A COMPARISON or memo ACADEMI¢ AND PERSONAL CHARACTEmsms or REGULARLY ' ‘EN‘ROLLED AND COMMUNITHUMOR Como: ’ TRANSFER STUDENT-SAT MICZHIGAN STATE uuiveksmr Thom for Hip Degree of Ed. D. MICHIGAN STATE UNIVERSlTY‘I Francis Joseph Hennessy - 1960 ' “TIN. ”WIHIWWWHW 31293 01070 6301 This is to certify that the thesis entitled A COMPARISON OF SEIECTED ACADEMIC AND PERSONAL CHARACTERISTICS OF REGULARLY ENROLLE'D AND COMMUNITY-JUNIOR COLLEGE TRANSFER STUDENTS AT MICHIGAN STATE UNIVERSITY presented by Francis Joseph Harmony has been accepted towards fulfillment of the requirements for M degree in m 9442, f£4m Major flofessor Date May 11) 1960 O~169 L [I Mich Ur. A COMPARISON OF SELECTED ACADB'IIC AN) PERSONAL CHARACTERISTICS OF REGULARLY ENROLLED AND COMMUNITY-JUMOR COLLEGE TRAIBFER STUDENTS AT MICHIGAN STATE UNIVERSITY BY Francis Joseph Hermessy AN ABSTRACT Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and APPIiOd Science in partial fulfillment of the requirements for the degree of DOCTOR OF EDUCATION Depart-am; of Administrative and Educational Services 1960 1 Francis J. Hennessy ' Abstract This study is concerned With a comparison between community— junior college transfer students and regularly enrolled students at ‘Michigan State University. The primary objective of the study was to determine how these two groups of students compare with regard to selec- ted academic and personal characteristics. In addition, the study in- vestigated the predictive efficiency of selected educational variables for community-junior college transfer students. The samples consisted of 173 community-junior college transfer students and 173 regularly enrolled students. Each group included 137 males and 36 females. The two groups had earned a comparable number of credits prior to the Fall term, 1958 and were enrolled for classes at the beginning of the Fall term 1958. Data relative to the selected academic and personal variables were secured from permanent records in the Office of the Registrar, Idichigan State University. The differences between the two groups were tested by application of Fisher's "t" test or chi-square technique as appropriate. The predictive value of the selected variables which either singly or in combination maximize predictive efficiency were tested by application of appropriate correlation techniques. It was discovered that these two groups of students were quite similar in many respects but significantly different with regard to some of the variables. Community-junior college transfer students achieved grade-point-averages which were slightly lower than those achieved by the regularly enrolled students. Female community-junior college trans- fer students experience severe "grade-point-losses" during their first term at‘Michigan State University. A significantly greater number of t. ..u vs w I 2 Francis J. Hennessy Abstract community-junior college transfer students than non-transfer students fail to maintain passing (2.0) grade-point-averages. This is also true for those enrolled in the College of Business and Public Service. A significantly greater number of the community-junior college transfer students than non-transfer students were married and/or veterans. A significant difference was also found between the two groups with re- spect to the occupational status of the fathers. The fathers of non- transfer students generally possessed higher status positions than those of the community-junior college transfer group. The best single pre- dictor of academic success at Michigan State University for community- junior college transfer students was found to be grade-point-average earned previously at the community-junior college. None of the other variables tested proved useful for predictive purposes. The results of this study led to the conclusion that these two groups of students were quite similar in many respects. However, there would seem to be cause for reviewing orientation procedures as applied to community-junior college transfer students. It would also seem appropriate for the various Colleges at Michigan State University to review their policies with regard to the advisement of the community- junior college transfer student. It was further concluded that a com- prehensive study of drop-outs should be made in the near future. A C(MPARISON OF 831.8an ACADEMIC All) PERSONAL CHARACTERISTICS OF REGULARLY smousn AN) COMMUNITY-JUNIOR COLLEGE 111mm STUDENTS AT MICHIGAN STAR UNIVERSITY BY Francis Joseph Bennessy A THESIS Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF EDUCATION Department of Administrative and Educational Services 1960 ii Francis Joseph Bennessy Candidate for the degree of Doctor of Education Final Examination: May 11, 1960, Room 427, College of Education Dissertation: A Comparison of Selected Academic and Personal Charac- teristics of Regularly Enrolled and Comunity-Junior College Transfer Students at Michigan State University Outline of Studies : Major Subject: Counseling and Guidance Minor Subject: Higher Education Cognate Area: Sociolog Biographical Items: Born: March 25, 1929, Springfield, Massachusetts Undergraduate Studies: State Teachers College, Bridgewater, Massachusetts , 1946-50 . Graduate Studies: Boston College, 1950-54 Boston University. 8.8.. 1955 Michigan State University 1956-60 Experience: Military: United States Arley. Company Comander, 1952-54 Education: High School Teacher-Coach, Bast Bridgewater, Massachusetts, 1950-54; Instructor in Education, University of Dayton, 1954-56: Graduate Teaching Assistant, Michigan State University, 1956-57; Director of Guidance, Winds-ere School District, Lansing. Michigan, 1957-58: Director of Student Personnel Services, Muskegon Comunity College, Musbegon, Michigan, 1958-59; Assistant Instructor of Guidance, Michigan State University, 1959-60. Member of American Personnel and Guidance Association, American College Personnel Association, Kappa Delta Pi. iii ACKNOWLEDGEMENT The writer wishes to express his sincere appreciation to Dr. Walter F. Johnson, Chairman of his Guidance Committee and Co-Director of the thesis, who provided constant encouragement and wise direction throughout the course of the doctoral program. In addition, he desires to express appreciation to Dr. Willard Harrington, who as Co-Director of the thesis, provided much inspiration and many helpful suggestions. To the other members of the Guidance Committee, Dr. Robert Hopper and Dr. William Form the writer is indebted for their genuine cooper- ation and helpful suggestions. ‘ Grateful acknowledgement is also due to Dr. Eldon Nonnamaker, .Assistant Director,‘Men's Division of student Affairs for his helpful suggestions. Mr. John Paterson, Instructor in the Bureau of Educa- tional Research, Michigan State University rendered valuable assistance in the statistical analysis of the data. Sincere appreciation is also extended to the staff of the Office of the Registrar, Michigan State University, for their cooperation and assistance in gathering the data for this study. Final acknowledgement must be made to my wife, Ruth Jeffery Hennessy, who assisted materially with many phases of the research and whose unwavering confidence in the writer made this project possible. TABLE OF CONTENTS P ERS 0 ML DAM . O O C 0 . O O O O O O O O O O O O O O O O O O O ACKNWIIE mm N13 0 O O O O O O O O O O O O O O O 0 O O O O O 0 LI S T OF TABLES O O O O O O C O O O O C O O O O O O O O O O O 0 CHAPTER I. II. III. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . Statement of the Problem . . . . . . . . . . . . . . . The Hypotheses . . . . . . . . . . . . . . . . . . . Purpose of the Study . . . . . . . . . . . . . . . . . Importance of the Problem . . . . . . . . . . . . . . Limitations of the Study . . . . . . . . . . . . . . . Definition of Terms . . . . . . . . . . . . . . . . . The Plan of the Dissertation . . . . . . . . . . . . . REVIEW OF THE LITERATURE . . . . . . . . . . . . . . . . Literature concerning the Academic and Personal Characteristics of the Junior College Transfer Student Literature Concerning the Prediction of Academic Success in Four Year Colleges and Universities . . . . General Review of Prediction Studies in Education 0 O O O O O O O I O O O O O O O O O 0 Review of Prediction Studies Directly Related to Junior College Transfer Students . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . METHODS.AND PROCEDURES . . . . . . . . . . . . . . . . . The Samples . . . . . . . . . . . . . . . . . . . . . Procedures Used in Collecting and Tabulating the Data Techniques of Analysis . . . . . . . . . . . . . . . . Fisher's "t” Test . . . . . . . . . . . . . . . . . . Chi-square Computations e e e e o a a e e e e e e e 0 iv Page . ii .iii .vii .21 .21 .23 .25 .28 .28 .30 .32 32 .33 Multiple Correlation Coefficients . . . . . . . . . . 8m” 0 O O O O O O O O O O O O O O O O O O I O O O 0 IV. ANALYSIS OF THE DATA ON THE DIFFERENIES BETWEEN THE “CUPS O O O O O O O O O O O O O O O O O O O O O O O O O The Differences and Similarities Between the T140 craps O C O O 0 C O O O O O O O O O O O O O O O O O O Resultsofthe"t"tests. ...... . .. . Results of the Chi—square Analyses . . . . . . . . . Other Data Illustrating Differences and Similarities BetweentheGroups................. smry O O O O O O O O O O O O O O O O O O O O O O O V. ANALYSIS OF THE DATA ON TEE RELATIOI‘EHIP BETWEEN CERTAIN SELECTED EDUCATIONAL VARIABLES AN) ACADEMIC SUCCESS . . The Zero-order Coefficients of Correlation for selected variables believed to be related to Academic Success for All Courses Taken and for Social Studies andNaturalScience ................. The Relationship Between Selected Variables and Academic Success of Male Students . . . . . . . The Relationship Between Selected Variables and Academic Success of Female Students . . . . . . The Relationship Between Selected Variables and Academic Success in Social Studies for “‘1. Stm.nt. O O O O O O O O O O O O O O O O O O O The Relationship Between‘Selected Variables and Academic Success in Natural Science for MaleStudents ................... Predicting Cumulative Grade Point Average . . . . . . The Multiple Correlation Coefficient . . . . . . . . The Affect of Reduced Course Lead on Grade POint Awrag. O O O O O O O O O 9 O O O O O O O O O 0 SM” 0 O O O O O O O O O O O I O O O O O O 0 O O O 34 34 35 37 37 38 42 49 51 54 55 59 60 63 66 69 69 72 74 vi VI. summer, comiusmrs am sucsss'nors FOR FURTHER “mm C O O C O O O O O O O C O O O O O O O O O O O O O O O 76 m PrOblem O O O O O O O O C O O O O O O O O O O O O O O O 7 6 m th°d°1°gy O O O O O O O C O O O O O O O O 0 O O O O O O O 7 7 Fimtm' O O O O O O O O O O O 0 O O O O O O 0 O O O O O 0 7 9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . 82 Implications for Further Research . . . . . . . . . . . . . 87 BIBLIMeeeeeeeeeeeeee'eeeeeeeeeeeeee89 APPENDIX A Procedures for Preparing Data for Machine Analysis . . . 94 B Courses used For Computation of Grade Point Averages in Social Studies and Natural Science . . . . . . . . . .104 C Classification of Curriculum and of Fathers Occupational Status . . . . . . . . . . . . . . . . . .106 D Standard Error of The Zero-Order Correlation coeffiCi-ents O O O O O O C C O O O O O O O O O 0 O O 0 .108 E Frequency Data Related to the Effect of Sex, Marital Status and‘Military Status on Success-Failure . . . . . .110 F Computer Laboratory Library Routine KBAM . . . . . . . .112 G Community College of Origin For Transfer Students Included in This Study . . . . . . . . . . . . . . . . .116 III. IV. VII. VIII. IX. LIST OF TABLES Means, Standard Deviations and "E'Values on Selected Variables for Male Students . . . . . . Means, Standard Deviations and"t"Values on Selected Variables for Female Students . . . . . The Success-Failure Ratio Between Comunity- Junior College Transfer Students and Non- Transfer Students Regardless of Curriculum or College . . . . . . . . . . . . . . . . . . . The Success-Failure Ratio Between Camunity- Junior College Transfer Students and Non- Tnnsfer Students in the College of Busi- nessandPublicService............. The Success-Failure Ratio Between Comunity- Junior College Transfer Students and Non- Transfer Students in the College of Engineering................... The Success-Failure Ratio Between Comunity- Junior College Transfer Students and Non- Transfer Students in the College of Education . . The Success-Failure Ratio Between Comunity- Junior College Transfer Students and Non- Transfer Students in the College of Science .w Art. 0 0 O O O O O O O O O O O O O O O O O O The Drop-Out Ratio Between Cos-lunity-Junior College Transfer Students and Non-Transfer Stm.nt. O O O O O O O O O O O O O O O O O O O O The Married-single Ratio Between Co—unity-Junior College Transfer Students and the Non- TransferStudent................. The Veteran-Non-Veteran Ratio Between Co-unity- Junior College Transfer Students and Non- TransferStudents................ The Ratio of Fathers in Six Different Occupational Classifications Between the Conunity-Junior College Transfer Students and the Non-Transfer Students .................... vii Page 38 39 43 43 44 45 46 46 47 48 49 ~ TA VA Table XII.a. XII.b. XIII.a. XIII.b. XIV.a. XIV.b. XV.b. XVII. XVIII. The Average Number of Term.Hours Taken . . . . . Means, Standard Deviations and Zero-Order Corre— lation.Matrix for Sample of 123 Male Communitquunior College Transfer Students . . Means, Standard Deviations and Zero-Order Corre- lation Matrix for Sample of 128 Male Non- Transfer Students . . . . . . . . . . . . . . Means, Standard Deviations and Zero-Order Corre- lation‘Matrix for Sample of 32 Female Communitquunior College Transfer Students . . Means, Standard Deviations and Zero-Order Corre- lation Matrix for Sample of 32 Female Non- Transfer Students . . . . . . . . . . . . . . Means, Standard Deviations and Zero-Order Corre- lation Matrix for Sample of 60 Male Community- Junior College Transfer Students in Social Studies . . . . . . . . . . . . . . . . . . . Means, Standard Deviations and Zero-Order Corre- lation Matrix for Sample of 50 Male Non- Transfer Students in Social Studies . . . . . Means, Standard Deviations and Zero-Order Corre— lation Matrix for Sample of 51 Male Community- Junior College Transfer Students in Natural Science . . . . . . . . . . . . . . . . . . . Means, Standard Deviations and Zero-Order Corre- lation.Matrix for Sample of 51 Male Non- Transfer Students in Natural Science . . . . . Summary Data Used in the Prediction of Cumulative Grade-Point Average for All Courses Taken by 123 Male CommunityaJunior College Transfer Students and 128 Male Non-Transfer Students . The Relationship Between Reduction of Course Load and Grade-Point Average for CommunityaJunior College Transfer Students . . . . . . . . . . The Relationship Between Reduction of Course Load and Grade-Point Average for Non-Transfer Students viii Page 50 57 58 61 62 64 65 67 68 71 73 73 CHAPTER I INTRODUCTION The community-junior college transfer student is fast becoming a proportionally larger segment of the new student population at Michigan State University. The future holds promise that ultimately a majority of new students in any given year will come from the community-junior colleges of‘Michigan.(36) At the present time these students are ac- tually a small minority of the entering groups. This fact may in part explain the lack of any systematic institutional studies of this group of students and of their progress. However, they have not been total- ly neglected since the Office of Community-Junior College Cooperation and the Office of Evaluation Services have undertaken several analy- ses of the Michigan.community-junior college transfer student. The exact nature of the studies undertaken by these offices will be re- viewed in Chapter II. There is general agreement among University officials that, in fact, very little is known about the community-junior college trans- fer student. Por example, it has not been systematically determined at Michigan State University whether or not he differs in any signifi- . cant way from his non-transfer counterpart even though differences might be expected between two such groups. The fact remains that if wide dif- ferences do separate these two groups there would be need to revise thinking with regard to present practices in admissions, orientation, testing and counseling. These and other student personnel practices are expected to focus on the needs of the student body. However, the univer- sity must know the student in order to provide adequate and appropriate services to meet his particular needs. In this regard it would seem that several questions require answers.» Who is the community-junior college transfer student? What are his distinguishing characteristics? How do these characteristics affect his university work? Specifically, what is the influence of certain educational and personal background factors on his university achievement? Finally, is there a need to develop new emphasis in student personnel practices as they relate to the community-junior college transfer student? The literature as reviewed in.Chapter II indicates the presence on the national scene of certain trends regarding these questions. It remains for this study to focus on these trends at Michigan State Uni- versity and to investigate in greater detail certain selected variables that might be expected to influence the progress of community-junior college transfer students at this institution. For example, with one exception the studies reported in the literature did not investigate the possible influence of sex differences and/or other background factone. The following general hypothesis are introduced in order to focus attention on the major objectives of this study. They were formu-’ lated with reference to the findings of studies of a similar nature accomplished outside the state of Michigan. Statement of the Problem ,EZEE_§_potheses The major hypotheses to be tested in this investigation are as follows: H0: The comunity-junior college transfer student is differ- ent from his non-transfer counterpart in respect to a number of educa- tional and personal characteristics. H0: Previous grades from community-junior college are the most efficient forecasters of the achievement of community-junior college transfer student after they transfer to the university. TE Purpgse _o_f_ the m It is the purpose of this study to provide a description of the cousunity-junior college transfer student and to com- pare him with his non-transfer counterpart and in the process point out differences and similarities between the groups. The study will also endeavor to measure the degree of relationship between certain educa- tional variables and the academic success of the comunity-junior college transfer student. More, specifically, this study seeks to determine the differences between the two groups with respect to: 1. first tam G.P.A.s 2. second term G.P.A.s 3. third term cumulative G.P.A.s 4. G.P.A.s for the first two years of 'college work 5. social studies G.P.A.s 6. natural science G.P.A.s 7. high school class rank 8. occupational level of the fathers In order to develop a more complete description of the continuity-- junior college transfer student, and compare him with the non-transfer student, the study further seeks to determine the ratio of: 1. success or failure within and between the two groups as a V whole. ‘ 2. success or failure within and betweenathe two groups by curricula. 3. veteran to non-veteran within and between the two groups. 4. married students to non-married students within and between the two groups. 5. drop-outs to non-drop-outs within and between the two groups. The following data also contributes to the description and comp parison. l. The average number of term hours carried by the two groups each term. 2. Term hours earned previous to Fall term, 1958. 3. Total college credits earned up to and including spring term, 1959. 4. Age In considering the advisability of academic load adjustment for comunity-junior college transfer students, the following analysis was undertaken: The effect of decreased course load from‘Fall to Winter terms on the academic performance of community-junior college transfer students. In addition, this study seeks to determine the relationships between: 1. The college qualification test derived scores of the C.J.C. transfer students and their first term G.P.A.s at Michigan State University. 2. The college qualification test derived scores of the C.J.C. transfer students and their cumulative G.P.A.s at Michigan State Univer- sity for three terms.1 3. The reading test derived scores of the 0.3.0. transfer students and their cunulative G.P.A.s at Michigan State University for thee terms. 4. The English test derived scores of the 0.3.0. transfer students and their cumulative G.P.A.s at Michigan State University for three terms. 5. The high school class rank of the C.J.C. transfer stu- dents and their cumulative G.P.A.s at Michigan State University for three terns. 6. 0.51.0. G.P.A.s and Michigan State University G.P.A.s. 7. Transfer students first term Michigan State University G.P.A.s and their third term cumulative G.P.A.s. 8. Test scores, high school class rank, C.J.C. social studies G.P.A., C.J.C. cutlative G.§.A. and social studies G.P.A. at Michigan State University. 9. Test scores, high school class rank, C.J.C. cumulative G.P.A., C.J.C. natural science G.P.A. and natural science G.P.A. at Michigan State University. The non-transfer group will be analysed in somewhat similar fash- ion» with the following exceptions: 1Throughout this study the three terms referred to are: Pall term, 1958; Winter term, 1959; Spring term, 1959. a. The test data will differ as explained on p. 8 of this chapter. b. Since community-junior college G.P.A.s are non existent for non-transfer students, their basic college G.P.A.s will be used. Importance of the Problem At the present time community-junior college transfer students are admitted to the University providing that they can present an accep- table community-junior college transcript. The criteria for admission include: 1. a minimum 2.0 grade point average. 2. an acceptable community-junior college course pattern to fit the curriculum to be followed. 3. Community-junior college courses certified for college credit. Teaminal courses are not usually acceptable. On the other hand, selection of students entering from high school is done in view of many factors which focus on the estimated capa- city of a particular applicant to succeed in the academic program. This practice reflects institutional consideration for the applicant and works to reduce the rate of attrition. Bow adequate are the criteria for community-junior college transfer students, as compared with those for nonptransfers in assisting the personnel worker in working with the trans- fer student? ' There is a dearth of research in the literature which specificallyb/ analyzes the factors for the prediction of academic success of the commu- nity-junior college transfer student. A fsw'such studies are reported in Chapter II. A very small amount of general information regarding the differences and similarities between the community-junior college transfer and the non-transfer is available. No controlled studies have been undertaken which compare community-junior college transfers with a comparable group of regularly enrolled students. Thus this investigation was undertaken to ascertain the differences and similarities between two such groups, to investigate the predictive efficiency of selected educational \ariables and to indicate the need for the exten- sion of orientation, testing and counseling programs to augment and facilitate the early adjustment of the community-junior college trans- fer student at Michigan State University. Limitations of the Study 1. The study is limited to Michigan State University students. Thus the findings are more applicable to this institution than any other. Therefore the results should not be generalized beyond the boundaries of this institution and it would semm desirable to obtain similar findings from other institutions since it is not known whether or not significant differences exist between those who transfer to ZMichigan State University and those who transfer to other four year institutions. 2. The study is not directly concerned with the mechanics of admissions and does not attempt to evaluate or make recommendations in‘ this area. 3. The study concerns itself only withfiMichigan community-junior college transfer students who accumulated at least 75 term hours of community-junior college credits. It does not refer to those students who transferred with less than 75 hours. 4. Comparisons between the transfer students and the non- transfer students are limited insofar as aptitude, reading and English test scores are concerned because: a. The orientation test battery was changed in the Fall of 1958, giving the transfer students an entirely different set of test scores.1 The non-transfer group was tested in the Fall of 1956. b. The non-transfer group was on the average three years younger at the time of testing in 1956 than the transfer group at the time of their testing in 1958. c. The transfer group had the benefit of added educational experience and maturity before being tested. Therefore, test scores cannot be considered comparable in anything other than very general terms. d. Community-junior college transfer students derived test scores are based on freshman norms. 5. It must be assumed that the information contained in the students’ records is accurate. 6. The possibilities of errors inherent in the random selective process enters into the results obtained relative to the non-transfer sample. 1The 1956 orientation test battery consisted of the ACE Psychological examination, a reading test and an English usage test. The 1958 orientation test battery consisted of the College Qualification test and new reading and English usage tests. 7. The follow-up of the two groups was limited to the first three years of college work. 8. No attempt was made to determine what happened to those who dropped out. 9. Certain data did not yield large enough N's to be useful for study and were not reported. Definition of Terms The Michigan Community-Junior College transfer student. For the purposes of this study, the student shall have attended one of thirteen community-junior colleges in Michigan and shall have accumulated 75 term hours of transferrable credit at that institution. They enrolled at Michigan State University in the Fall of 1958. They will be referred to in this study as C.J.C. transfer students. The Non-transfer student. Students who enrolled at Michigan State University directly following their graduation from high school in 1956 and were still enrolled in the Fall of 1958. College Qualification Test. A commonly used general academic aptitude test for college freshmen. The test gives verbal, numerical, 1 information and total scores. It is a measure of general college ability administered to transfer students in 1958. The test 25 English Usage, 1958. It is designed to test the students capabilities in basic English skills.‘ It is a measure of pro- ficiency in grammar and expression administered to transfer students in 1Includes social studies and natural science. 10 1L958. This is a different form of the test administered to non-trans- fer students in 1956. The Michigan State University Reading Test, 1958. furnishes a score in reading comprehension. It is a measure of general reading ability. The American Council 22 Education Psychological Examination is a commonly used general academic aptitude test for college freshmen. The test gives measures of quantitive and linguistic abilities and a total score which is the sum of Q and L. It is a measure of general college ability which has given way to other tests in recent years, particularly the College Qualification Test mentioned earlier. It was administered to non-transfers in 1956. ThepMichigan State University Reading Test, 1956. The test fur- nishes three scores: Vocabulary, comprehension andtotal score. It is a measure of general reading ability administered to non-transfers in 1956. Derived Scores. Each of the tests mentioned above are reported in terms of derived scores. Derived scores are based upon a standard ten point scale ranging from 1 (the lowest) to 10 (the highest). The scale reduces the percent of students in the extremes and increases the per- cent in the middle of the scale. Under this system extreme scores be- come much more significant in indicating superior and inferior ability. The derived score distributions for tests, Fall 1956 and Fall 1958, as established by the Michigan State University Board of Examiners, were utilized in this study. These are not decile scores. Eigh School 5225. A derived score distribution was devised for high school ranks in order to provide standard score units to facilitate 11 analysis. The rank in class was divided by the number in the class. giving a per cent. The table for proportions of area under the nor- ‘mal curve of distribution gave the number of observations per derived score unit. [Fifteen derived score units represent high school rank from 1 (the lowest) to 15 (the highest). (Example: A student who ranked 1 in a class of 64 would be seen as .015 per cent. The derived score unit for this percentage in this distribution would be 13. At the same time a student who ranked 5 in a class of 328 would also be seen as .015 per cent. The derived score unit in this case would a1- so be 13.) §£292_£212£ Average. The abbreviation G.P.A. refers to a sum calculated by dividing the number of honor pOints by the number of semester hours. The marking systems included in this study all utilize the point system ranging from 0.00 to 4.00. Thus honor points are cal- culated by multiplying the number value for the grade by the number of term hours granted for the specific course. (The total honor points for all courses taken divided by the total term hours for all courses taken yields the G.P.A. Cumulative Grade Point Average. Throughout this study the term Cumulative G.P.A. will refer to the average based on term hours and honor points earned during Fall, Winter and Spring terms, 1958-59. Social Studies Grade 22325 Average. All grades in the list of courses shown in Appendix B were recorded from the students' records and a grade point average determined as explained above.1 1Grade point averages were computed when the students' records indicated 6 or more term hours in the subject. 12 Natural Science Grade Point Average. All grades in the list of courses shown in appendix B were recorded from the students‘ records and a G.P.A. determined as explained above. Chapter Chapter Chapter Chapter Chapter Chapter II III IV The Plan of the Dissertation Introduction to the Problem, The Purpose of the Study, Importance of the Problem, Limitations of the Study, Definition of terms. Review of the Literature on Academic characteristics of the C.J.C. transfer student and Studies of Prediction in Education Appropriate to this study. Methods and Procedures Analysis of the Data on the Differences between the Groups Analysis of the Data on the Relationship between Selected Variables and Academic Performance Summary, Conclusions and Suggestions for Further Research CHAPTER II REVIEW OF THE LITERATURE The major purposes of this investigation are: (1) to investigate the differences1 and similarities between C.J.C. transfer students and non-transfer students at Michigan State University. (2) to investigate the predictive efficiency of selected educa- tional variables for C.J.C. transfer students at‘Michigan State University. I The literature which has been reviewed was selected for its pertinence to the particular aspects of this study as stated above. It is the purpose of this review to bring together the important studies which have appeared within the past 30'years concerned with the follow» ing: (1) the academic and personal characteristics of the junior college transfer student. ‘ (2) the prediction of academic success in four year colleges for junior college transfer students. The literature reviewed in this chapter might seem to cover the subject of this study rather thoroughly. However, close examination reveals several inadequacies in the research.which has been.comp1eted to 1Primarily academic differences 14 date. First, comparisons of junior college transfer students and regu- larly enrolled students have not been accomplished by matching compar~ able groups of junior college transfer students and regularly enrolled students. Secondly, sex differences and other background factors have not been taken into account in the several analyses reported on in this chapter. Third, and most important for consideration, is the fact that the results of the studies reported here are mainly applicable to the institutions in which they were undertaken. The Academic Characteristics of the Junior College Transfer Student .393 Achievement BEESES.2£.EEE Communityegggigg College Transfer Student 33 Compared E that 93 92h}: Students. Reeves (41:95), reporting on factors affecting scholarship in the transferred groups at the University of Chicago (1933), concluded that junior college transfer students made lower G.P.A.s after transfer than were made by the respective control students who had had all their work at the University. However, the difference was not statistically significant. He reported further that students transferring from junior colleges made the best records among the transferring groups. From his study it would seem that junior college transfer students achieve nearly as well as their non-transfer counterparts. A prominent claim made by junior colleges is that they can offer two years of work acceptable to colleges and universities. Perhaps the best way to substantiate this claim would be to determine whether or not the junior college graduate has received a training which will qualify 15 him to pursue advanced college work in a four year degree-granting insitution with a degree of proficiency equal to students who have re- ceived their first two years of college training in standard four year colleges and universities. Over the years since the study by Reeves (41), several of our larger State Universities have conducted studies which shed some light on the situation. The University Examiner, University of Illinois (26:303) con- ducted a study of the scholarship records for the junior and senior years in the University of Illinois of those students who entered the University with junior standing (60-72 semester hours) during a four year period. The.conclusions of this study indicate that the trans- ferring group from junior colleges exailed the other groups from dif- ferent types of institutions. The University Examiner goes on to state that: ”From.the facts presented in this report it may be said that without doubt junior college graduates are able to pursue advanced college courses in the junior and senior years at the University of Illinois with a degree of proficiency equal to and in.some cases superior to that of students who have received their first two years of training in standard colleges and universities.' Although the University of Illinois findings are dated 1934, it is significant to note that the findings arein close agreement with other studies of a similar character which have been made since that time in other universities where junior college graduates are received in substantial numbers. ‘Martorana (37:415), in a study of 251 transfer students at the State College of washington, concluded that: “....when students are considered in groups, there is no significant difference between the academic success of the students who come from.the junior colleges and that achieved by students who begin as freshmen in the institution.