of the Engineering and Physical Science Aptitude Test as a Predictor of Academic Achievement of Freshman Engineering Students," Journal of Educational Research, 43:129-33, 1949. Trueblood, Dennis L. "Effects of Employment on Academic Achievement," Personnel and Guidance Journal. 36:112-15, 1957. Turrell, A. M. "Relationship of Personality to Scholarship," Junior College Journal, 5:355-57, 1935. Van Tuyl, K., and Eurich, A. C. "Measuring the Interest of College Students with Different Major Subjects," Journal of Applied Psychology, 18:27-44, 1934. Vaughan, William H. "Are Academic Subjects in High School the Most Desirable Preparation for College Entrance?" Peabody Journal of Education, 25:94-99, 1947. Wagner, M. E. "Prediction of College Success," University of Buffalo Studies, 9:125-44, 1934. Wagner, M. E., and Strabel, E. "Homogeneous Grouping as a Means of Prediction of Academic Performance," Journal of Applied Psychology. 19:426-46, 1935. 141 234. 235. 236. 237. 238. 239. 240. 241. 242. Wallace, W. L. "Differential Predictive Value of the ACE Psychological Examination," School and Society, 70:23-25, 1949. Weber, John Robert. "Relationship of Physical Fitness to Success in College and to Personality," Research Quarterly of the American Association for Health. Physical Education and Recreation, 24:471-74, 1953. Weigand, George. "Goal Aspiration and Academic Success," Personnel and Guidance Journal, 31: 458-61, 1953. Weintraub, Ruth G., and Salley, Ruth E. "Hunter College Reports on Its Veterans," School and Society. 68:59-63, 1948. Weitz, Henry, Clarke, Mary, and Jones, Ora. "The Relationship Between Choice of a Major Field of Study and Academic Preparation and Performance," Educational and Psychological Measurement, 15:28-38, 1955. Weitz, Henry, and Wilkinson, H. Jean. "The Relationship between Certain Nonintellective Factors and Academic Success in College," Journal of Counseling Psychology, 4:54-60, 1957. Wheeler, Lester R. "Summary of a Study of the Intelligence of University of Miami Freshmen," Journal of Educational Research. 43:307-08, 1949. Williams, Henrietta V., and McQuary, John P. "The High-School Performance of College Freshmen," Educational Administration and Supervision, 39: 303-08, 1953. Wolf, Ralph R. "Differential Forecasts of Achieve ment," Psychology Monographs, 51:No. 227, 1939. 142 243. Woodman, Everett M. "Description of a Guidance Instrument Designed to Measure Attitudes Related to Academic Success in College," Educational and Psychological Measurement, 12:275-84, 1952. 244. Worell, Leonard. "Level of Aspiration and Academic Success," Journal of Educational Psychology, 50:47-54, 1959.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The Delinquent'S Self-Concept

PDF

The Influence Of Certain Characteristics Of Teachers And Teacher Raters On The Quality Of Formal Teacher Appraisal

PDF

The Effects Of Unification On The Instructional Service Functions Of The Intermediate Unit In California

PDF

Librarians' Perceptions Of Librarianship

PDF

Differential Effects Of A Reading Instructional System Using Different Reading Schedules On Low And Middle Socioeconomic Ethnic Groups Of Varying Ability In Grades One And Two

PDF

Opinions Of Beginning High School Business Teachers Regarding Their College Preparation For Teaching

PDF

The Relationship Of Tested Creative Abilities And Selected Factors Of Academic Achievement, Intelligence, Sex, Socioeconomic Status And Pupil Attitudes

PDF

A Survey Of Summer School Programs In Los Angeles County-1959

PDF

Tibullus I: Theme And Technique: "Di Tibi Divitias Dederunt"

PDF

An Analysis Of The Position Of The Elementary School Guidance Counselor

PDF

A Study Of The Effectiveness Of Group Counseling In Achieving The Goals Of Guidance In Education, Using Two Contrasting Formats

PDF

Principals' Perceptions Of Changes In Their Role Resulting From Unification

PDF

The Effects Of Individual Counseling On Selected Groups Of Underachievingstudents With High Ability

PDF

Effects Of Short-Term Confrontational Group Counseling On The Risk-Takingbehavior Of Counselor Education Graduate Students

PDF

Achievement In Junior College As Related To Eligibility Level On Entrance

PDF

Factors Related To Teacher Longevity Of Employment

PDF

The Stability Of The Self-Concept In Junior College Students

PDF

Studies In The Influence Of The 'Commedia Dell'Arte' On English Drama: 1605-1800

PDF

Conflicts In Role Expectations For Academic Deans In Church-Related Colleges

PDF

The Effects Of Group Experiences On The Aged

Asset Metadata

Creator Murray, Mary Etta Brookes (author)

Core Title The Effect Of Roommates On The Scholastic Achievement Of College Students

Degree Doctor of Philosophy

Degree Program Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, educational psychology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Carnes, Earl F. (committee chair), O'Neil, Edward N. (committee member), Stoops, Emery (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-85288

Unique identifier UC11357570

Identifier 6006211.pdf (filename),usctheses-c18-85288 (legacy record id)

Legacy Identifier 6006211.pdf

Dmrecord 85288

Document Type Dissertation

Rights Murray, Mary Etta Brookes

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA