A STUDY OF THE EFFECT OF PERSONAL AND NONaPERSONAL CONTACTS. BY RELIGIOUS WORKERS ON THE RELIGIOUS INVOLVEMENT OF LUTHERAN COLLEGE FRESHMEN mesis fer the Degree 0! Ph. I]. MlfiHIGAN SIATE UNIVERSITY Donald Warren Herb 1966 TH £95 This is to certify that the thesis entitled A STUDY OF THE EFFECT OF PERSONAL AND NON-PERSONAL CONTACTS BY RELIGIOUS WORKERS ON THE RELIGIOUS INVOLVEMENT OF LUTHERAN COLLEGE FRESHMEN presented by Donald Warren Herb has been accepted towards fulfillment of t , requirements for Phil}, . degree in Counseling and Guidance / 1'// fix 2 CLO/44¢ Majcyfi'rofessor 5' Date June 37. 1966 0-169 3 IIIII 1 II III III IIII IIII IIII III III IIII IIII III IIIII IIIII 293 .In. Ii i .Alynrgmnmfin..£ sh JI ABSTRACT A STUDY OF THE EFFECT OF PERSONAL AND NON-PERSONAL CONTACTS BY RELIGIOUS WORKERS ON THE RELIGIOUS INVOLVEMENT OF LUTHERAN COLLEGE FRESHMEN by Donald warren Herb Objectives This study is an experimental investigation of the effect on the religious involvement of college students of various fbrms of contact with campus religious workers, comparing the impact of personal face-to- face contacts, a series of contacts by regular personalized mail, and a contact by a printed publicity brochure. An attempt was made to answer the following questions: 1. Hill personal contact between religious workers and students produce greater religious involvement in the students than personalized mail? 2. Will personal contact and personalized mail produce greater religious involvement in students than impersonal mail? 3. Hill different types of contact produce different effects in male and female students? 4. Hill different types of contact produce different effects when used by different workers? The Sample The sample consisted of 180 freshman at Michigan State University in East Lansing who indicated a religious preference for the Lutheran Church on the IBM cards voluntarily completed by students and collected DONALD WARREN HERB by the University at registration. The subjects were classified into categories for sex, size of home town, campus residence, and church membership and randomly divided into six groups of 30 each. No married person nor anyone over 18 years of age was included in the sample. Methodolo The six groups were assigned as follows: two groups were allocated to the campus pastor, two groups were apportioned to the lay campus worker, and two groups were given to neither worker. Each worker contacted the subjects in one assigned group by a series of personal con- tacts consisting of a visit to each student in his residence hall, an invitation to visit the worker in his office at the local church, and an invitation for hOSpitality at the worker's home. Both workers contacted the subjects in their other assigned group with a series of personalized messages by mail. The local church sent a printed publicity brochure to the subjects in the remaining two groups. These treatments were admin- istered over a period of approximately four months. Data were collected by means of registration cards at 17 services at the local church and a questionnaire was sent by mail to the 161 subjects who remained at the university during the sixth month. Follows up procedures extended into the seventh month. The questionnaire was designed to measure religious involvement separately for the action, attitude, participation, and belief variables. The analysis employed the statistical procedures of analysis of variance, chi square, a means test using chi square, and Goodman-- Kruskal correlations as appropriate. DONALD WARREN HERB Findings Personal contact and personalized mail produced statistically significant increases in religious involvement when compared to impersonal mail for the action (operationally defined as attend- ing church Services, communions, and religious programs, contri- buting money, initiating contacts with religious workers, and visiting student religious centers) and attitude variables. Personal contact produced statistically significant increases in religious involvement when compared to personalized mail for the action and attitude variables. No statistically significant treatment effects were found for the participation and belief variables. Males and females responded to the treatments in a significantly different manner for the action variable only. There was no significant difference in the effect of the treat- ments when used by different workers. A STUDY OF THE EFFECT OF PERSONAL AND NON-PERSONAL CONTACTS BY RELIGIOUS WORKERS ON THE RELIGIOUS INVOLVEMENT OF LUTHERAN COLLEGE FRESHMEN By Donald warren Herb A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services, and Educational Psychology College of Education 1966 ACKNOWLEDGMENTS The writer wishes to express his sincere appreciation to Dr. John Jordan, who acted as advisor for this research, and to Dr. William Farguhar, Dr. Walter Johnson, and Dr. Alfred Dietze, members of the committee. He is deeply indebted to his colleague, Miss Tecla Sund, who gave unselfishly of her time and effOrt as the lay campus worker in this study. He is grateful to the staff and members of University Lutheran Church, East Lansing, Michigan, the local church involved in this research. Many people have helped in ways too numerous to mention, but special thanks is hereby expressed to Dr. Dana Quads of the University of North Carolina, who acted as a consultant on statistical procedures and to the Reverend Otto Bremer, who as a member of the national staff of the Division of College and University Whrk of the National Lutheran Council persistently stressed the value of calling in a campus ministry and inspired this investigation. The work of the typists is also appreciated: Miss Sandra Herb, who typed the early drafts and Mrs. Erenda Jarman, whose skill and craftsmanship is displayed here. ii VITA Donald warren