‘ 16 Of special note is the fact that in the subject areas of engineering and physical sciences, the transfer students, as a group, outdid their nonstransfer counterparts. The Office of Educational Research and Service at the FlOrida State University had been studying students who transfer from.Florida Junior Colleges to the Florida State University. (2.3.21) Three studies have been completed, each covering a two year span of time. Each of the studies has given positive evidence that the junior college transfer student who transferred after completion of the junior college course of study (60 semester hours) did essentially the same quality of work in the University as did the native student. The Office of Evaluation Services at Michigan State University reports several significant findings in their study of the transfer stu- dent there. (53) This study revealed that junior college transfer stu- dents are doing about as well as students who entered as freshmen. How» ever, higher ability students seem.1ess numerous among these junior college groups at Michigan State University. The junior college trans- fer students at Michigan State University generally rank in the upper half of their junior college class with a smaller than expected number ranking hear the top. It is further reported that junior level trans- fers in Engineering had higher G.P.A.s than all Michigan State Univer- sity Engineering juniors, but junior level transfers in Business and Public Service, Science and Arts, and Education are lower on.theee G.P.A. measures. All of these findings seem to indicate that the junior college transfer student has kept pace with his non-transfer counterpart over 17 the years in many different institutions in many widely separated loca- tions. The Failure and Dr__g_p-_o__ut m of the W College Transfer Student 2 ared with 95325 Students. Reeves (41:93) found a relatively higher failure rate among transferred students and attributed it in part to the differances in grading standards at the University of Chicago (and at the various other institutions. attended). It was reported that most of the with- drawals took place relatively soon after matriculation and that a higher percentage of withdrawals come from the transfer group. Grossman (26:301) reported that a larger percentage of the entrants from junior colleges were placed on probation for low scholarship than other students at the University of Illinois. The number of students who dropped from school was also reported higher from the junior college transfer group. In the study conducted by Martorana (37:413) at the State College of Washington, the drop-out rate for junior college transfers was 34.7 per cent as compared with 23.9 per cent of the non-transfers. According to Martorana, “The evidence, though not conclusive, shows that the percentages of drop-outs due to low scholarship were less among transfers than among non-transfers. Lower financial EHth on the part of transfer students to meet tin cost of education away from home may be a partial explanation. ' The University Achievement 3239.59. 3.1. the M College Transfer Student 23 Compared with his Junior College Record. The Office of Educational Research and Service, Florida State University (3:7) states that: 18 'It seems safe to conclude that the student who transferred to Florida State University were among the more able academically of all the students in the respective junior alleges from which they transferred.“ It is reported further that transfers with 60 or more semester hours of junior college credit did essentially the same quality of work in the University as they did in junior college before transfer. The Office of Evaluation Services, Michigan State University (53:1) states that in the case of the 0.3.0. transfer student: ‘College records rather than high school records must now be evaluated. Where entrance test scores were previously a convenient supplement to the school record, the test score for transfer students with collegiate experience is less meaningful. The evaluation of transcripts from previous colleges becomes a particularly difficult and sensitive issue.” Other studies (38,40,46,50) reviewed point to the fact that the quality of work done in junior college compares favorably with that done subsequently in the University although.G.P.A.s are some- what lower in the University than they were in junior college. This factor may be partially accounted for in terms of observations made at Florida State University (2:2). 'During their first semester of enrollment in the Florida State University, junior college transfer students quite frequently suffer substantial 'G.P.A. shocks'. There is good reason to believe, however, that after the first semester of enrollment these students rapidly recover from these G.P.A. losses and go on to do University work comparable in quality to the junior college work they did before transfer.‘ The Academic Abilities 2f the Junior College Transfer Student, 53 Measured Pl Aptitude and Achievement Tests. In a study by Kirk (34), a comparison was made of junior college transfer students and students from other sources of origin on the College Qualification test. In comparing the transfer students with 19 freshmen, Kirk found that both.men and women score higher than do entering freshmen students in total scores. It was also indicated that junior transfers score about one half a standard deviation better than freshmen and sophmore transfer students. Seashore (44) compared junior college freshmen students classi- fied as transfers with freshmen in senior colleges and universities on the College qualification test. He found that the median score for junior college transfer freshmen is near the 25th percentile for senior college freshmen. I These findings are further indication that the better students matriculate to the universities from junior colleges as suggested in the Florida State University study. (21:4) At'Michigan.State University, the Office of Evaluation Services reports: (53:7) ‘It appears that transfer students entering Michigan State University at more advanced levels receive progressively higher median scores on the orientation tests. This is particularly evident for the mere verbal tests.” They state further that this may be a result of extended college experience or the intrusion of a selective process which tends to elimi- nate larger proportions of students at the lower ability levels. At Florida State University (3:4-5), the junior college transfer students equalled at the time of transfer the test norms for freshmen entering four year colleges and exceeded nomms of freshmen entering junior colleges. Curriculum Choices 25 the Junior College Transfer Student. The most popular schools selected by junior college transfer students entering Florida State University seem to be Education, Science and Arts, and Business. At the State College of Washington, the 2O choices seem to favor Engineering, Science and Arts, and Business. These appear to be customary choices although the order may change in some institutions as at Michigan State University (53) where prefer- ence is highest for Business, Science and Arts, and Engineering in that order. Tbse are also the most popular choices of non-transfer fresh- men. Personal and background factors Medsker (38:41) has summarized the literature dealing with per- sonal background factors as they relate to junior college students. He re ports that: 'Public junior colleges, being primarily local and inexpensive to attend, draw heavily from.the lower half of the socio-economic dis- tribution, as shown by various studies..... An analysis of data procurred from six public junior colleges reported the occupations of the parents of almost five thousand stu- dents enrolled over a period of three years. Only one fourth of the group came from the higher level in an arbitrary high-low classifica- tion. The largest group of students (almost a third) came from a skilled labor background. Only a tenth came from.families in the professional category. An index of marital status is also available from this same group of six public junior colleges. Of more than eight thousand students enrolled in the six colleges, 23 per cent were married. In 75 two year colleges studied by“Medsker, the ratio of men to semen was three to one. A ratio of two men to one woman in junior colleges was found in the study of college entrants in Minnesota. (13) 21 The Prediction of Academic Success in Four Year Colleges General Review pf Prediction Studies _i_t_1 Education. In order to provide perspective for viewing the prediction stu- dies related to junior college transfer students, it would seem ‘ appropriate to review some of the studies which have been made in gen- eral in the field of education. ‘A complete review is not necessary since several authors (6.16.24) have summarized these studies making such a procedure unnecessary. The reliability of teacher grades. Grades or G.P.A.s are the principal criteria of academic success used in this study. The ques- tion is, how reliable are teacher grades? Bohan (5) points out that it is practically impossible to make comparisons from teacher to teacher. Williamson (55) found that there has been a failure to ad- just grades to changes in aptitude or ability level of students in the Arts College of the University of Minnesota. Generally speaking, according to Johnson (32:23): The chief factors which tend to reduce the reliability of academic grades revolve, first of all, around the subjective elements contained within the instructors' estimation of performance or achievement, and secondly, the student's actual performance. With respect to the first factor. Feder (20:108) states that: 'Most college instruction proceeds upon the tacit assumption that all students are equal in ability, approximately equally con- ditioned by past experiences, and therefore equally able to profit from the learning opportunities offered in higher education. With respect to the second factor, Borow (6) points out that an impor- tant part of scholastic performance can be accounted for by aspects of student behavior which are not associated with intellectual aptitude for 22 college work. He has defined and described six adjustment categories which affect academic achievement: curricular adjustment; maturity of goals and level of aspiration; personal efficiency; planning and use of time; study skills and practices; mental health; personal relations (with faculty and associates). The conclusion, then, is that the same criticism that is leveled at other prediction studies which employ teacher grades as the criterion of academic success may be leveled at the present study. . Scholastic aptitude test scores as predictors of academic success in college. Johnson (32:21), reporting in 1950, indicates: It may be said that the findings regarding the efficiency of intelligence of scholastic aptitude tests as predictors of college scholarship vary with such factors as the curricula in which the students are enrolled. the particular tests being studied, and the nature of the population groups being studied. It can be further stated that conclusions reached by authors reviewing the literature periodically with respect to the efficiency of tests of this type for prediction of scholastic success were in close agreement (median correlation coefficients of approximately .44--.45). High school rank 32‘s predictor of scholastic success in college. High school rank is a widely used predictive device. Froehlich (23) reports an r of .62 between G.P.A. and high school rank. Borow (6) found that rank standing in the graduating class has yielded about (.55), a slightly higher correlation than individual tests. Garrett (24) concludes that high school rank yields a high correlation (.49) when compared with G.P.A. in college. Johnson (32) states that: There seems to be rather consistent agreement that high school record or rank. in spite of factors of size of the school, pattern of courses. and variations in marking practices. is of considerable value for predicting subsequent scholastic achievement and still remains the best single index for prediction of academic success in college. 23 Combination of factors for general scholastic prediction. Most investigators are in agreement that prediction should not be a matter of selecting any single measure but rather of multiple-correlation or re- gression line techniques. It is possible to determine by these methods, combinations of factors that will prove most efficient for predictive purposes. Johnston (33) reports that by use of a combination of college aptitude test scores and high school rank it is possible to fix a threshold which.will select out students who are not likely to succeed in college work. The foregoing review indicates that the individual items con- sidered in the prediction of success in college have value, and that combined, their value becomes more pronounced. However, it is impor- tant to recognize that their efficiency, either singly or in combination, still falls far short of the ideal. The studies referred to above were directly concerned with predicting college success for those students enrolling immediately upon completion of their high school course of study. Following is a review of studies relating to the junior college transfer student. Prediction studies directly related to junior college transfer students. Aptitude and Achievement test scores 2£_predictors. Rodes (43:22) reports that the University of California is constantly re- fining its method of admission to the junior year in Engineering. In 1949, a battery of tests lasting one full day was required of all appli- cants for admission to upper division courses in‘Engineering. These tests attempted to measure achievement in five subject fields-«dEnglish, Mathematics, Physics, Chemistry, and Engineering Drawing. Prediction 24 studies for transfer students admitted to the junior year in Engineer- ing during 1947, revealed a correlation coefficient of .63 between the total scores on these tests and subsequent grades in Engineering courses. ' In a recent study at Florida State University (21) total scores on the Florida StateAWide Twelfth Grade Testing Program (F.S.T.G.T.P.) battery of tests were correlated with grades earned at junior college and at Florida State University. The scores were as follows: F.S.T.G.T.P. scores and junior college G.P.A.s, .3188; F.S.T.G.T.P. scores and University G.P.A.s, .1092. The F.S.T.G.T.P. Scores seem useful in helping to predict achievement in junior college work but are much less useful as long range indicators for predicting the quality of work expected from the student after he transfers to the University. Seashore (44:76) reports correlations between College Qualifi- cation Test -- total scores and first semester junior college transfer students' G.P.A.s ranging from..26 to .60. The Psychological corpora- tion reports the total score.has greater predictive efficiency of first semester college grades than any of the sub scores. Kirk (34:220) reported correlations between College Qualification Test -- total score and first term G.P.A.s for junior college transfer students at the University of California ranging from .31 to .44. High school and Junior College grade point averages £§_Predictors. It is almost an axion that the best index of a student's probable record in college is a record which he has previously made in college. Reeves (41:121), utilizing grades of junior college transfer students from high school records and from previous college records to predict 25 University of Chicago G.P.A., obtained a coefficient of multiple corre- lation of .685. He indicates that if the high school records and previous college records are known, individual University of Chicago records could be predicted with an accuracy such that 50 per cent of the predicted G.P.A.s would be within approximately half a grade point. Similarly, Siemens (46:27) indicates that the best factors for pre- dicting success after transfer were found to be the G.P.A. in all lower division work and G.P.A. of the first semester work after transfer. It is further stated that through the use of prediction equations it was found possible to forecast upper division academic success for trans- fers such that the predicted G.P.A. would not vary on the average from the actual G»P.A. by more than about .25 of a grade point unit. ‘Rodes (43:126) found that the correlation between grades in lower division Engineering courses and subsequent grades in upper division courses for junior college transfers was .64. He combined previous grades and the 1 to obtain a co- total score of the junior status engineering test efficient of multiple correlation of ;702. The coefficient of correlation between junior college G.P.A.s and University G.P.A.s (.47) reported at Florida State University (21:5) would seem.to indicate that junior college grades are probably the best measure for predicting success, all other factors considered. Summary Since the emergence of the junior colleges in American Education 1See page 23. 