Herb Candidate for the degree of Doctor of Philosophy Date of EMamination: June 3, 1966 Dissertation: A Study of the Effect of Personal and Non-Personal Contacts by Religious WOrkers on the Religious Involvement of Lutheran College Freshmen Outline of Studies: Major area - Counseling and Personnel Services Minor area - Psychology Biographical Items: Birthdate - March 23, 1923, Reading, Pennsylvania Undergraduate Studies: Gettysburg College Gettysburg, Pennsylvania Graduate Studies: Lutheran Theological Seminary Gettysburg, Pennsylvania - 1943-1944 Lutheran Theological Seminary Philadelphia, Pennsylvania - B.D., 1944-1945 Columbia University New York City - 1951-1954 Michigan State University East Lansing, Michigan - 1959-1966 Experience: United States Navy Chaplain, USNR - 1945—1947 St. Matthew Lutheran Church, Lancaster, Pennsylvania Pastor - 1947-1950 Lutheran Student Foundation of Greater New York New York City Lutheran Pastor to Students - 1950-1957 University Lutheran Church, East Lansing, Michigan Pastor and Campus Pastor - 1957-1962 Division of College and University Work, National Lutheran Council, Chapel Hill, North Carolina Regional Secretary - 1962-1966 Lutheran Theological Seminary, Gettysburg, Pennsylvania Associate Professor - beginning July 1, 1966 Membership held in the North Carolina Synod of the Lutheran Church in America. iii Dedicated to My wife, Nancy, who sacrificed whatever was necessary to bring this study to completion, my children, Sandra, Paul, and Peter, who shared in the deprivations and hopes, and to my parents, because there are so few ways in which to repay the investment of their lives in mine. iv TABLE OF CONTENTS VITA . . ....................... . ..... DEDICATION . . . . . . . . ................. LIST OF TABLES ......... . .......... . . . LIST OF FIGURES . . . . . . . . . ..... . . . . . . . . LIST OF APPENDICES . . . . . . ...... . . . ...... Chapter I. THE PROBLEM . . . . ....... . . . . . . Purpose . . . . . . . . . ..... . . . . . . . . Definitions . . . . . . . ....... . . . . . . . Limitations of this Study . ............... Basic Assumptions . . . . . . . . . . . . . . . . Research Hypotheses . . . . . . . . . . . . . . . . The Need for the Study ......... . ....... The Importance of this Study ...... . . . . . . Organization of this Thesis . . . . . . . . . . . . II. A REVIEW OF RELATED LITERATURE . . . . . . . . Attitudinal Changes of College Students . . . . . Motivation to Participate . . . . . . . . Contact . . . . . . . . ....... . . . Theoretical Considerations Regarding Personal Contact . . . . . . . . ......... . . Research on Personal Contact . . . . . . . . . . . . . . Personal Contact by Church workers . . ......... Contact by Mail . . . . . . . . . . . . . . . . . . Mbasurement of Religious Involvement . . . . . . . . Summary . . . . . . . . . ................ III. RESEARCH DESIGN AND METHODOLOGY . . . . . . . . . . . The Setting . . . . . ...... . . . . The Sample . . ...... . . ..... . . . . Grouping . . . ....... . . ...... . The werkers . . . .......... . . . . . . . . (I) O‘UIUOWWNHH H Chapter Page The Instrument . . . . . . ..... . . ....... . 38 Experimental Treatment . . . . . . . . . . . . . . . . . 39 The Gathering of the Data . . . . . . . . . . . . . . . . 43 Statistical Procedures . . . . . . . . . . . . ..... 45 Summary . . . . . . . . . . . . . . . ....... . . . 47 IV. RESULTS OF THE STUDY . . . . ....... . ....... 48 Sample Size after Treatment . . . . . ........ . . 48 Action Data . . . . ...... . . . . ...... . 49 Attendance at University Lutheran Church ....... . 50 Other Action Data . . . . . . . . . . . . . . ..... . 55 Attitude Data . . . . . . . . . . . . . . . . . . . 59 Participation and Belief Data . . . . . . . . . . . . . . 66 Summary . . . . . . . . . . . ...... . . . . . . . . 68 V. CONCLUSICNS, DISCUSSIONS, RECOMMENDATIONS, IMPLICATIONS FOR FUTURE RESEARCH AND SUMMARY . . . . . . . . . . . . . 70 Conclusions . . . . . . . . . . . . . . . . . ..... . 70 Discussion . . . . . . . . . . . . . . ..... . . 70 Recommendations for Campus Ministry Personnel . . . . . . 77 Implications for Future Research . . . . . . . . . . . . 78 Summary . . . . . . . . . . . . . . ...... . . . . . 79 REFERENCES . . ...... . ..... . . . . . . . . . . . . . . 81 APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 vi Table 10. ll. 12. 13. LIST OF TABLES Assignment of groups by type of treatment and workers 0 O O 0 0 O O O O O O O O O O O 0 Composition of each treatment group of 30 individuals by sex, church membership, type of home town, and area of campus residence . . . . . . . . . . . . . . . Statistical design for analysis of variance computations by treatment groups, sex, and workers . . . . . . . . . . Size of sample and groups at end of treatment and number and percentage of questionnaires returned . . . . . . Types of reaponses given by telephone by subjects who did not return questionnaire . . . . . . . . . . . . Analysis of variance for attendance at University Lutheran Church as determined from registration cards by treatment groups, sex, and workers . . . . . . . . . . . . Means of treatment groups for attendance at University Lutheran Church as determined by registration cards by sex and workers . . . . . . . . . . . . . . . . . . . . . Analysis of variance for attendance at University Lutheran Church as determined by the questionnaire by treatment groups, sex, and workers . . . . . . . Means of treatment groups for at tendance at University Lutheran Church as determined by the questionnaire by sex and workers . . . . . n . . . . . . . . . . . . Analysis of variance for attendance at students' home church by treatment groups, sex, and workers . . . . '.' ‘Means of treatment groups for attendance at students' home church by sex and workers . . . . . . . . . . . . . Analysis of variance for total church attendance by treatment groups, sex, and workers . . . . . . . . . Means of treatment groups for total church attendance by sex and workers . . . . . . . . . . . . . . . . . . . . vii Page 36 37 45 49 49 5O 51 54 55 55 56 56 56 Table Page 14. Analysis of variance for number of times communion was received by treatment groups, sex, and workers . . . . . . 57 15. Means of treatment groups for number of times communion was received by sex and workers . . . . . . . . . . . . . 57 16. Joint frequency of contacts initiated toward lay worker by assigned treatment groups . . . . . . . . . . . . . . . 58 17. Joint frequency of visits to a local church or student center by treatment groups . . . . . . . . . . . . . . . . 58 18. Goodman--Kruskal correlation for responses to the question about the extent of the conflict between science and religion . . . . . . . . . . . . . . . . . . . 59 19. Goodman--Kruskal correlation for responses to the question asking what difference coming to believe there were no God would make in various opinions . . . . . 6O 20. Joint frequency of evaluation of lay campus worker by assigned treatment groups . . . . . . . . . . . . . . 61 21. Median test: joint frequency fer person preferred as a helper in a serious emergency by assigned treatment groups 0 O O O O O I O O O O O O O I O O O O O O O O O O O 62 22. Analysis of variance for scores on the Puritan Morality-- Relativistic scale by treatment groups, sex, and workers 0 I O O O O I O O O O O O O O O O O O O O O O O O 63 23. Means of treatment groups for scores on the Puritan Morality--RelatiVISth 308.16 0 e e o e o e o eeeeee o 63 24. Joint frequency of types of essay responses about the adequacy of the campus ministry and fulfilling students' religious needs by treatment groups . . . . . . . . . . . 64 25. Joint frequency of reaponses to the question about the degree of religious influence in students' upbringing by treatment groups 0 O I O O O l O O I I O O O O O O O 65 26. Joint frequency of responses to the question about maintaining religious workers by treatment groups . . . . 66 27. Means of books and articles read on religious themes by treatment groups . . . . . . . . . . . . . . . . . . . 67 28. Joint frequency of reading books and articles on religious themes by treatment groups . . . . . . . . . . . . . . . . 67 viii Table 29. Joint frequency of responses to question about the Nature of Christ by treatment groups . . . . . . . . . . . . . . 30. Analysis of variance for religious beliefs test by treatment groups, sex, and workers . . . . . . . . . 31. Means of treatment groups for religious beliefs test . . . ix Page 67 LIST OF FIGURES Figure Page 1. Sunday morning church attendance by males at University Lutheran Church as recorded on registration cards . . . . . 52 2. Sunday morning church attendance by females at University Lutheran Church as recorded on registration cuda O O O O O O ...... O ....... O O O O I O I 53 Appendix A. The Instrument and Registration Card . . . . . . . B. Letters Used in Requesting the Questionnaire Return and FOllOW-Up o o o o o eeeee e e e e e o e e o C. Types of Letters Used With Personalized Mail Groups 0 O O O O O O O O O O O O O O O O O O D. The Brochure . . . . . . . . . . . . . . . . . . . . . . LIST OF APPENDICES Page . 88 . . . 94 . 98 104 CHAPTER I THE PROBLEM Purpose There are a limited number of ways in which campus religious workers may contact their constituents for religious purposes. The traditional methods involve personal face-to-face contact. Also, a great deal of material is sent through the mail. Some mail seeks to convey a message from a specific worker to a specific student. Other mailed material is mass-produced and,however cleverly designed, is readily seen by the receiving student to be non-personal. The purpose of this study is to investigate the effect of these various forms of contact on the religious involvement of students. Definitions For the purpose of this study, the following terms are operationally defined: Religious involvement - a measurable behavior or cognition toward the belief system.perceived to be associated with, the institution perceived to be associated with, the persons perceived to be related to and/or the ethics and values perceived to be associated with a traditional Christian concept of God. This has been called religiosity in many studies. The measurable behaviors and cognitions are defined as: Action (number of Services attended, amount and frequency of= contributions, number of communions, number of religious programs attended, number of contacts initiated by a student toward a religious worker, visits to student religious=centers), Attitudes (toward the clergy, campus religious workers, the Church and the churches, theology, church activities, ethical questions), Participation (think- ing about religioum and ethical problems, reading religious literature and the Bible, efforts to evangelize others), and Beliefs (about God and related matters, the Protestant religion). 