26 periodic studies have been.made to ascertain whether the junior colleges are adequately preparing students for further study in higher institu- tions. The general conclusions of these studies has been that junior colleges have been successful in this regard. Almost invariably, the group of junior college transfers considered has been found to do at least as well academically in the latter years at a higher institution as do students in the same fields who have spent all four years at the same insitution. The literature reviewed on the preceding pages tends to support this fact. Junior college transfer students seem to achieve slightly below the level of the non-transfer students but not to any significant degree. The research indicates that junior college transfer students did essentially the same quality of work in the universities as did the non-transfer students. However, most studies on the subject reveal high- er failure and drop-out rates for junior college transfer students than for non-transfers. Among other significant findings in the research is the fact that junior college transfer students go on to do University work com- parable in quality to the junior college work they did before transfer. The research concerning the academic abilities of the junior college transfer student indicates that the better students matriculate to the universities from junior colleges. The best comparisons that can be made between test scores of junior college transfer students and non- transfer students indicates that only small differences exist in favor of the nonptransfer group. The most popular curriculum choices among junior college trans- 27 fer students appear to be Business, Science and Arts, Engineering and Education. (Not necessarily in that order.) The literature seems to assign limited value to aptitude and achievement tests as predictors of academic success for junior college transfer students after they have transferred to the University. According to the literature, a coefficient of multiple correlation utilizing high school record and junior college record and certain test scores seems to offer the best predictive indices. Junior college record was regarded in all studies as the best single predictor of the achievement of the junior college transfer student after transfer. CHAPTER III METHODS AND PROCEDURES This study involves two groups of students: (1) the community- junior college transfer students and (2) the control group consisting of non-transfer students. This chapter describes the procedures used for selecting the samples, collecting and organizing the data, and techniques for analysis. The Samples The total on-campus Michigan State University population for the Fall term, 1958, included 19,516 students. There were 13,139 men and 6.377 women. The junior class consisted of 3,663 students. Data in not available on the total number of transfer students. However, 334 students enrolled from Michigan Community-Junior colleges. The C.J.C. transfer sample was drawn from this group. The non-transfer sam- ple was drawn from the group of 3,663 juniors. _'1he criteria used in selecting the sample are set forth below. :11: smug-m college transfer student 33221.3. The Regis- trar's list of new students for the Fall term, 1958 provided the C.J.C. transfer student sample. This list was coded to indicate "transfer" and "non-transfer" and number of hours transferred. C.J.C. transfer students were selected from this list providing that they had transferred 75 or more term hours. The 1958 group was the largest group of C.J.C. 29 transfer students to enter in a given year. A total of 173 students met the 75 hour requirement, including 137 male and 36 femal students. The control 23521:. The control group involved sampling the non- transfer population since it would hardly seem feasible to utilize the entire population. The criteria for selection included: (1) admission to the University for the Fall term, 1956. (2) being enrolled in a University program for the Fall term, 1958. (3) no transfer credits from other institutions of higher learning. (4) the same sex ratio as the community-junior college transfer group.1 The Registrar's alphabetical list of students enrolled for the Fall term, 1958, was utilized in selecting the control sample. This list provided the names of all students in school for the Fall team, 1958. Student numbers identified those who had originally entered the University for the Fall term, 1956. The list was coded to enable eli- mination of those with transfer credits. Separate lists for male and female students were constructed from the total listing of students enrolled for the Fall term, 1958. The samples were randomly selected utilizing random numbers to allow each individual on the composite lists an equal opportunity to enter the sample. A total of 173 students was selected, including 137 male and 36 female students. The control sample was selected in this manner and in lCoueunity- junior college sex ratio favors the male to a greater degree than the native ratio. 30 accord with the above listed criteria in order to make it comparable to the C.J.C. transfer student sample, first in terms of total college ex- perience and secondly, in terms of sex differences that are known to exist. .An unequal sex ratio: between groups would seriously distort any comparisons and greatly affect measures of differences. For ex- ample, as reported in the literature, grade point averages and test scores are likely to be significantly higher for female students. It should also be pointed out that the statistical analysis will not utilize the entire samples since drop-outs give incomplete data in some cases. Procedures Used in Collecting and Tabulating the Data Sources for collection of £22 2253. Permission was granted by the Registrar for the use of records needed in gathering the data, providing the information was to be handled in a confidential manner. Two sources were required for gathering the information pertinent to this study. (1) the permanent record cards on file in the main records office.1 (2) the record folder containing the application blank and previous school records on file in the records vault. ‘Method used in tabulating. A work sheet was developed for trans- ferring data from.the records. Each individual work sheet was identi- fied by student number. Data for each student we recorded on these 1The permanent record cards are the source of the official transcript of grades 31 individual work sheets. Each item on the work sheet was then coded1 and transferred to an IBM master tabulating form. The data for each indi- vidualwere then key punched on IBM cards. All items were checked to insure accurate recording of data on the work sheets and all tabulations were verified by machine by two different key punch operators. A visual spot check was also made. The specific information and its source. The permanent record cards provided the following information: test scores (derived) hours carried each term grade point averages by term and cumulatively 'Michigan.State University social studies grades IMichigan State University natural science grades military status curriculum drop-out number of transfer credits somqoxynmurow The record folders and the application blank, in particular, pro- vided the following information: 1. age 2. marital status 3. father's occupation 4. high school rank Previous school and college records provided the following infor- nation: 1. community-junior college G.P.A.s 2. community-junior college grades in social studies 3. community-junior college grades in natural science 4. identification of the community-junior college of origin Classification of fathers' occupation and University curriculum. It was necessary to classify certain data for analysis. 2Appendix A 32 Father's occupation: The classification used here was adapted from the Edwards Occupational Index. (4:172) Certain modifications were made to render the instrument more sensitive to differences, par- ticularly in the upper portion of the index. However, the number of classifications was reduced in the statistical analysis because of the nature of the tabulations. The application blank asks for a state- ment of father's occupation. However, several students indicated no knowledge of father's occupation or did not clearly define the position. These and others who indicated "retired" or ”deceased" were placed in a separate class. The classification can be found in Appendix 6. Michigan State University Curriculum. {Assignments were made to major college programs only. In general the classification follows college boundaries with a breakdown for Science and Arts into Linguis- tie and Scientific-Computational. A further breakdown is provided for Science and Arts, Home Economics and Agriculture students working for teaching certificates.1 The classification can be found in.Appendix C. Facsimiles of data sheets and IBM cards are included in the Appendix A. Coding procedures can also be found in Appendix A. Techniques of Analysis Those variables given as continuous data are analyzed by use of the "t” test. Variables given as non-continuous data are analyzed by use of chi-square. Comparison 25 the two groups by ”E? test. The primary objective 1Certain of these classifications were combined for statistical analysis because of the nature of the tabulations. 33 of this study is to determine whether or not the two groups differ sig- nificantly with respect to any of the educational variables and back- ground information. The C.J.C. transfer students and non-transfer stu- dents were compared by "t" test on the following variables: 1. grade point averages a. Fall b. Winter c. Spring d. Cumulative . high school ranks 1 . 7G.P.A.s previous to Fall term, 1958 . Michigan State University social studies G.P.A.s . Michigan State University natural science G.P.A.s “#0419 The ratio 22 differences between the two groups by chi-square. Differences between the two groups with respect to certain background information were obtained by the application of chi-square. Transfers and non-transfer students were compared relative to the following information: . success or failure2 success or failure according to curriculum drop-out veteranp-non-veteran . married-~single . father's occupation GU§UNH e The same chi-square technique was applied to comparisons within the groups themselves. This part of the analysis is a study of the relationships between: 1. number of credit hours transferred and academic success 2. decrease in course load from Fall to‘Winter terms and academic success Winter term. 1Community junior college grade point average for transfers. Michigan State University grade point average for first two years for nonptransfers ZSuccesa: 2.0 G.P.A. or better. Failure: Less than 2.0 G.P.A. 34 Analysis of the Relationship between certain educational variables and academic performance. The study is also concerned with the relationship between certain background factors in the college and university life of the C.J.C. transfer student and his academic success at Michigan State University. The following variables have been analy— zed by use of Pearson's Product4Moment Correlation technique.1 1. the orientation test battery scores with Michigan State University G.P.A.s 2. high school rank with Michigan State University G.P.A. 3. community-junior college G.P.A. with Michigan State University G.P.A. 4. ‘Michigan State University G.P.A. for Bhe first two years of college of the control group with cumulative Michigan State University G.P.A. 5. Michigan State University first, second and third term G.P.A.s with cumulative Michigan State University G.P.A. for the 1958-59 school year. 6. Community-junior college social studies and natural science G.P.A.s with Michigan State University social studies and natural science G.P.A.s for males only. All results will be presented in tabular form in Chapter V. Levels of confidence were derived from tables by Wallace and Snedecor.3 Analysis of variables in combination. It seemed advisable to also analyze the predictive value of certain of these variables in combina- tion. The calculation of the multiple correlation coefficient was . . . . 4 . accomplished through the application of the Doolittle Technique. This technique asks what regression weights best predict the criterion from 1Michigan State University intergral computer operated by the Michigan State University computer laboratory was employed. 2Averages prior to Fall term, 1958. 3See Guilford (27:538-9). 35 the other variables combined and what the correlation of those predic- tions with obtained criterian values would be. The unknowns are the Beta coefficients and there are as many equations as unknowns. The problempin this study involves the following five variables: 12 iMichigan.State University Reading test-~total score x3 AeCeEe Or C.Q.T.--t0t81 score x4 high school rank x5 previous G.P.A. These variables were selected for study because they appeared to be the most significant factors in this study and are also reported as such in the literature. The statistics are presented in tabular form in Chapter V. Summary 1. The population consists of two groups of students: 1) the C.J.C. transfer student and 2) the non-transfer student a) the C.J.C. transfer student sample consists of all those who enrolled at Michigan State University for the Fall term, 1958, and were at the same time granted 75 or more term hours of transfer credit. b) the non-transfer student sample was selected at random from the Registrar's alphabetical list of students enrolled for the Fall term, 1958. Only those who originally enrolled for the Fall term, 1956 were included. Thus, the two groups have been engaged in a college program over a comparable period of time. c) Each sample included 137 males and 36 females. Nermal sex differences in connection with educational variables dictated separate analysis by sex. 36 2. Data were gathered from the permanent record cards and the records folders in the office of the Registrar. 3. All of the data were key punched on IBM cards and prepared for machine analysis. 4. The following points concern the statistical procedures employed in this study. a) Basic statistical tabulations were made by use of IBM equipment. b) "t" tests were used to test the differences between the groups on these same variables. c) The ratio of differences between the two groups and within the groups relative to specific background information were analyzed by application of chiusquare. d) The zero order correlations were computed to measure the 1 degree of relationship between a number of educational variables and overall academic success. Zero-order ocrrelations were also computed to measure the degree of relationship between the educational variables and academic success in social studies and natural science. Means and variances were computed for all of the variables. Appropriate tables were consulted to determine significant; of r. The standard error of r was derived from the appropriate formula. e) Further statistical analysis involves computation of mul- tiple correlation coefficients, beta weights and the multiple regression equation for predicting individual G.P.A.s 1test scores, high school rank, community-junior college Fall, Winter, Spring and Cumulative G.P.A.s, social studies G.P.A.s, natural science G.P.A.s, Basic College G.P.A.s. CHAPTER IV ANALYSIS OF THE DATA ON THE DIFFERENCES BETWEEN THE GROUPS It has been hypothesized in Chapter I (p.3) that the C.J.C. transfer student is different from his non-transfer counterpart in respect to a number of educational and personal characteristics. This chapter is concerned with the analysis of the differences between the C.J.C. transfer student and the non-transfer student relative to certain educational variables and background factors. The Differences Between the Groups Fisher’s "t? test Fisher's "t" test was applied to determine whether or not the two groups differed significantly with regard to the following: (1) G.P.A.s for Fall, Winter and Spring terms, 1958-59. (2) cumulative G.P.A.s for the three terms under study. (3) high school rank (4) college G.P.A.s for work done prior to Fall term, 1958. (5) Michigan State University social studies G.P.A.s (For male only) (6) Michigan State University natural science G.P.A.s. (For male only) Results of this analysis are summarized for the male and female students, respectively, in Tables I-A and I-B. TABLE I-A Means, Standard Deviations and t Values on Selected Variables for Male Students 38 Transfer Non-Transfer t1 N ‘3 0r N 'i C’ M.S.U. Fall 123 2.38 .71‘ 127 2.44 .71 1.78 GOP 0A0 ‘M.S.U.‘Winter 123 2.54 .65 127 2.52 .65 - G.P.A. u.s.u. Spring 123 2.51 .71 127 2.54 .66 - G.P.A. M.S.U.Cumulative 123 2.48 .53 127 2.49 .50 .- G.P.A. High School Rank 123 7.58 2.67 127 8.15 2.32 1.81 College G.P.A. Prior to Fall, 123 2.54 .44 127 2.53 .47 - 1958 Social Studies 60 2.38 .69 50 2.37 .68 - G.P.A. Natural Science 51 2.35 .75 551 2.18 .78 1.13 G.P.A. 1Values less than 1 are not reported. *5 per cent Level of Significance ** l per cent Level of Significance Results of the ”t" tests. Table I-A indicates that none of the "t" tests were significant for males, although some were of a borderline nature. The "t" test value of 1.78 between the Fall term G.P.A.s of the two groups borders on significane and may be interpreted as an indi- 39 cation of a trend of higher achievement for non-transfer students than TABLE I-B Means, Standard Deviations and t Values on Selected Variables for Female Students Transfer Noinransfer t1 N 3E 0' N '1? U M.S.U. Fall 32 2.18 .85 32 2.63 .70 2.32* G.P.A. e M.S.U. Winter 32 2.44 .60 32 2.59 .48 1.12 G.P.A. M.S.U. Spring 32 2.59 .63 32 2.66 .54 - G.P.A. M.S.U. Cumulative 32 2.41 .55 32 2.61 .44 1.54 G.P.A. High School Rank 32 8.50 2.80 32 9.13 2.05 1.00 College G.P.A. Prior to Fall 32 2.69 .46 32 2.50 .43 1.71 1958 1Values less than 1 are not reported. * 5 per cent level of significance. **1 per cent level of significance. for C.J.C. transfer students. Similarly, the fit” test value of 1.81 between the high school ranks of the two groups borders on significance and may be interpreted as a trend of higher high school achievement for non-transfer students than for C.J.C. transfer students. Table I-B indicates that Fall term G.P.A.s for females, are significantly different between the two groups. The "t" of 2.32 was found to be significant beyond the five per cent level of confidence. .321 in! mfiw - YE 40 None of the other "t" tests were found to be significant. However, the same trends which were found for the males appear to be operating in the case of the females. The ”t" tests would seem to infer the existence of certain similarities between the two groups. These results tend to support the findings of other studies reported in Chapter II which indicate that junior college transfer students achieve at approximately the same level as the non-transfer student. However, the significant dif- ference between the Fall term G.P.A.s of the two female groups will require further comment in Chapter VI. Chi-ssuare Analysis Chi-square technique was applied to determine whether or not the two groups differed significantly with regard to the following I variables: (1) The number of students who fell below a 2.0 G.P.A. for the three teams under study (All University). (2) The number of students in the College of Business and Public Service who fell below a 2.0 G.P.A. for the three terms under study. (3) The number of students in the College of Engineering who fell below a 2.0 G.P.A. for the three terms under study. (4) The number of students in the College of Education who fell below a 2.0 G.P.A. for the three terms under study. (5) The number of students in the College of Science and Arts who fell below a 2.0 G.P.A. for the three terms under study. (6) The number of drop-outs. 41 . (7)‘Military status (8) Marital status (9) Father's occupation The tables that follow make it possible to present the actual frequency distributions of the students used in the study in each of the variables listed above. Explanation of the chi-square tables. The chi-square tables were constructed as follows: For example, in Table II, the factors success-failure are listed in the first column one above the other, (above 2,0, below 2.0). The second column presents the actual or observed frequencies of the transfers. The third column is identical to the second except that these data are for the non-transfer group. The third colmmn gives the total number of frequencies found in the sample. Indication of significance levels. Chi-square totals bearing a double asterisk (**) indicate that the result is significant at or beyond the one per cent level of confidence. A single.asterisk (*) indicates the five per cent level of confidence. Throughout the study only the l and 5 per cent levels of confidence were considered. Thblss without asterisks present factors that cannot be considered significant for use in describing differences between the two groups. levels of significancewere taken from Table III of Fisher's "Statistical Methods for Research Workers”.1 lSee Guilford (27:540). 