1 Personal contact - a face-to—face confrontation initiated by a campus ' religious worker with a student. Calling - a visit by a worker to a student at his home or residence hall. Office interview - a visit by a student to the worker's church office upon the invitation of the worker. Home hospitality - a visit to the campus religious worker's home for food and conversation at the worker's invitation. Personalized mail - a written or printed communication addressed to a student by name from a campus worker who identifies himself by name. Impersonal mail - a printed general communication sent to a student without any personalized reference to the name of the sender or recipient except as part of the formal printed message in the former instance and as part of the mailing address in the latter. Campus ministry — a program whereby churches seek to promote their aims at colleges and universities and to serve their constituents among the faculty and student body. Campps pastor - an ordained minister, usually male, who is paid a salary for work performed in the campus ministry. Lgy campus worker - an unordained person, usually female, who is paid a salary for work performed in the campus ministry. Limitations of this Study The population from which experimental group samples were drawn was composed of all freshmen entering Michigan State University in a selected year who indicated a religious preference for a Lutheran Church related to the National Lutheran Council. Only limited generalizations should be made to other populations, such as freshman in other years, older students at or above the sophomore level, other colleges and univer- sities, and students from other denominations. Tentative conclusions drawn from.this study about other populations should be tested by further experimentation. Basic Assumptions Two basic assumptions are made: 1. Religious involvement can be changed or modified. 2. Religious involvement can be measured by measuring closely related cognitions and behaviors as operationally defined else- where in this chapter. Resegych Hypotheses It is hypothesized that: 1. Personal contact and personalized mail will produce greater religious involvement than impersonal mail. 2. Personal contact will produce greater religious involvement than personalized mail. 3. Different ferms of contact will produce identical religious involvement in subjects of either sex. 4. Different forms of contact will produce identical religious involvement when used by workers of different professional status. The Need for the Study Religious groups traditionally have used personal contact of various kinds in promoting the religious involvement of their members and the unchurched. The face-to-face nature of worship and preaching, the minister's warm greeting at the door, small instruction and Sunday School classes, social and recreational activities all involve, potentially at least, personal contact. Although sociologists have usually classified churches as "secondary groups" (Young, 1942, p. 23), they attempt to become as much as possible "primary groups," seeing themselves at best as the "People of God." Ministers have been exhorted and trained to use intense personal contact in the pastoral care of their parishioners (Jackson, 1963). Counseling, calling on parishioners in their homes, and the care of the sick, dying, and bereaved, provide opportunities for intimate face-to-face contact. Yet the value of these activities has 'usually been assessed only intuitively and empirically. There are signs that modern young ministers have lost their belief that calling in the homes of their parishioners is valuable enough to expend the time and effort involved (Jackson, p. 185), but there are no reports of scientific research to use as guides in deciding to call or not to call. Beginning in the 1890's a special area of service developed as religious groups tried to minister to students through the Young Men's Christian Association and the Young WOmen's Christian Association. Particularly after the post-World War II period, Christian denominations became increasingly interested in serving their constituents at colleges and universities, provided their own workers, and began to spend huge amounts on salaries, programs,and buildings. Home hospitality, calling, and other forms of personal contact were stressed by the early pioneers. The Lutheran churches (Arthur, 1966, p. 55) described the task of their campus workers as: (a) preaching, (b) teaching, (c) counseling students and.members of the faculty, (d) calling upon students and.members of the faculty, and (e) recruiting students and faculty members to Christ and the Church. All of these tasks presuppose and involve substantial face-to- face encounter. Many pressures, such as burgeoning student enrollment, commuters, a shift to a small group emphasis to counter student resist- ance to large groups, etc., have forced the typical campus minister to abandon some of the traditional forms of personal contact previously used. Yet.no research has been available to help evaluate the "traditional" strategy compared to other possibilities. Calling, in particular has been neglected according to the reports of those responsible for leadership in the denominational campus ministries. Workers claim that calling is too time consuming, that students have changed their habits and spend much of their time outside their rooms where they cannot be found, that student addresses are almost impossible to keep up-to-date, and that students are threatened by an ecclesiastical visitor, especially in the presence of roommates. They report that students fail to reapond to invitations for interviews or home hospitality after the first few weeks of their freshmen year. It seems to be a growing belief that calling, home hospitality, and interviewing is part of a bygone era and that students will no longer respond to these "traditional" methods. Again, no research is available to guide these decisions. Often, as a result, campus ministry contact is largely by mail for the vast majority of students. Term announcements, letters, devotional booklets, study materials, and student religious magazines are published and mailed by the thousands in the hope that students will be encouraged and helped in their involvement with the Faith. Personal contact is often limited to those who respond to mail. Theplpportance of this Stpgy It is of crucial importance for studies to be made of the effects of various methods of contacting students since the outcome affects: (a) the self image and guilt feelings of the campus worker, (b) the strategy of investments of thme, abilities, and energy, and (0) financial invest- ment. 6 When dedicated employees find their practice at Variance with their employer's expectations and basic job descriptions, they have feelings of guilt and conflict and a wholesome and effective selfeimage is impaired. If the campus worker could be certain that personal contact is ineffective, he could feel justified in making contact in other ways and churches would be encouraged to revise their expectationso If convinced that personal contact were productive, the worker would be encouraged to invest himself in it and the churches would stress even more insistently that willingness and competence in the face-to-faoe aspects of the campus ministry are essential to the successful fulfill- ment of the religious worker's task. Personal contact costs, in dollars, more than a ministry by mail since more personnel mustlbe involved. If mail is just as effective, the student load per worker could be substan- tially increased. The churches have limited funds to invest and need to know whether personal contact is worth the greatly increased cost it involves. Yet with hundreds of workers in the campus ministry and millions of dollars being spent annually, there is little or no research available which can guide decisions or determine strategy. Churches must decide where their limited resources of manpower and money can be most effectively used. Careful studies in Specific areas under consideration are not only helpful, but essential. Organization of this Thesis In Chapter I a discussion of personal and non-personal forms of contact used by the church in promoting religious involvement, espe- cially in the campus ministry, is presented. Definitions, limitations of the study, basic assumptions, and research hypotheses are given. The need for the study is discussed. Chapter II is a review of the literature related to this study. 7 Chapter III contains a discussion of the general design and methodology of the study. In Chapter IV the analysis of the data and the results of the study are reported. Chapter V contains a summary, conclusions, discussion, recommenda— tions, and the implications of this study for further research. The theoretical bases of this study will become more clear in the review of pertinent literature in the next chapter. CHAPTER II A REVIEW OF RELATED LITERATURE Attitudinal Changes of College Students A large number of both longitudinal and cross-sectional studies support the notion that students in college can and do change, an assumption on which the present experiment is based. Freedman (1960) in a report on a longitudinal study underway at Vassar College claimed there were substantial changes in students between the freshman year and graduation. Seniors tended to be more mature, less stable, less "feminine", less authoritarian, displayed greater religious liberalism, demonstrated greater acceptance of intellectual values and greater internal conflict than freshmen. Nelson (1938) studied the four under- graduate classes at 18 institutions and found that the freshmen were more homogeneous in attitudes than seniors, that they were more conserva- tive, more favorably disposed towards religion, more likely to indicate their belief in God, and more likely to report that they attended worship. Both Arsenian (1943) who studied males at Springfield College and Newcomb (1943) who studied students at Bannington agreed that changes take place from.the freshman to the senior year. Jacob's (1957) study and exhaustive review of studies on college students' attitudes remains one of the definitive studies to date. He concluded that education had little effect on student values, that the overall result of a college education was to bring about general accept- ance of a body of standards and attitudes common to college graduates in 8 the American community, that only a few "potent" colleges (small, liberal arts, residential, highly selective) had any significant impact on student values, and that instructors had little or no influence. Heisman (1958) seriously questioned both the methodology and conclusions of Jacob. Wise wrote: The recent literature on the question has rendered a substantial service to the higher education by challenging the assumption that our colleges are successful in educating for values. What- ever the limitation of the research methodologies and in spite of the tendency to generalize too much from the data, most of us are now convinced that college has less influence on students than previously assumed. One reaction to this disillusionment is to assert that colleges have no important influence on student values, but such a reaction fails to recognize the substantial data which strongly suggest that some colleges do influence student values (Wise, Hodgkinson, Rogers, & Shafer, 1964, p. iii). A careful longitudinal study at Michigan State University, the same setting as in the present study, continued to challenge Jacob's conclusions. "...it is difficult to assume as did Jacob that neither courses, nor instructors, nor instructional methods have a marked impact on students' values" (Lehmann & Dressel, 1962, p. 19). The results of the Lehmann and Dressel study must be taken into account in the interpre- tation of related research, eSpecially at Michigan State University. Among the major findings, the fellowing seemed to be relevant to the present experiment in particular to the control of certain variables: (a) IMales were more stereotypic and dogmatic (in which respect they showed a relationship to their rural/urban background), scored higher in the theoretical, economic, and political scales on the Study of Values Test, and were less receptive to new ideas than females. (b) No differences were feund between children of native and fereign-born parents. (c) Freshman students exhibited the greatest magnitude of