4? Results 25 the flésquare Analysis While all evidence points to difference in performance between male and female, the sexes were treated as a group in the chi—square analysis. Separate chi-square analysis was made to determine the influence of sex differences on the success-failure ratio of the two groups. These analyses indicated that sex differences had little bearing on the success-failure ratio of the two groups. It must also be remem- bered that there is an equal proportion of male to female in the samples. The frequency data for the above may be found in Appendix E. It is also known that veterans and married students generally achieve higher than other students. Chi-square analysis indicated that there was little or no influence on the part of these variables in so far as the success-failure ratio is concerned for these groups. It is possible that with larger samples some differences might be found with respect to these variables. In view of these findings the following chi-square analyses treats male and female as a group and will not con- trol marital or military status. The ratio of success-failure in the university program between C.J.C. transfer students and non-transfer students. The chi-square total, presented in Teble II, of 8.163 indicates that this factor was significant beyond the one per cent level of confidence. The data indicate that C.J.C. transfer students had a significantly higher inci- dence of failure than the non-transfer student. This fact would seem to contradict earlier findings of no significant differences in achieve- ment between the two groups. 43 TABLE II The Success-Failure Ratio Between.C.J.C. Transfer Students and NonJTransfer Students Regardless of Curriculum or College WM Transfer Non-Transfer Totals G.P.A. Above 2.0 129 150 279 G.P.A. Below 2.0 44 23 67 Totals 173 1 73 346 2 X = 8.163“? The ratio of success-failure in the college of Business and Public I“ Service between C.J.C. transfer students and Eon-transfer students. The ‘chi-square total, presented in Table III, of 6.703 indicates this factor is significant beyond the one per cent leveltof confidence. The data indicate that C.J.C. transfer students have a significantly higher inci- dence of failure in the College of Business and Public Service than do the non—transfer students. TABLE III The Success-Failure Ratio Between C.J.C. Transfer Students and Non—Transfer Students in the College of Business and Public Service Transfer Non-Transfer Totals G.P.A. Above 2.0 30 40 70 G.P.A. Below 2.0 18 6 24 Totals 48 46 94 7t 2 a 6.703** 44 The ratio 2£.success-failure in the College 35 Engineering between C.J.C. transfer students and non-transfer students. The chi- square total, presented in Table IV, of 1.621 indicates that this value should not be used to describe difference between the two groups. The data indicate a slightly greater incidence of failure in the College of Bngimering for the non-transfer group. Interpretation is limited considerably because of the small sample involved. TABLE IV The Success-Failure Ratio Between C.J.C. Transfer Students and Non-Transfer Students in the College of Engineering Transfer Non-Transfer Totals G.P.A. Above 2.0 25 13 38 G.P.A. Below 2.0 5 5 10 Totals 30 18 43 x2 - 1.6211 1Yates correction used in the computation of X2. (27 ) The ratio of success-failure in the College 3_f_ Education between Eh: 931:2. transfer students 1'31 Egg-transfer students. The chi-square total, presented in Table V, of 1983 indicates that this value should not be used to describe difference between the two groups. The data indicate a slightly greater incidence of failure in the College of Educa- tion for the C.J.C. transfer student. Interpretation is again restricted by the small sample. . 45 TABLE V The Success-Failure Ratio Between C.J.C. Transfer Students and Non-Transfer Students in the College of Education Transfer Non-Transfer Totals G.P.A. Above 2.0 34 46 80 G.P.A. Below 2.0 6 2 8 Totals 40 48 88 2 X t 1.9831 1Yates correction used in the computation of x2. (27) £22225 success-failure 3.2 3112 College of Science and 523 between £2129; transfer students and Ben-transfer students. The chi- square total, presented in Table VI, of 3.120 indicates that this value should be interpreted with caution. The data indicate a some- what higher incidence of failure in the College of Science and Arts for the C.J.C. transfer group. 46 TABLE VI The Success-Failure Ratio Between C.J.C. Transfer Students and Non-Transfer Students in the College of Science and Arts ‘_.— ”W L” —-~—-——‘ 4—— IN Transfer Non-Transfer Totals G.P.A. Above 2.0 40 45 85 G.P.A. Below 2.0 15 7 . 22 Totals 55 52 107 X2 = 3.120 The ratio of drop-outs between C.J.C. transfer students and non- transfer students. The data in Table VII indicate that there is a slightly higher drop-out rate for C.J.C. transfer students than for non-transfer students. However, the chi-square total of 2.694 is not significant and the data should be interpreted accordingly. TABLE VII The Drop-out Ratio Between C.J.C. Transfer Students and Non-Transfer Students Transfer Noinransfer Totals Drop-outs 32 21 53 Non-drop-outs 141 152 296 Totals 173 173 346 x2 = 2.694 47 The ratio 25 married and single students between the C.J.C. transfer students and the Egg-transfer students. The chi-square total of 14.89 in Table VIII is significant beyond the one per cent level of confidence. The data indicate a much greater proportion of married students in the C.J.C. transfer group. In view of earlier findings (p:“b, this difference, while requiring interpretation, does not greatly influence the other variables being studied. Further discussion of this factor will be included in Chapter VI. TABLE VIII The Married-Single Ratio Between C.J.C. Transfer Students and the Non-Transfer Students Transfer Noinransfer Totals Married 29 7 36 Single 144 166 310 Totals _ 173 173 346 X2 :- l4.89** The; £223 21: veteran and lion-veteran students between the C.J.C. transfer students and non-transfer students. It was anticipated that the non-transfer group would have a greater ratio of veterans to non- veterans. However, the exact opposite seems to be true. The chi- square total of 15.04 in Table IX is significant beyond the one per cent level of confidence. The data indicate a much larger proportion of veterans in the C.J.C. transfer group. Further, this factor does not appear to influence the success-failure ratio. This factor will be 48 discussed further in Chapter VI. TABLE IX The Veteran-NonFVeteran Ratio Between the C.J.C. Students and the NOn-Transfer Students L Transfer Non-Transfer Totals Veteran 46 18 64 Noneveteran 127 155 282 Totals 173 173 346 X 2 = 15.04** The ratio of fathers in six different occupational classifications between the C.J.C. transfer students and the non-transfer students. The choice of a community-junior college for the first two years is often a matter of financial necessity. For this reason a difference would be . expected between the socio-economic status of the two groups. The chi- square total in Table x of 32.85 is significant beyond the one per cent level of confidence. The data indicate a significantly higher occupa- tional ranking for the fathers of the non-transfer students. The table is influenced in favor of the nonétransfer group partly because of the 20-2 ratio in category VII (Unknown). However, there is reason to believe that these unknowns would not reduce the chi-square value by any significant degree, if known. Indications are that most would fall in categories III to V1. This would tend to increase the value of Chi— square. 49 TABLE X The Ratio and Percent of Fathers in Six Different Occupational Classifications Between the C.J.C.1 Transfer Students and the Non-Transfer Students Transfer Non-Transfer Totals N Per Cent N Per Cent I Professional 21 12 52 30 73 II Managerial & Farm 43 25 53 31 96 III White Collar 22 13 17 '10 39 IV Skilled Labor 35 20, ~22 13 57 V Semi-Skilled 10 6 l2 7 22 Labor VIA Unskilled Labor 22 13 15 s 37 VII Unknown 20 11 2 1 22 Totals 173 100 175 100 346 )12 = 32.854** In addition to the foregoing analysis, certain other data was deemed important to the study. Therefore, the following information is included to aid in the description of the population. Age. The age of each student was computed as of October 1, 1958, utilizing the date of birth which was recorded on the application form. 1Classifications I and II from the original data were combined for purposes of analysis since the ”N" for classification I was quite small. The same procedure holds for classification III and TV from the original data. 50 The average age for the male C.J.C. transfer student was twenty-two years, two months. The average for the male non-transfer student was twenty years, seven months. The average age for the female C.J.C. transfer students was twenty years, three months. The average for the female non-transfer student was twenty years, one month. The data indicate a one year and five months difference between the males and a two month difference between the females. Torm.hours carried. Table x1 indicates the average number of term hours taken by term and total term hours earned for the three terms. The data indicate the two groups carried approximately the same number of term hours per term and earned the same number through the third term. TABLEXI The Average number of Term Hours Taken C.J.C. Transfers Non-Transfers Fall 15.7 15.3 Winter 15.3 15.1 Spring 15.6 15.9 Cumulative 44.4 44.9 Term hours earned previous to Fall term, 1958. The C.J.C. trans- fer studenttransferred an average of 93.9 term hours of credit from the community-junior college. The non-transfer student earned an average of 97.5 term.hours of credit previous to the Fall term, 1958. The two 51' groups seem comparable with regard to college credits earned previous to Fall term, 1958. 32311 college credits gigging SM 3352, 1222. The C.J.C. transfer student completed the third year with an average of 138.5 term.hours to his credit. The non-transfer student completed the third year with an average of 142.7 term hours to his credit. The two groups seem to be comparable in this respect. Summary The differences between C.J.C. students and non-transfer students were tested and sumarized. The testing of the differences involved the application of Fisher's "t" and chi-square techniques. The "t" tests were computed on six variables with one set of computations for males and a separate set for females. The "t" tests between the means of Fall term G.P.A.s for females (2.32) was found to be significant at the 5 per cent level of confidence. No other differ- ences were found to be significant by this method. Further analysis of the differences between the groups was accomplished by application of the chi-square technique. Several of the chi—square ratios were found to be significant beyond the one per cent level of confidence. The differences revealed by this method are as follows: (1) A significantly greater number of C.J.C. transfer students than nonptransfer students fell below 2.0 G.P.A.s for the three terms under study. (2) A significantly greater number of C.J.C. transfer students than non-transfer students in the College of Business and Public Service 52 fall below 2.0 G.P.A.s for the three terms under study. (3) There is a significantly greater ratio of veterans in the C.J.C. transfer group than in the non-transfer group. (4) There is a significantly greater ratio of married students in the C.J.C. transfer group than in the non-transfer group. (5) The fathers of the non-transfers as a group have higher occupational status than the fathers of the C.J.C. transfer students. Some of the chi-square ratios revealed differences which are not to be considered significant. The following differences are not significant: (1) The number of C.J.C. transfer students who fell below 2.0 in the College of Engineering is not significantly different from the number of non-transfer students who fell below 2.0. (2) The number of C.J.C. students who fell below 2.0 in the College of Education is not significantly different from.the number of 'non-transfer students who fell below 2.0. (3) The number of C.J.C. transfer students who fell below 2.0 in the College of Science and Arts is not significantly different from the number of nonptransfer students who fell below 2.0. (4) The number of C.J.C. transfer students who drop out of the University is not significantly different fromzthe number of non-trans- fer students who drop out. (5) Sex, marital status and military status do not appear to affect the success-failure ratios of the two groups. The male C.J.C. transfer students are on the average one year and five months older than the male non-transfer students. The female students show a slight variation in age between the two groups. ti: 53 The two groups earned approximately the same number of term hour credits during the period under study and had also accumulated approxi- mately the same number of term hour credits previous to Fall term, 1958. Small differences are seen in the comparison of curriculum elec- tions between the two groups. CHAPTER V ANALYSIS OF THE DATA ON THE RELATIOI‘SHIP BE'IWEEN SELECTED EDUCATIONAL VARIABLES AND ACADEMIC SUCCESS The Procedure The secondary objective of this study was to measure the degree of relationship between selected educational variables and the 1 The procedures academic success of the C.J.C. transfer student. utilized here for determining the predictive value of the selected variables which, either singly or in combination, maximize predictive efficiency are: (l) calculations of zero-order coefficients of correlation (2) calculation of standard partial regression coefficients (Beta weights) (3) calculation of the multiple correlation coefficient (4) setting up of the multiple regression equation which predicts the G.P.A. This chapter will present the results of these computations in an effort to shed some light on the predictive value of certain variables where C.J.C. transfer students are concerned. 1Parallel study accomplished for non-transfer sample. 55 The Zero-order Correlation Coefficients for Selected Variables Related to Academic Success for All Courses Taken The variables As stated previously, the independent variables that were considered in comparison with the Michigan State University cumulative G.P.A. are : (1) (2) (3) (4) (5) (6) (7) derived scores on the college Qualification test-- V score, I score, N score and Total score.1 derived scores on the American Council on Education Psychological Examination--Q score, L score and Total score.2 derived score on the Michigan State University Reading Test.3 derived score on the Michigan State University Test of English Usage.4 high school rank previously acquired G.P.A.s5 Fall, Winter and Spring term.G.P.A.s. 1C.J.C. transfers only. 2 Non-transfers only. 3 4 Different form used for C.J.C. transfer students C.J.C. transfers only 5C.J.C. G.P.A. or Basic College G.P.A. 56 The Zero-Order Correlation Coefficients. Means, standard deviations and zero-order correlation coefficients for all courses taken are presented in Tables XII-A, XII-B, XIII-A and XIII-B. Male and female were treated separately since differences in performance might be expected between the sexes, particularly with respect to test scores. Before considering the findings, these correlations should be considered in light of other studies, reviewed in Chapter II, which have dealt with the relationship to college academic achievement of the variables being considered here. Care should be taken to remember that in this instance the dependent variable, cumulative G.P.A., is the average for the three terms of the junior year. The significance of the correlation coefficients found in this study were determined from Snedecor's table, "Correlation Coefficients at the 5 per cent and 1 per cent Levels of Significance";1 utilizing degrees of freedom as indicated by "N" in each table. Correlation coefficients followed by "a" are not considered significant at either 1 per cent level or the 5 per cent level. Correlation coefficients followed by "b" are significant at the 5 per cent level but not at the l per cent level. All others are significant at or beyond the one per cent level. The standard errors of the zero-order coefficients of correlation are listed in Appendix D. 1see Guilford (27:538-9). 57 .Ho>mH peso woe .ao>oH have use H on» vacuum unmoauwewwm one euomuo HH< us no: use ueoo pom m on» as unmouuwmwwm ea umowoflumooo cowumaouuooa .~o>o~ ammo use m may no unmowuacwwe no: aw umowommmooo mowumaouuoOm me. me. he. mm. om. em. sea. mm. ewe. an. mm. mm. we.m .<.m.e a>auaeaa=o «a . em. mm. mm. see. «we. «mo. «he. «oh. «a. we“. as. Hm.~ .<.m.e museum ”a 1 cm. mm. em. new. ma. em. mam. em. ma. me. em.~ .<.m.o nausea oh . as. am. mm. «he. mm. on. an. an. as. umm.~ .<.m.e Heme o . sq. see. who. am. mom. on. em”. ee. em.~ .<.m.o .o.n.o m a me. mm. mm. mm. an. mm. ae.~ mm.a emu a n no. me. as. mm. on. on.“ em.e 9.900 e u me. moo. mm. and. mo.~ mn.e 2.900 m u as. an. an. nn.a nm.e HJsoo a a we. on. “as.” ~o.e >Jaoo m 1 mm. on.” mu.e unga.w=aeaom m .. we.“ E uuomwfifiwcu H 2 S a m a. e m s n m a .c.m m. WHZHQPHm Egg menu—HOD mow—29732330 NASA nun mo 33m mom x3942 zogfigoo mun—gnomes 92 E05535 323$ .93.? (Id Hun saga. 58 .123; 0:00 use d on» .Ho>o~ uses use a on» nachos unmowuwewwe one euoauo n~< an ac: use name use m one as usaowuwcwwe ea unowowuuooo soauesoueooc .He>o~ undo use m can as usaowuwemwm no: ow unohowmuoOo sowuoaouuoom om. on. on. he. on. an. mm. mm. mm. saw. nae. on. as.” m>suawsmsw «a . oe. om. mm. am. an. ma. em. em. new. now. on. em.m .<.m.o waauem a“ - mm. on. an. on. mm. mm. nae. am“. as”. no. mm.~ .<.m.o saunas ca . me. mm. mm. em. «a. «we. «we. mod. as. e¢.~ .<.m.o same a .ou» w pea . on. an. we. as. as. om. mm. es. mm.~ .<.m.e pm: m u an. as. on. mm. on” ma. mm.~ ma.m mm; s . am. oo. «a. me. we. oo.~ mn.m euwehemom o u an. ac. no. He. an.“ oe.m owmasemam m u we. on. me. an." me.m swmeheaom e . mm. as. on.” ms.n anmo< m - as. ee.a ne.n enmo< m n so.” as.