changes in value orientation and there was evidence that most major 10 changes took place in the first two years. (d) Religious changes took place mainly toward greater liberality and realism and a large proportion of the students felt that they had become less attached to a religion. (e) A small percentage of students changed in a negative direction. (f) Experiences in living units (particularly discussions and "bull sessions") were the most significant for the students. (g) Peers were a consider- able influence to all, but to women more than men. (h) Instructors seemed to have a reinforcing or modifying effect on students' attitudes and values beginning in the junior years (Lehmann and Dressel, 1962, pp. 265-269). Summarizing recent studies (Edwards, 1957; Goode & Hatt, 1952; Green, 1954; Guttman, 1950; Guttman, 1954; Riley, Riley & Jackson, 1954; waisanen, 1960), Felty (1965) concluded that the methodological problems related to attitude measurement could be overcome by the use of an appropriate scaling method: Scale analysis provides a method for determining whether a set of items can be ordered along a single dimension. If a particular attitude universe is really one-dimensional, any sampling of items from it should also be one-dimensional, and should provide an ordering of respondents essentially the same as that provided by any other sampling of items from the universe. If the predicted ordering does not occur, the universe is judged to be multi- dimensional and consequently not scalable. . . . If items do suggest an underlying single dimension, it is meaningful to describe a respondent with a higher total score as possessing more of the characteristic being measured than someone with a lower total score. Most important, if scale properties are obtained, this provides evidence for the existence of a defined body of opinion in the respondent group in respect to the particular area of measurement involved (Felty, 1965, pp. 44-45). Felty also feund experimental evidence (Foa, 1950; Foa, 1961; Guttman, 1947; Guttman, 1950; Guttman & Foa, 1951; Guttman&Suchman, 1947; Suchman, 1950; Suchman & Guttman, 1947) that: intensity will usually form a quasi-scale which, when plotted against the content dimension, will reveal the point on the content 11 scale of the lowest intensity of response. This point has been empirically established as a point of indifference in respect to the item content. Attitudes become favorable on one side of the point and unfavorable on the other side of the point. It then becomes possible to state in respect to a particular group about what per cent of the respondents are actually favorable, neutral, or unfavorable as defined by an objective and invariant referent point (Felty, 1965, pp. 47-48). Motivation to Participatg G. W. Allport (1945) drew a sharp distinction between action and participation. He contrasted mere performance of an activity with "ego- involved participation." The major fields of action were identified as economic, educational, recreational, political, religious, and domestic. Although willing to grant each person a blind spot or two, he held that full participation in several important areas of life was essential to the welfare of the individual himself and ultimately to society: "But unless he is in some areas ego-engaged and participant his life is crippled and his existence is a blemish on democracy" (Allport, 1945, p. 125). Any study which attempts to measure participation must include controls for the variables which affect involvement in groups. One writer (Goldhammer, 1942) has suggested that they are: Sex Age Maturity Religion Education Marital status Residence status (living alone, etc.) Length of residence at present address Length of residence in present city Rental level Occupation Church attendance Social status Income Neuroticism ersssssaqamssms 12 Goldhammer found significant differences for the three variables which he tested: age, education, and certain personality characteristics. A. W. Smith (1948) surveyed the organizational participation records of a group of college graduates and found that their undergraduate experiences, particularly with faculty, were significantly related to the number of postgraduate affliations: The implication here is that the most significant factors which were related to participation in postgraduation organizations of a social or civic nature were the human relations of a constructive, forceful, and friendly nature that the student had as an under- graduate. Evidence would also tend to show that, although experiences involving fellow students were frequently mentioned, it was the faculty that made the greater impression (A. W. Smith, 1948, p. 53). Psychological literature suggests that individuals-participate in organizations in order to satisfy their needs. Maslow (1943) theorized that human needs are pursued sequentially--"lower" ones first and, only after these are satisfied, the "higher" orders are pursued. His hierarchy of needs follows: 1. Physiological needs, e.g. hunger, thirst. 2. Safety needs, e.g. security, order. 3. Belonginess and love needs, e.g. affection, identification. 4. Esteem needs, e.g., prestige, success, selfLrespect. 