“ oumo< a nu an a MW 5 o HHMW‘ e n u H, .n.m 11hr eoanmwue> Sam mom KHMHSA zogjmmmoo MMQMOIOMNN % EOEMG DES—um .Wm mazmnpam ummmzema essence monphrastezzoo mq;l,‘.l h.-I-IP:. W‘ VQUI.)I.IA.‘/h(§\yl. 2e .7. v...’\ith-rni e n . e a . a e . . . n o . o - o v o a . I c . a . . . 62 .HgoH usmo use H emu um uoe use Hc>oH uemo use m meu um unmouuustm mu usouoummooe scuuoHouuoo .He>oH ueso use m oeu um unmouuusmHm uo: mu usapHmuooo eoHumHoupoom .Hs>oH usoo use H oeu secede unmouuwmwum one museuo HH< e as. as. me. use. can. «ma. non. emu. mun. «mo. sq. ue.~ o>anmusmsw me . «mm. mm. men. es. nos. ms. ems. ems. use. em. oe.~ .<.m.o wnuuem Ha u as. can. «no. «mo. moo. «mo. cue. mo. we. om.~ .<.m.e houses oh as. say. new. mom. nan. an”. new. «mo. on. no.~ .<.m.e Haas a .eu» N ueH n as. Hm. es. en. ems. em. sac. me. on.~ .<.m.o ems. m n «me. see. new. man. «on. now. so.~ nu.a ems a n as. no. on. as. an». we.u «a.m sawsaeaom e u on. em. we. can. so.H «s.m cumuaeaom m n as. om. non. e«.H Ne.m swweueaoa s u am. as. ne.H he.“ ensue n a an. no.u ee.m sumo< a u on.“ -.n onmu< a "a oh 0 a e . n e n a H .e.m m. nouausuo> Emnam NEWAEIZE a N». m0 wage mom x352 239558 genomes 92 mzoE‘Ee gm .932 QIH HHX SQSH 63 perform in subsequent college work. The correlations between cumula- tive G.P.A. and previous G.P.A. (.72, .65) are high enough to be considered as a good predictive measure. The zero-order correlation coefficients between cumulative G.P.A., high school rank and derived test scores show trends which indicate somewhat better predictive value for these variables when applied to female students than for male students. It must be remembered that the small samples place definite restrictions on the interpretation. The Relationship Between Selected Variables and Social Studies Grade~Point Averages The zero-order correlation coefficients. The following analysis is restricted to samples of 60 male C.J.C. transfer students and 50 male non-transfer students (G.P.A.s were computed only for students who had earned a minimum of 6 term hours in Social Studies). Tables XIV-A and XIV-B present means, stan- dard deviations and zero-order correlation coefficients for these samples. The data are presented for the purpose of determining the relationship between selected variables and future achievement of stu- dents in social studies. The small sample size places definite restrictions on the interpretation. The analysis was accomplished for the purpose of indicating trends. These results indicate that previous grades are the most efficient forecasters of future achievement. The relationship between C.J.C. social studies G.P.A. and University social studies G.P.A. (.49) and between previous G.P.A. for all courses taken and University social studies G.P.A. (.64, .73) seem to bear this out. The zero-order correla- 64 .HgoH uses use H meu nuance unsound—m: Pas auoeuo HH< .HgoH ueoo use H emu us uoaa use uaaoo use m eeu um 333.25? 3 useuouuuooo souusHeuaooe .HeesH usoo use m eeu um useoHuHaawHe uoe ea useHoHuuooo soHumHonuoom. .<.m.U mowvsum am. em. as. am. so. new. sea. ems. sea. now. an. new. as. ma.~ asaoom pmz.na .<.m.e anaesum . an. as. as. as. on. an. sea. as. as. mm. nan. an. mm.~ aaaoom .o.n.o «a - as. no. as. an". aoa. aaa. ,uoo. «no. sea. «ma. me. as.“ .<.m.e maausm aa 3 es. an. nan. «an. «aw. aha. cam. on. non. so. om.~ .<.m.e aou=a3.oa - mm. saw. emu. sea. aaa. as". nan. nan. me. nn.~ .<.a.o aaam a a an. msa. «so. asm. ema. on. mam. on. mm.~ .<.m.o .o.no m - an. on. as. on. sn. emu. ne.~ os.a ems a u so. as. ms. 1 an. an. en.a ma.s auaoo e . ss. use. ssa. mma. mm.a se.n 2.900 m u as. oo. as. a~.a so.e aueoo s n am. no. oe.a on.e nsweoo n . mm. as.a as.» page waaeaom a . as.a ~n.n some aaaawau a «a aa 3 m m a e m s n m a .e.m m. moaning 288m .508 3 E235 smegma. Meade... dog—7322380 mass 8 mo Seem mom x222 53558 350.8% 92 Eoflfisme geafim .235 HX Saga. 65 .Ho>oH ueoo .agoa 2.8 use a flu Baas. «sundae»: 8a 823 aa< use H eeu uos use usoo use m aeu us ueaoHuuewHe eH unoQOHmuooo moHuoHeauooe .He>eH ueeo use n eeu um useoHuaste uom eH ueeaoHuueoo soHueHeuuoow sac «nausea on. now. ms. me. an. as. nan. ms. sea. asa. sea. on. sn.~ auaoom smz_~a - ass. pan. on. mom. an. nan. an. on. sun. on. so. as.~ .<.m.o maauam aa pea. on. new. pea. men. can. ona. ana. ans. on. ~n.~ .<.m.o enemas oa . saw. saa. «aw. «so. «so. aao. new. no. os.~ .s.a.o aaam s .eh» N ueH . an. an. ms. an. as. «s. as. ss. os.~ «as am: e s an. as. ss. nan. am. can. ~n.~ no.5 ems a u mm. «a. on. so. an. «n.a oe.m auwaaeaom e a op. an. mm. as. mm.a «o.n ouweaeace m n me. as. an. ue.a on.m swwuaeaam s n as. «o. ss.a «o.m e-uu< n u on. as.a os.m numus a u am.a «o.m ounce a HH 2 a m a. m n e n N H .9. m M eeHeoHue> mama—Pam 300m ZH ENE-Hm $253322 an MC Sea eon sags: 2395538 $59.83 92 flange 9325.» .262 ml>HK Ema 66 tion coefficients between University social studies G.P.A. and derived test scores are relatively low and difficult to interpret in terms of the small samples involved. The Michigan State University reading test scores with r's of .33 and .45 would seem to be of greatest value. In general the test scores seem to have limited predictive value for this level of achievement. The Relationship Between Selected Variables and Natural Science Grade Point Average The zero-order correlation coefficients. The following analysis is restricted to sample of 51 C.J.C. transfer students and 51 non-transfer students. G.P.A.s were computed only for students who had earned a minimum of 6 term hours in natural science. Tables XV-A and XV-B present means, standard deviations and zero- crder correlation coefficients for these samples. These data are presented for this purpose of determining the relationship between selected variables and future achievement in natural science. The small sample size again places definite restrictions on the interpretation. The analysis was accomplished for the purpose of indicating trends. These results indicate further that previous grades are the most efficient forecasters of future achievement. The relationship between C.J.C. natural science G.P.A. (.42) and between previous G.P.A. for all courses taken and University natural science G.P.A. (.51, .48) are considerably higher than for any other variables included in the analysis. For example, the derived test scores provided low rela~ tionships with zero-order correlation coefficients, ranging from .18 to .28. 67 .22: .253 uses use H seas accuse unsouuHaawHe sus euseuo HH< usso use H eeu uos use uses use m eeu us unscHuHewus 3 333380 mOHusHsuuooe .Hs>sH uses use n seu us unscHuusmHu uce sH useHOHuueoo eoHusHsuuoos «s. on. «ma. sm. an. mam. «ma. a8. new. sea. «on. n... ms. and am; pm: na 55 «830m .. saa. saa. on. no. as. on. «mm. as. ass. «8. pan. on. as.~ ahaz 08 «a .. nan. as. R. ama. mac. «3.. sea. «3. 63. «no. 3. ssé .<.u.s scram aa .. saa. «2. sea. «3. sea. «as. «so. «nu. «ma. as. and .<.m.c use; 3 .. an. «nu. ma. sea. an. an. an. an. 2.. and .<.m.s aaau a .. .3. R. saw. on. new. new. can. an. and .s...a.e .o.a..o a. .. sun. can. 1a. «2. sea. an. as." 2..» mm: a. .. an. as. as. an. as. 9.1a 2.6 uuuoo e .. new. «ma. saw. now. so; as.» 2:98 a .. on. an. R. on; and Teen s u as. as. 8.« SK 2.98 n .. an. $.a and «spa. weave-m a .. aha sa.n use. auaauau a «a aa 3 a o a o n s n n a .a.m x 333...; MguHOm ASH—Hag 2H augm Egg M03900 gagpflHH—g Hm mo saga mom mamas: Sausages $5983 92 982.538 gm .235 41.3" an. 68 .HsbsH usso use H Bu accuse unscHuHust sus suseuo HH< .Hs>sH uses use H eeu uos use uses use m eeu us unsound—mus sH ussHsumusco souusHsuuooe .Hs>sH uaaso use m eeu us unsouuuewus use nu ussHoHuusoo soHusHsuuous . <8 sonsH cm as. ss. as. as. os. asa. ana. soa. sou. ssa. sou. as. oa.u a.u.z ems. ua . as. on. us. eon. sua. sea. «ma. aua. «so. qua. ss. um.u sue waauum aa - mm. os. on. sau. sea. «ma. ssu. snu. smu. as. ns.u sue uoaaus oa - as. msu. «nu. «ma. ssu. sou. sea. ssa. em. os.u sue aaau a .uus u uua u as. as. as. ms. ss. on. sun. ss. ne.u sue an: m a sun. on. emu. on. as. s». oa.u as.» mm: s u as. us. as. us. an. so.a ms.n saunasaau s 1 ms. . as. us. an. os.a ne.n onwaaeaou m a as. ns. ns. ns.a sn.n smwaasscu s . co. es. as.a ms.n muses n - mm. na.a nm.n awuus u n ma.a ms.n ounce a aa 3 u o s e n s n u a .né N 333.55 mEmHom gaz zH gunman $3.22 Hnmo 53m so... fleas: 53.3538 Eagauu Es macausasun 533.5 .25.: mien Mama 69 The Multiple Correlation Coefficient For Predicting Cumulative Grade Point Average The multiple correlation coefficient (R) has been computed for males only since the female sample does not have a sufficient "N" to utilize this technique. IE: variables. The variables'used in the solution of "R" are as follows: x2 --‘Michigan State University Reading test-~tota1 score X3 -- College Qualification Test--total score or ACE Psychologi- cal Examination —- total score X4 -- High School Rank X1 --'Michigan State University cumulative G.P.A. Computation of "R". The Doolittle method was utilized in de- l riving the Beta weights shown in.Table XVI. The coefficient of multi- ple correlation is expressed as: R2 3 p12 r12 4- p15 4- p14 1‘14 + 315 r15 R2 is therefore the sum of the products of Beta times its corresponding r. The multiple correlation coefficient for C.J.C. transfer students (.583) indicates a small amount of contribution from variables X2, X3 and X4. The difference between r1.5 (.526) and r (.583) being 1 .2345 .057. The multiple correlation coefficient for non-transfer students (.621) also indicates a small amount of contribution from variables X2, x and x . The difference between r1 5 (.612) and r1.2345 (.621) is 3 4 .009. The data indicate that previous grades have the highest degree 18cc Guilford (27:406-10). 70 of relationship with cumulative G.P.A. and that the addition of other variables did not change the magnitude of the correlation greatly. The contribution of each of the selected variables. (1) Comparison of the Beta weights, or standard partial regression coefficients, in Table XVI reveals that weighting attached to previous G.P.A. (.46 and .56) is large and considerably greater than for the other independent variables. (2) The weighting attached to the Michigan State University Reading Test (.23 and .16) is relatively small and contributes little to the equation. (3) The weighting attached to the College Qualification test and the American Council on.Education Psychological Examination (.06 and -.11) is extremely small and contributes very little to the equation. (4) The weighting attached to high school rank (-.03 and .03) is extremely small and does not contribute to the equation. This study of values in Table XVI indicates clearly that the weighting to be attached to the previous G.P.A. (.46 and .56) for pre- dicting cumulative G.P.A. in all courses taken is much greater than for the other three variables, reading test scores, CQT—T scores or ACE-T scores or high school rank. Estimation of the cumulative G.P.A. of individual students: The General equation Estimating the cumulative G.P.A. of individual students is made possible through the construction of regression equations utilizing the basic data provided in Table XVI. 71 TABLE XVI Summary Data Used in the Prediction of Cumulative Grade Point Average for All Courses Taken by 123 Male Communitquunior College Transfer Students and 128 Male Non-Transfer Students M.S.U. CQT High School Previous Cumulative Statistic Reading Test Total Rank G.P.A. G.P.A. (A) x2 x3 x4 x5 x1 Mean 6.29 6.37 7.58 2.54 2.48 3.0. 1.58 1.50 2.67 .44 .53 13 .2326 .0628 -.0291 .4574 b .0781 .0222 -.0057 .5520 r .3941 .2617 .2873 .5258 R1.2345 = .583* a = .4899 (B) Mean 5.55 5.73 8.15 2.53 2.49 s.n. 1.69 1.59 2.32 .47 .50 a .1557 -.1138 .0299 .5612 b .0464 —.0359 .0065 .6099 r .3695 .2305 .3625 .6115 R1.2345 = .621* a = .8416 (A) Community-junior College transfer (B) Non-transfer "a" constant a in the multiple regression equation or the mean of the X1 values minus the products of other means times their corresponding b weights as, a I M1 - b12 M2 - bn‘M3 - b14M4 - b M 15 5 * "R" significant beyond the one per cent level of confidence. I?! 72 Effect of Reduced Course Load on G.P.A. _Th_e. Chi-square ana lysis The analysis of this factor was accomplished by use of chi— square technique since the data seemed most adaptable to that method. The chi-square ratio of G.P.A. fluctuations between students who took at least three fewer term hours in the Winter term than they had taken in the Fall and students who took the same number or more term hours in the Winter than they took in the Fall is presented in Tables XVII and XVIII. Column 1 describes the trend for G.P.A.s (increased in.Winter over Fall, remained the same or decreased in Winter over Fall.) Column 2 presents the number of actual or observed frequencies for those taking less hours in'Winter than in Fall term. Column 3 is the same except that it deals with those who took the same or more hours in Winter term than they did in the Fall term. Column 4 gives the totals of the observed frequencies. Th: results gg‘thgsghi-square analysis. Table XVII indicates a chi-square total of 13.72 for the C.J.C. transfer student. The data indicate a significant chi-square value between reduction of course load and increased G.P.A. The chi-square total was found to be significant beyond the l per cent level of con~ fidence. Table XVIII indicates a chi-square total of 13.68 for the non- transfer student. The data indicates a significant relationship between reduction of course load and increased G.P.A. The chi-square total was found to be signficant beyond the l per cent level of confidence. TABLE XVII The Relationship Between Reduction of Course Load and G.P.A. for C.J.C. Students 73 Increase in G.P.A. Fa11AWinter Less hours in Winter term than in Fall term Same or more hours in Winter term as in Fall Term Totals 31 59 90 No Change or ‘Decrease in G.P.A. 8 70 78 FallAWinter Totals 39 129 168 7(2 a 13.72** TABLE XVIII The Relationship Between Reduction Reduction of Course Load and G.P.A. For Non-Transfer Students :_: T Less hours in Winter term than in Fall term I T Same or more hours in Winter term as in Fall term Totals Increase in G.P.A. FalléWinter 21 53 74 No Change or Decrease in G.P.A. 6 83 89 Fall-Winter Totals 27 , 136 163 742 13.68" 74 Summary The material in this chapter has been presented in an effort to study the predictive efficiency for students included in this study of the following variables, individually and in combination: orientation test derived scores, high school rank, previously acquired G.P.A.s and Fall term G.P.A. The dependent variable was cumulative G.P.A. for all courses taken during the 1958-59 school year. Analysis of the zero-order correlations for males revealed that previous G.P.A.s and Fall term G.P.A.s correlated higher with the de- pendent variable than did any of the other single independent vari- ables. When selected independent variables were combined into multi- ple regression equations, the weighting attached to previous G.P.A. was greater than for any of the other independent variables included. The Michigan State University Reading test total score, the College Qualification test total score of the American Council on Education Psychological Examination total score and high school rank made only negligible contributions to the multiple correlation coefficients. Analysis of the zero-order correlations for females revealed larger and more significant relationships between variables for the C.J.C. transfer female students. Previous G.P.A.s correlated higher with the dependent variable than did any other single independent vari- able. The data for females must be carefully interpreted because of the small "N" involved. Among other findings in this chapter is the fact that previous G.P.A.s show a higher relationship to achievement in.social studies and natural science than any of the other variables considered. 75 The effect of reducing the students' course load on his academic achievement was also noted in this chapter. It is evident that reduced course load influences the improvement of G.P.A. to some degree. 76 CHAPTER VI SUMMARY, COI‘L’LUSIOI‘S AN) SUGGESTIOI‘S FOR FURTHER RESEARCH The Problem Community-junior college students are transferring to Michigan State University each year in ever-increasing numbers. In the near future these students will constitute a large portion of each entering class. However, it is generally agreed that knowledge of the charac- teristics and needs of this relatively important group of students is limited. Several institutions have undertaken studies of the junior college transfer student in order that they might better comprehend and provide for his needs. In this regard, the present study investi- gates the distinguishing characteristics of the C.J.C. transfer student and the effect of these characteristics on his University work. The study considers the following questions: (1) Are community-junior college transfer students different and therefore distinguishable from non-transfer students? (2) To what extent can existing criteria be used as efficient tools for prediction of academic success for these students after trans- fer to the University? The present investigation is concerned with the differences and 77 similarities between a group of 173 C.J.C. transfer students and a comparable group of 173 non-transfer students at Michigan State Univer- sity. The differences and similarities revealed by this comparison provide the perspective required to identify and explore the areas of need of the C.J.C. transfer student at Michigan State University. In addition, this investigation considers the predictive efficiency of selected educational variables now available for the C.J.C. transfer student after transfer to the University. Methodology This study involves two groups of students: the C.J.C. trans- fer students and the non-transfer students at Michigan State University. The 173 C.J.C. transfer students were selected from the total transfer population according to the criteria mentioned earlier (p 28). The non- transfer students are a random sampling of students who were admitted as freshmen for Fall term, 1956 and were enrolled for Fall term, 1958. Both groups include 137 males and 36 females. Therefore the two groups can be considered comparable in terms of sex ratio and amount of pre- vious college experience. The data were gathered from the permanent record cards, application blanks and record folders in the Registrar's office. The pertinent information was assembled on individual work sheets and punched on IBM cards. Summary data used in statistical calculations were secured by IBM tabulating equipment. Three major objectives were involved in the statistical calcu- lations. The first was to test the differences between the groups; the 78 second, to determine the relationship of selected independent variables with the cumulative G.P.A. for the respective terms of the junior year; and the third, to investigate the predictive efficiency of certain variables, singly and in combination. The analyses of the data could be criticized in certain respects. First, derived scores were utilized on the orientation tests instead of raw scores. However, derived scores were used by the personnel workers who worked with these students and the statistical error has been found to be slight---the r between derived scores and raw scores has been reported by the Michigan State University Board of Examiners to be in excess of .96. Second, the fallibility of teacher grades and resulting G.P.A.s has been shown in the literature. However, the G.P.A. is the only measure available of the student's achievement. Third, the probabilities and frequencies presented are based on group performances and do not necessarily hold for individual prediction. Fourth, the measure of high school rank was reported as a derived score and is at best a rough estimate which can be considered useful in terms of group analysis. Fifth, two assumptions underlying the use of "r", homo- scedasticity and linearity of regression, were not tested, but merely assumed to be operative. Sixth, small N's in some instances affect the interpretation of particular statistical results. Derived scores, such as were used in this study, become most meaningful at the extremes---l, 2, and 3 on the lower end of the scale and 8, 9 and 10 on the upper end of the scale. Zero-order coefficients of correlation were computed by the Michigan State University Integral Computer. (MISTIC) 79 Findings 0n the basis of the analyses presented in Chapters IV and V the following findings are listed: (1) The "t" tables and findings reported in Chapter IV (con- cerning G.P.A.s and high school rank) generally point up the similari- ties rather than any significant differences between the two groups. The most significant difference between the groups was found to be Fall term.G.P.A. for females. The "t" (2.32) in this instance was found to be significant beyond the five per cent level of confidence. The mean Fall G.P.A. for C.J.C. female transfer students was 2.18 as com- pared with 2.63 for female non-transfer students, a difference of .45 of a grade point. The Fall term G.P.A.s for female transfers (2.18) also constitutes a rather large decrease from their previous G.P.A. which was 2.69, a difference of .51 of a grade point. The "t" test did not reveal significant differences on any of the other variables. However, in the case of high school rank, the non-transfer students seem to rank somewhat higher. No significant differences were found between groups with regard to social studies and natural science G.P.A.s. (2) The chi-square analyses point up a number of differences and similarities between the two groups: A. A significantly greater number of C.J.C. transfer students than non-transfer students received cumulative G.P.A.s below 2.0. B. A significantly greater number of C.J.C. transfer stu- dents than non-transfer students in the College of Business and Public Service received G.P.A.s below 2.0. 80 C. No significant differences exist between the ratio of students receiving G.P.A.s below 2.0 in the Colleges of Engineering, Science and Arts, and Education for the two groups. However, C.J.C. transfer students appear to achieve at a slightly higher level in Engineering than the non-transfer students. Non-transfer students appear to achieve at a slightly higher level that C.J.C. transfer students in Science and Arts and Education. D. The drop-out rate is not significantly different between the two groups. It may be noted, however, that a greater percentage of C.J.C. transfer students dropped out than did non-transfer student. E. A significantly greater number of the C.J.C. transfer students than non-transfer students were married. F. A significantly greater number of the C.J.C. transfer students than non-transfer students are veterans. C. There is a highly significant difference between the occupational status of the fathers of the two groups, the non-transfer group generally having higher status than the C.J.C. group. (5) The computation of the zero-order correlation coefficients revealed that G.P.A. earned previously at a C.J.C. or at Michigan State University is the best indicator of how a student will perform in the future. High school rank and orientation test derived scores generally provided low correlations with cumulative G.P.A. and therefore do not appear to have significant predictive value for these students. Howe ever, the orientation test battery derived scores can be useful in 81 working with individuals who have scored at the extremes of the derived score scale. The correlations obtained for females indicated the same general trends as those formales.~ The small sample size of the female groups seriously restricts interpretation of the zero-order correlation coefficients. The zero-order correlation coefficients between social studies G.P.A. and selected variables again indicates that previous grades are the most efficient predictors of future academic achievement. The Michigan State University Reading test derived score also shows some limited predictive value in relationship to social studies G.P.A. In the case of natural science G.P.A. correlations were generally low and therefore of limited predictive value. Previous grades again appear to be the most useful measure. The small N's utilized in the fore- going analysis seriously restrict interpretation. (4) The multiple correlation coefficients (.585, .612) for the male groups, utilizing five variables, (Michigan State University Reading test--total score, College Qualification Test-~total score1 or A.C.E. Psychological Examination--total score2 , previous G.P.A. and Michigan State University cumulative G.P.A.) indicates that the greatest contribution to the coefficient of multiple correlation (R) is made °by previous G.P.A. and that the contribution of the other variables is small. That is, the magnitude of the correlation is not changed greatly by additional variables. The r's between cumulative G.P.A. and 1Community-junior college transfer students. 2Non-transfer students . 82 previous G.P.A. were .53 and .61 respectively. The R's were .58 and .62 respectively. Weighting seems to favor previous G.P.A. as the most efficient predictive measure. All other variables considered here would seem.to have extremely limited predictive value. (5) Chi—square analysis indicated that the reduction of class load, by three or more hours, between the Fall and Winter terms results in increased G.P.A.s for a significant number of students. It would appear that students who reduce their class load have greater success in raising their G.P.A.s than students who maintain the same load or increase the number of term hours. (6) It was further noted that: The male C.J.C. transfer student is, on the average, approximately one year and five months older than his non-transfer counterpart. The female C.J.C. transfer student is, on the average, approximately two months older than her non-transfer counterpart. The two groups carry approximately the same number of hours per term. They also earned approximately the same number of term hours previous to Fall term, 1958, and completed the Spring term, 1959, with approximately the same nmmber of term.hours to their credit and seem to favor the same curricula as reported in Chapter II. Conclusions and Implications for Further Research This investigation has endeavored to study the characteristics of the C.J.C. transfer student at Michigan State University by comparing him with the nonstransfer student. The findings indicate that these two groups of students are quite similar in certain respects and quite differ- 83 ent in other respects. On the basis of these findings, certain con- clusions can be drawn. (1) One of the most significant findings of this study is the fact that female C.J.C. transfer students suffer extreme "grade point average shock" dunng their first term at Michigan State University. Male C.J.C. transfer students suffered very mild "G.P.A. shocks" during the same period. Since the female group presented a superior achievement record and ability level (as measured by the orientation test battery), the conclusion would seem to rest on some non-academic factor such as difficulty in making the "living adjustment" to campus life. Further research into this matter is indicated since greater expectations should be required of the female C.J.C. transfer student. If such research were undertaken, it should probably involve "case study" procedures. (2) Fisher's "t" test indicated that the two groups did not differ significantly in over-all achievement. However, it is impor- tant to note that a significantly greater number of C.J.C. transfer students than non—transfer students failed to maintain 2.0 cumulative G.P.A.s. Similarly, a significantly greater number of C.J.C. transfer students than non-transfer students in the College of Business and Public Service failed to maintain 2.0 cumulative G.P.A.s. It must be remembered that the non-transfer students maintain some superiority in general ability and achievement. However, it would not seem unreasonable to hold higher expectations for the C.J.C. transfer group. More inten- sive study of "failing” C.J.C. transfer students is indicated. What are the factors which contribute to the failure on the part of these students to maintain 2.0 G.P.A.s in selected curricula? The "case study" 84 approach would seem appropriate in this case also. The apparently excessive failure rate of C.J.C. transfer students in the College of Business and Public Service may possibly be the result of higher grading standards in the University Business program than in the C.J.C. Business programs. 0n the other hand, a more serious implication might be basic differences in the instructional programs. Further research is indicated to determine cause in this case. Intensive communications between the staff of the college of Business and Public Service and community-junior college personnel is indicated. (3) While the drop—out rate is not significantly greater for C.J.C. transfer students than for non-transfer students, it amounted to 24 per cent of the C.J.C. transfer group. No attempt has been made in this study to determine the causes of drop-out. Further research is indicated in this area, particularly when the findings of Martorana (37) are considered. It may well be that a significant percentage of the total number of drop-outs result from factors other than academic failure. (4) It may be noted that a significantly greater number of C.J.C. transfer students than non-transfer students were married. It is also noted that a significantly greater number of C.J.C. transfer students than non-transfer students were veterans. The number of veterans in college will decrease rapidly in the next few years, which would render this factor of small consequence for future study. How- ever, the increasing trend toward early marriage dictates that serious thought be given to this factor. Young peOple who marry at the con- clusion of their high school careers are quite likely to utilize the 85 low cost advantages of the community-junior college for the first two years of their college training. Implications for further research are quite evident in this area. (5) The fathers of non-transfer students, as a group, seem to occupy higher occupational and/or socio—economic status. This fact confirms the general feeling that community-junior college attendance is to some degree influenced by economic need on the part of indivi- dual students and their families. This fact, as well as others re- vealed here, indicates that community-junior colleges serve a variety of students with a variety of educational and personal needs. (6) In general, the C.J.C. transfer student seems to compare favorably in terms of the achievement measures utilized in this study. It is apparent that the non-transfer is slightly superior in terms of ability as measured by common scholastic aptitude tests. These findings would tend to indicate the presence of certain selective fac- tors where the C.J.C. transfer student is concerned. That is, the students with poor high school and community-junior college records are not likely to continue on in a four year program. It seems to follow that the C.J.C. transfer group is actually comparable in ability and achievement to the non-transfer students. It must be remembered, howa ever, that there isa-uch greater percentage of higher ability students in the non-transfer group. This investigation has endeavored to study academic prediction by using G.P.A. for all courses taken and for classes belonging to two subject areas (i.e. social studies and natural science). The findings indicate that prediction does not vary greatly between the two groups 86 except in the case of certain variables for the female groups. (1) The findings indicate that one variable stands out above all others as a relatively efficient predictor of academic success. Previous college G.P.A. seems to have a high degree of relationship with subsequent G.P.A. As far as the C.J.C. transfer student is con- cerned, the C.J.C. grade point average is the best single indicator of his expected success in the University program. Other factors such as high school rank, reading test scores, and scholastic aptitude test scores did not contribute greatly to the multiple regression equation. This fact indicates that community-junior college records should probably receive greatest weight in considering probabilities of success where C.J.C. transfer students are concerned. (2) It is apparent that A.C.E. test scores have limited pre- dictive value for junior level non-transfer students. It may be noted that very few'test devices of this type serve well as long range pre- dictors. Consequently, it would not be expected that they would predict well for C.J.C. transfer students. As noted throughout these conclusions, many educational and nunseducational factors may be operating singly or in combination to reduce the predictive value of the test scores for these students. Further research might possibly assist in planning and developing an appropriate orientation testing program.for C.J.C. transfer students. A series of achievement tests in specific subject matter fields might prove more useful to personnel work- ers where the C.J.C. transfer student is concerned. (3) The high verbal content of course work taken by most of the students studied here probably accounts for the relatively high corre- lations between G.P.A. and the Michigan State University Reading test 87 scores. The Reading test score would be useful in working with indi- vidual students. (4) All of the derived test scores can be utilized effectively in selecting out the very low and very high ability students from these groups. (5) The time interval since high school graduation for these students has reduced the usual predictive value of high school rank. (6) Previous grades and Reading Test scores stand out as the only variables which would seem to relate to social studies and natural science G.P.A.s Again, in the case of specific subject matter, the community-junior college record seems to be the best single predictor. In the highly verbal social studies, the Michigan State University Reading Test seems to have considerable relationship to the G.P.A. Further study of these test devices is indicated. (7) All of the correlation coefficients derived for the female samples must be considered carefully since the N's for these groups were very small. No definite conclusions can be drawn from these correlations. It is recommended that further study be made of female C.J.C. transfer students utilizing larger samples. (8) The data for non-transfers compares favorably with the results of studies conducted by the Michigan State University Office of Evaluation Services. (52, 53) The patterns of the correlation analyses are similar but less significant due to the small sample size. (9) Reduction of class load could be an effective means of assisting C.J.C. transfer students to recover from G.P.A. shock. The following specific recommendations for further research and action seem appropriate in view of the findings. 88 (l) A "Case Study“ approach to the academic problems of the IMichigan.C.J.C. transfer student. (At.Michigan State University and possibly in other Michigan colleges and Universities.) (2) A study of drop-outs: Comparison of the characteristics of community-junior college drop-outs with non-transfer drop-outs. (3) A study of the non-academic adjustment problems of the C.J.C. transfer student. (4) Development of test norms for the C.J.C. transfer student. (5) A study of academic prediction (utilizing larger samples than the present study). (6) A comparative study of transfer students from.the various Michigan comunity-junior colleges . (7) A study of the influence of selected orientation practices on the early adjustment of C.J.C. transfer students at Michigan State University. (8) A study of the effect of personal background factors, such as marital status and socio-economic status on the academic achievement record. (9) A comparison between programs of study at Michigan State University and the various Michigan Community-Junior Colleges. (10) A comparison of C.J.C. Transfer students at the various four year colleges and universities in Michigan. 10. 11. 12. BIBLIOGRAPHY Abelson, R. P., "Sex Differences in Predictability of College Grades", Educational and Psychological Measurement, Vol. 12, No. 4: 638-44{TI952. A Study of Florida Junior College Transfer Students in the Florida State University: Fall Semester, 1956-1957, Prepared in the Office of Educational Research and Service, Florida State University, April, 1957. A Study of Students Who Transferred From Florida State University, Prepared in the Office of Educational Research and Service, Florida State University, May, 1955. Barber, Bernard. Social Stratification: A Comparative Analysis of Structure and Procesg New York: Harcourt, Brace and Company, I957. Bohan, J. 8., "Students“Marks in College Courses", PhD. thesis, on file University of Minnesota Library, 1926. Borow, H., "Current Problems in the Prediction of College Performance", Journal of the American Association of Collegiate Registrars, 1946, 22:14-26. Bouge, Jesse P., The Community College, New York: McGraw- Hill Book Company;_l950. Brown,‘H. 8., "Differential Prediction by the A.C.E.", Journal of Educational Research, 44: 116-21, October, 1950. Caplow, Theodore. The Sociology of Work,‘Minnesapolis: University of Minnesota Press, 1954. Clark, Burton R. The Open Door College: A Case Study. (New York: McGrawAHill Book Company, inc.,—I965. Cohan, L., "Predicting Academic Success in an Engineering College and Suggestions for an Objective Evaluation of High School Marks", Journal of Educational Psychology, 37: 381-4, September, 1946. Crawford, Albert B. and Burnham, Paul S., Forecasting College Achievement, New'HavenzYale University Press, 1946. .n ._J ‘L‘ 15. ‘A ks It 23 A, C'- 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 90 Barley, John G. "Factors Associated With College Careers in ‘Minnesota," unpublished manuscript, Center for the Study of Higher Education, Berkeley, California, 1959. Dixon, Wilfrid J. and Massey, Frank J., Iflgrodugtion to Statis- tical Analysis, New York: McGraw-Hill Book Company, Inc., 1957. Douglass, Harl R., "How Can A Junior College Best Serve the Needs of a Student Going on to Senior College", North Central Association Quarterly, 28: 404-410, April, 1954. Durflinger, G.W., ”The Prediction of College Success--A Summary of Recent Findings", Journal of the American Association of Collegiate Registrars, 19: 68-78, 1943. Bells, Walter C., The Junior College, Boston: Houghton Mifflin Co., 1931. Ezekiel, Mordecai and Fox, Karl A., Methods of Correlation and Regresion Analysis, New York: John Wiley and Sons, Inc., 1959. Farwell, Gail P., "An.Analysis of Factors and Criteria Related to the Admission of Borderline Cases at Michigan State College, Fall Quarter, 1952.", (Unpublished Ed.D. Dissertation,‘Michigan State College, East Lansing, 1954.) Feder, D. D., "Factors Which Affect Achievement and Its Prediction at the College Level", Journal of the American Association of Collegiate Registrars, 15:107-118, 1940. Florida Junior College Transfer Students in the Florida State University: Fall Semester, 1958-1959, Prepared in the Office of Institutional Research and Service, Florida State Univeristy, April, 1959. French, J. W. and Others, "A Factor Analysis of Aptitude and Achievement Entrance Tests and Course Grades at the United States Coast Guard Academy, The Journal of Educational Psychology, 43: 65-80, February, 1952. Froehlich, G. J., ”Academic Prediction at the University of Wisconsin", Journal of the American Association of Collegiate Registrars, 17: 65-76, 1941. Garrett, Harley F., "A Review and Interpretation of Investigations of Factors Related to Scholastic Success in Colleges of Arts and Sciences and Teachers Colleges", (Unpublished Doctor's Dissertation, The University of Colorado, Boulder, 1949.) Garrett, Henry 8., Statistics in Psychology and Education, New York: Longmans, Green and Company, 1955. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 91 Grossman,‘D. A., "Junior College Transfers at Illinois", Junior College Journal, 4: 297-303, March, 1934. Guilford, J. P., Fundamental Statistics in Psychology and_Educa- tion, New York: McGraw-Hill Book Company, Inc., 1956. Havinghurst, Robert J. and Neugarten. Society and Education, Englewood Cliffs, NuJ.: Allyn and Bacon, Inc., 1957. Hertel, J. P. and DiVesta, F. J., "Evaluation of Five Factors for Predicting the Success of Students Entering the New York State College of Agriculture", Educational and Psychological Measurement, Vol. 8, No. 3: 389-95, 1948. Hurd, A. W., "Problem of the Prediction of College Success", Journal of Educational Research, 38: 217-19, November, 1944. Johnson, Palmer 0., Statistical Methods in Research, New York: Prentice-Hall, Inc.,'1950. Johnson, Walter P., "A Study of the Efficiency of Certain Factors for Predicting Achievement of Veterans at the Junior College Level in the College of Science, Literature and the Arts at the University of Minnesota.“ (Unpublished PhD. Dissertation, University of Minnesota, Minneapolis, 1950.) Johnston, J. B., "Predicting Success or Failure in College at the Time of Entrance,” School and Society, 1926, 23: 82-88. Kirk, Barbara A., "Comparison of Transfer Students by Source of Origin with Entering Students on the College Qualification Test", Junior College Journal, 29: 218-221, December, 1958. LaFauci, H.‘M., and Richter, P.E., "Academic Success Beyond Junior College: The Identification and Selection of the Four Year Student", Junior College Journal, 29:123-127, November, 1958. iMartorana, S. V., ”The Community College in.Michigan", The Survey of Higher Education in Michigan, Prepared for the Michigan Legislative Study Committee on Higher Education, December, 1956. ‘Martorana, S. V. and Williams, L. L., ”Academic Success of Junior College Transfers at the State College of Washington", Junior College Journal, 24: 402-415, March, 1954. Medsker, Leland. The Junior College: Progress and Prospect, lbw York: McGraw-Hill Book Co. Inc., 1960. Odell, C. W.. Predicting College Success of College Students, University of Illinois College of Education, Bureau of Educational Research, Bulletin No. 52, Urbana, 1930. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 92 Peterson, B. H., "Admission Practices in Relation to Junior College Education”, California Journal of Secondary Education, 28: 40-42, January, 1953. Reeves, Floyd‘W. and Russell, John D., Admission and Retention of University Students, Chicago: The University of Chicago Press, 1933. Remmers, H. H. and Others, "Curricular Differences in Predicting Scholastic Achievement: Applications to Counseling", Journal of Educational Psychology, 40: 385-94, November, 1949. Rodes, H. P., "Succesful Transfer in Engineering", Junior College Journal, 20: 121-127, NOvember, 1949. Seashore, Harold, "Academic Abilities of Junior College Students", Junior College Journal, 29: 74-80, October, 1958. Segal, D., Prediction of Success in College, United States Office of Education Bulletin, 1934, No. 15, Government printing Office, Washington, D. C., 1934. Siemens, Cornelius H., "Predicting Success of Transfer Students", Junior College Journal, 14: 24-28, September, 1943. Smith, Francis F., "Use of Previous Record in Estimating College Success", Journal of Educational Psychology, 36: 167-76, March, 1945. Starrak, J. A. and Hughes, R. M., The Community College in the United States, Ames, Iowa: Iowa State College Press, 1954. Stright, I. L., "Some Factors Affecting College Success", Journal of Educational Psychology, 38: 232-40, April, 1947. Success of 832 Community-Junior College Transfer Students at Michigan State University: Winter Quarter, “1959, Prepared in the Office of Community-Junior College Cooperation,‘Michigan State University, 1959. The Committee on Junior and Senior Colleges of the Association of American Colleges, The American Association of Junior Colleges and the American.Association of Collegiate Registrars and Admissions Officers, "A Report of Progress of the Committee on Junior and Senior Colleges", Junior College Newsletter, 15: 2-3, February, 1960. 52. 53. 54. 55. 56. 93 The Orientation Tests and Long-Range Predictions: A One Year Follow-up of New Freshmen, Part 1, Prepared in the Office of Evaluation Services, Michigan State University, 1959. The Transfer Student at Michigan State University: His Attain- ments and_Abilities, Prepared in the Office of Evaluation Services, Michigan State University, 1959. Votow, D. F., "Comparison of Test Scores of Entering College Freshmen as Instruments for Predicting Subsequent Scholarship", Journal of Educational Research, 40: 215-18, November, 1946. \ Williamson, E. G., "The Decreasing Accuracy of Scholastic Predictions", Journal of Educational Psychology, 28: 1-16, 1937. . , Wise, Max W. They Come for the Best Reasons - College Students Today, Washington: American Council on Education, 1958. APPENDIX A 94 PROCEDURES FOR PREPARING DATA FOR MACHINE ANALYSIS Data from the indxv1dual records and transcripts were trans- ferred to a "Data Sheet" prepared for each student and recorded in black pencil. Code numbers were then recorded in red pencil. The numbers recorded in red pencil were then transferred to IBM master worksheets. The numbers from the master worksheets were then punched directly on IBM cards by the key punch operator. The coded figures vere used for all subsequent analyses performed. A sample "Data Sheet" is found on the following page. Following the sample data sheet, the coding procedure is described in greater detail. A sample punched card completes Appendix A. DATA SHEET 95 He oe an mm nm on mm em mm wm an whomumm who szwo coo uo wommsz huwcsaaoo ono om mm mm .Imw mm mm mm mm mm Hm ha O H O O 11. 12. 13. 98 Unknown or other 22 Busines & Pub. Service Engineering Education (Ele. & Phy.) Science & Arts-L (Ed) Science & Arts-S & C (Ed.) Science & Arts-S & C Science & Arts-L Communications Arts Home Economics Agriculture Agriculture (Ed) 23* in school 21 Fall Winter Spring 16 17 18 19 2 33 40 20* Alpena 32 Battle Creek Bay City Benton'Harbor Flint Grand Rapids Gogebic Henry Ford Highland Park Jackson Muskegon Northwestern Port‘Huron 17 18 Reduced Class Load Winter term Increased GPA 99 1. Reduced 16-17 2. Same or More 1. Increased 12-13 2. Same or Less Columns 1 - 6 7 8 9 - 10 11 - 12 13 - 14 15 - 16 17 - 18 19 - 20 21 - 22 23 - 25 26 - 28 29 - 31 32 - 34 35 - 37 38 - 40 41 - 43 44 - 46 47 - 49 50 51 NON-TRANSFER STUDENTS Variable Student Number Group Designation Control 1. Complete data 2. Incomplete data ACE-Q ACE-L ACEJT Reading-V Reading-C Reading-T High School Rank MSU GPA up to Fall '58 MSU Fall '58 GPA MSU Winter '59 GPA MSU Spring ' 59 GPA MSU Social Studies GPA MSU Natural Science GPA MSU Mathematics GPA MSU Humanities GPA MSU Cumulative GPA Sex Marital Status Code 100 Place (Block No) 2 1. MCJC Transfer 41 2. Non-Transfer (derived scores 1-10) n n n u 7 O i 1 0" 0’. 6 10 11 (15 point scale) 39 places) 1. Male (GPA's computed to 2 decimal 2 . Female 1. Single 2. Married 12 13 14 24 25* 26* 27 28 38 101 52 Military Status 1. NOn-veteran 5 2. Veteran 53 - 54 Age to the nearest year as 4 of October 1, 1958 55 Fathers Occupation 1. Executive 36 2. Professional 3. Proprietors 4. Semi-Prof. 5. White Collar 6. Skilled Labor 7. Semi-Skilled 8. Unskilled 0. Unknown 56 Vocational Goal 1. Executive 37* 2. Professional 3. Proprietors 4. Sam-Prof. 5. White Collar 6. Skilled Labor 7. Semi-Skilled 8. Unskilled 0. Unknown 57 - 58 Curriculum 0. Unknown or other 22 1. Business & Pub. Service 2. Engineering 3. Education (Ele. & Phy.) 4. Science & Arts-L (Ed) 5. Science & Arts-S & C (Ed) 6. Science & Arts-S & C 7. Science & Arts-L 8. Communications Arts 9. Home Economics 10. Agriculture 11. Agriculture (Ed) 59 Curriculum Changes 23* 60 Drop-out 21 61 - 62 Hours Carried Fall '58 16 63 - 64 Hours Carried Winter ' 59 17 65 - 66 Hours Carried Spring '59 18 67 - 68 Hours Carried Total (FWS) 19 69 - 71 72 - 74 75 - 76 77 78 Total hours before Fall '58 Total College Credits Basic College Course Credits Reduced Class Load Fall to Winter Increased GPA Fall to Winter 102 33 4O 20* 16-17 12-13 103 I BM CARD 1:" C"? N DGO—DNuM."¢omDLDOF~wL-Ocmo nan—V‘Nnnnvmmfitflnf‘ 0 1‘" eezzsc.zczzez083383iamamsaxnwzfieszsosss:éaezaasazusuazesznafiac2:22:22:2 mmmmmcmommammmmmmmmm Bammmmmmmmwmmmmmmmmmmammmmmmm.mmmmmmmmmmwmmmmwnmmmfi _ . “Bani..aw.:N.:.::ms8.as.w3..s .L3 mn3N1..nmmnA:.nsm.3..sm.:a.e.vza-s.nsnaxnam.Nsmcm.NewN:.Nu Lam.e..2n.:..:.;2 m mmmmmmxn6.mmmnm $5» m mammamwmmmxmwm:mmmmmmmam6mmonwwmmmwm wmemmmm 3 2 a. m— 3 21.22 :LuNma3533333 3L3 mama .w 4... mm 3 Mn N... 233 $3 33.3.3"va 3.3 9, 3 NM 3 £32.. 2.. :27. mag 2 £22 an... NNNNNNNNNENNNNNNNNNLNNNNNNNNNNKNNNNNNNNNKNNNNHNNNNSNNNNNNNNrNNNNANN 3 L32 :fifl :22 IANSNBSSQENAS3.30.“.3333323"13:33:93...va3933:3333 fibufluwZfiRN433....«2 m. t 231 Emmmmmmmmmammmmmmmomwmmmmmmmmmdmmwmnmmmmammawumm mEmmwmmmmmmmmmwmmm eNienaaN.:.zema3.emaasnw3.aamn3..z “name.mgsvs..3n.:n.:.JQNNSNNfifinsz .mamNfiNNasquNm.:fim.s..2m.: mmmmmmmwmmm mm mnmmmmmhmmmmmmm :m m:mnmn nnmammmnnmm mammmmmmmmmfimmmmmm 822:3.3122 TYSQZJQJSBNLgnwmmN-onnnvnnmwn6.3333333“.va:L3mqwnfimnnmxnafiNaanmwm-~p~a~vummw~fiLRmL3:39! 3 L évqucvvvtvvvowevvvwevvvvevveeéovvvvvvvqsvvvvvvv v34v..cvv¢efivvv¢¢ev Lommifiu~2m~21221u$153332;3333333333..1933:3933:.vL....mmamNnR.nnvmnm~m_nLonar wmmnnvmnnn~3m~22:22! E: NNNNNNNN NN N:NNNN:NNNNNN NNNNNNNANNANNNNNNNNN N:N N:NNNN2NNNNNNNMN1 L 'flN C7) 44W ca— 10 u CC N G‘- ‘a C') G '39 n (T) :1 ’N g. ‘ “as 7D '1." LI? mm... (‘4 n O (v m v r. .6 m r .1) r N :3 cr~ "N H f" "I‘m 2: to 3;; 'T’ (I) to J. '-.a :: u 1 __' '- CD R IN . 1:.) I“) Ln :fi 115' LD 10 1n v-a ._ '1' it w w]. 41 v H M {eetzcxzvfnascmssrsas:nzsaxnwéjrscssx:3??:£:aaxnzmécmzanaaefiaee2222::2 NNNNNNNNNNNNNNNN NNNN NNNNNNNNN NNNNNNNNNNNNNNNNNNN NKNNNNNNNNTNNNNNNNNNNN {eczeczzezéssmssmaezia.sasxtm2339:3933?:Ssaasaxmmzimezcnzzsagee:2czce:5 :_~———————L—————__—_LL._p—_—_.—:—_———_—_—p:—_——__— __—_————p—L_———————_L_ Ere222:2::.283$883323§223$3322:Ssszvezficsyaaasaxaa: razecznaakrezeczczze Ema:aooaacaeococcocaacocaoccaouacaaecooaaoaaa:aooazcacaaaacmocooaoocooa L822 :3»... I 22 2233583333 SLSmnnmSwnmnxnan SLcnmvwomvwvnvagnv 33.835331”?~n$=fifl§§w~m~wnn~fiNnLoNE 2 :32! 2n— :2 N N—M— L) . .aa—aCQmm,¢¢LnO(Omr~ama-cnm -GCFCNmmuW-m~¢D-f~w hafiw-NCNNMN‘T'B QOv—vmvm.v.invwvhvm NDN—NNNMN‘QNWN‘DNHN nfin—flNq-nmnfl' -°"e—— 4...:1i. A.-. 11". I1. 5:? @1135. ... 3 9 3.32;. 3.1:: s .2 3.3.2.1.”. 2. a. ..N .-._.N are .NNN.JNN_N....N__2 2.....1ca..:_.:N..._.93.4.13. A0: ‘L‘ .>.. k r.rl. .Ll . . eflwsgnfafl APPENDIX B COURSES USED FOR COMPUTATION OF GRADE POINT AVERAGES IN SOCIAL STUDIES AND NATURAL SCIENCE I. Social Studies Anthropology Economics Geography History Political Science Sociology 11. Natural Science Anatomy Astronomy Biochemistry Biology Botony Chemistry Geology Physics Physiology Zoology 104 APPENDIX C 105 1. Classification of Curriculum 0 Unknown or other 1 Business and Public Service 2 Engineering 3 Education (Elementary and Physical) 4 Science and Arts - Linquistic (Education) 5 Science and Arts — Scientific (Education) 6 Science and Arts - Linquistic 7 Science and Arts - Scientific 8 Communications Arts 9 Home Economics 10 Agriculture 11 Agriculture (Education) II. Classification of Father's Occupational Status 1 Executive (Big Business) 2 Professional 3 Proprietors (Managers, Farmers, Small Business) 4 Semi—Professional (Technicians) 5 White Collar (Clerks, Sales) 6 Skilled Labor (Foreman) 7 Semi-Skilled Labor 8 Unskilled Labor 0 Unknown - deceased - retired APPENDIX D STANDARD ERROR OF THE ZERO-ORDER CORRELATION COEFFICIENTS Tables XIII-A and XIII-B .70 - .81 .04 .50 - .69 .06 .30 - .49 .08 .10 - .29 .09 Tables XIV-A and XIV-B .70 - .89 .09 .50 - .69 .12 .30 - .49 .15 .10 - .29 .17 Tables XV-A, XV-B. XVI-A and XVI-B .70 - .89 .06 .50 - .69 .08 .30 - .49 .10 010 - .29 .12 106 APPENDIX E 107 FREQUENCY DATA RELATED TO THE EFFECT OF SEX, MARITAL STATUS* AND MILITARY STATUS* ON SUCCESS-FAILURE CJC Transfer Non-Transfer Students Students Above Below Above Below 2.0 GPA 2.0 GPA Total 2.0 GPA 2.0 GPA Total Male 103 34 137 117 20 137 Female 26 10 36 33 3 36 Total 129 44 173 150 23 173 Married 23 6 29 6 1 7 Single 80 28 108 112 18 130 Total 103 34 137 118 19 137 Veteran 36 10 46 18 o 18 Noanet. 66 25 91 100 19 119 Total 102 35 137 118 19 137 *Male only APPENDIX F W WORM! Library Routine {SE-M TITLE: Product Manent Correlation, Means, Standard Deviation, Variances and Covariances, Card Input. TYPE: Entire Program MATIOII: Input: ioo Cards/minute maxim Computation: 53.3n2 + 60.2n Milliseconds Output: 25 p1n(n+l) milliseconds - for correlation mtrix 25(l+ )(5+n)n milliseconds - for mean, standard deviat on and variance - covariance matrices. where s . sample size n . number of variables p1- number of characters with which each correlation coefficient is punched. ' number of characters with which each mean, standard deviation, variance and covariance is to be punched. mums: 272 man or m: The program is read into the memory in the usual way followed by the parameter tape and lastly the data cards. Sane canputing is done after each row of the measurement matrix has been read into the memory. Since the correlation and variance-covariance matrices are symmetric, it is necessary to print only half the off-diagonal elements. 'lhe lover off-diagonal and diagonal elements are printed out row by row (this is equivalent, however, to printing out the upper aff- diagonal and diagonal elements column by column). First the correlation matrix is punched out, scaled down by a factor of ten, followed by an N. Next the mean and standard deviations appear in two parallel columns. finally, the variance-covariance matrix is punched out. A new problem can be begun by reading in new parameters. CAPACITY: Thirty-four variables: there is no limit on the amber of observations. W 01" m TAPES: For every problem four parameters are necessary. ‘lhey are as follows: 1. Let "s" be the sample size. Put as on the parameter tape. 2. Let ”n" be the number of variables. Put all on the parameter tape. KS-M CARDFORMAT: WON CARD: March 6, 1959 3. let "f" be the number of decimal places to which the correlation matrix is to be printed. Put fl" on the para- meter tape. If no print out is desired, f a O. h. let "I" be the number of decimal places to which the means, standard deviations and variance-covariance matrices are to be printed. Put IL on the parameter tape. If no print out is desired, 1 - 0. me eighty column card is to be punched so that at most seventy- two columns contain data. The remaining eight columns are not read by this routine, and. will cannonly be used for identifiers, etc. me eight columns can be any eight columns, and need not be continuous. However, these eight columns must be plugboard wired so that four of them go to aoa1a2a3 of A and four more to qoqhqg1 of A. Test column which is wired to read into so called column 0. 'lhere is a standard plug board already wired which puts colmnns l-h into ao-a3, 5-8 into qo-q3, and 941+ into au-a39, 155-80 into (ht-9393 The remaining 72 columns contain all fractions (-l