5. Need for self-actualization, i.e. the desire for self fulfillment. Other writers have found a relationship between the concept of self and participation in groups. In a study using the Twenty Statements Test (Kuhn & McPartland, 1954), 288 college students were asked to write 20 answers to the question, "Who am I?" They found that subjects tended to describe themselves as members of groups and classes (a Baptist, a hus- band) before they described themselves in evaluative terms (a loyal Christian, a good husband). A summary of’major social wants common to Western Man which affect participation in groups is offered by Krech, Crutchfield and Ballachey 13 (1962). The affiliation want draws people to associate with, or just be near other persons. Schachter (1959) found a relationship between the affliation want, anxiety level, and position in the family (first-born, etc.). He proposed that the company of others in a similar plight is anxiety-reducing and that first-born children, when they are made highly anxious strongly prefer the company of others. . . . the desire to belong to some activity or group larger than one's self, to be "accepted," to be a part of something significant, characterizes most people in society. This powerful want frequently motivates the "joining" activities of people. By joining a union the individual worker seeks to gain not only his economic goals, but also to satisfy his want to belong. The members of a church, of a college fraternity, of a scientific association, of the Knights of Columbus, of the American Veterans Committee--all these may find gratification of this want in their various groups (Kretch, et al., 1962, p. 294). The acguistive want drives men to acquire material objects and wealth. The prestige want encourages individuals to strive for status and to avoid social failure. The power want motivates people to desire to control other persons and objects, to obtain their obedience, and to come pel their actions. The altruistic want prompts men to attempt to help others, a phenomenon evident even in nursery school children (Murphy, 1937). The curiosit want, the desire to know creates an almost insatiable interest in exploring the world and manipulating objects in it. These writers concluded that "The thought and action of the individual reflect his wants and goals" (Kretch, et al., 1962, p. 70). If this be true, not only a person's participation, but his attitudes and actions as well are functions of his need to satisfy culturally acquired wants. Scientific literature, understandably, does not discuss a spiritual want. Many writers in the field of religion, however, refer to a longing fer God: "As the heart longs for flowing streams, so longs my soul for thee, 0 God" (Psahm 42). St. Augustine (A.D. 354-430) wrote, 14 ". . . Thou hast formed us for Thyself, and our hearts are restless till they find their rest in Thee" (Pilkington, 1943). Contact The experimental treatments used in this study are various forms of contact. A moderate amount of literature exists on theoretical considerations, scientific studies, types, etc. For the sake of convenience and clarity, contact will be discussed in the following sub- sections: Theoretical Considerations Regardipg Personal Contact The formation of theory about personal contact has been stated comprehensively. Studies have been made in the fields of ethnic attitudes, intercultural relationships, voting analysis, and marketing decisions. Pieces of a theory are often interspersed with the discussion. Lazars- feld, Berlson and Gaudet (1944) suggested that the great effectiveness of personal influence in inducing changes in voting decisions was related to the greater flexibility of face-to-face persuasion. E. Katz and Lazarsfeld (1955) stated: Formal media will influence mainly by representation or by indirect attraction, that is, by what they tell. People, however, can influence both this way and by control. People can induce each other to a variety of activities as a result of their interpersonal relations and thus their influence goes far beyond the content of their communications. This is probably the most important reason why we have found the impact of personal contact to be greater than the impact of formal media. . . . It is fair to say that persons have ppp_major avenues of influence while fermal mass media, like rggdo and print, have only one (E. Katz & Lazarsfeld, 1955, pp. 185, 1 . H. E. 0. James (1955) found personal contact indispensable in promoting favorable attitudes toward persons of different ethnic and cultural back- grounds and explained: 15 In personal contact there is response to oneself. Unlike events in books and films, what happens is affected by urgent interplay between the participants. This interaction is dynamic in that it is concerned with issues that matter to the participants, and exploratory in that the responses enable the participants to find out what they want to know. Because of these features, because it answered or seemed to answer the questions that mattered to them, personal contact was not only an effective maker and undoer of attitudes but often seemed to these children the fundamental, sometimes the only valid basis for ethnic attitudes. In personal contact they felt that they penetrated to a person's innermost "nature," as they called it; they really got to know persons and through them peoples (H.E.O. James, 1955, p. 68). He further hypothesized that if personal contact is to affect attitudes toward members of different ethnic groups, its representatives must be seen as nearly equally as possible as both persons in their own right and as members of the different group. McGuigan (1958) studied personality changes induced in a group of college students by a year of study, living with a native family, and travel in Europe and found so few changes that he concluded that modifications of personality as a result of this kind of personal contact are relatively rare. A more systematic theory was presented by Cook and Selltiz (1955) based on an analysis of over 30 studies concerning attitude change to- ward ethnic groups in the United States. They emphasized that personal contact could lead either to positive changes or to generally negative attitudes and even to actual violence depending on the manipulation, pintended or accidental, of the variables. The variables were analyzed according to three aspects of the contact. Regarding the contact situation, they found them to be: (a) extent in time, (b) implications as to the social acceptance by each individual of other persons partici- pating in the situation, (c) the extent to which the situation requires interdependence or independence or permits either, (d) the relative 16 ppgppp of the different ethnic groups represented, and (e) ppppp within each ethnic group with respect to attitudes and behavior toward the other group. The most important aspect in the contact situation was found to be "intimacy": A number of investigators have considered a characteristic of interaction which they have referred to loudly as its "quality" or "intimacy." All such studies have feund a clear relation between quality of contact and intergroup attitude--the greater the intimacy of the contact, the more favorable the intergroup attitude (Cook & Selltiz, 1955, p. 53). The predictors of "intimacy“ they found were the "acquaintance potential" (opportunity to get to know) and the "social acceptance implication" (willingness to accept as a social equal). Regarding the individual in the contact, the variables were (a) the extent that individuals differ from.commonly-held unfavorable sterotypes, and (b) the extent to which they resemble subject individuals in background characteristics, interests, etc. A set of variables (dependent) related to the manner in which changes have been looked for. Part of these were isolated by Cook and Selltiz by examining the measuring instruments used in previous studies. They cited: (a) beliefs, (b) feelings (especially those associated with ethnic groups), (c) policy orientation (appropriate way to treat), and (d) behavioral (observed actions). By examining the criterion variables in studies they found the following to be determinative: (a) specificity-- generality (investigators who used measures specific to particular individuals rather than ethnic groups have found favorable change with contact in beliefs, feelings, or policy orientation), (b) beliefs vs. feelings and policy orientations (many investigators have feund only changes in beliefs and the writers hypothesize that cognative aspects are. most easily changed by contact), (c) generalized feelings and/or policy l7 orientation (diverse results are found when reference is to a group as a whole rather than to individuals) and (d) behavior (all studies resulted in the contradictory findings that there was a greater change in behavior toward different ethnic groups as a result of personal contact than in attitudes toward them) (Cook & Selltiz, 1955, pp. 53-55). Some writers have attempted more general theories. Expanding on a suggestion from Krech and Crutchfield (1948), Festinger and Kelley (1951) summarized theories regarding attitude change. The perceptual control theory states that a person's knowledge about others is related to his attitude toward them and that changing knowledge may result in attitudinal change: The pptivational control theogy states that attitudes may be controlled through manipulating the need systems relevant to them. These writers pointed out the difficulty in inducing personality changes. The third, and most relevant to this study, is the social control theory—- which declares that attitudinal change may be induced by bringing a person into social contact with evaluations or attitudes that differ from his own. ". . . it is not information about the world that counts but what other people think about it, how they evaluate it, interpret it, what attitudes they hold about it" (Festinger & Kelley, 1951, pp. 8, 9). After the completion of their experiment in a housing project described below, they offered the following communication theory of attitude change: I. Contacts are effective in producing attitude changes only if they entail the transmission of social attitudes, i.e. evaluative statements by other persons about the objects of the attitude. Such evaluations do not need to be communicated verbally or explicitly. They can be evidenced in the behavior of the other persons toward the objects of the attitude. II. The effectiveness of a communicated attitude in producing change is a function of various aspects of the relationship 18 between the communicator and recipient, e.g. their inter- personal attractions, power relationships, etc. . . . Thus, for example, if a group of persons are held together by a common interest in community activities, and carry on communication about attitudes and opinions on which they differ, the conditions for attitude change are present. III. Contacts will not be effective in producing attitude change if they merely contain information about the objects of the attitude. Such information may change the recipient's cognitive structure of the world, but will not change his evaluations of it or emotional feelings about it (Festinger & Kelley, 1951, p. 76). Homans (1961) theorized about the basic nature of personal influence in the most elemental forms of social behavior which he describes as "face-to-face contacts between individuals." His conclu- sions are based on his previous research on small group situations (Homans, 1950). If elementary social behavior is the face-to-face contact between individuals, in which the reward each gets from the behavior of the others is relatively direct and immediate, then the study of elementary social behavior depends on observation of such face-to-face contacts, and these cannot easily be made of any large number of persons at one time. Accordingly, my subject is often called "the study of small groups"; I have called it so myself, but I am now persuaded that the name is misleading. Small groups are not what we study but where we study it (Homans, 1961, p. 7 . The same theorist described the variables of face-to-face relationships on which effective influence depends: (a) frequency, (b) sentiment (like a kiss, a sign), (c) interaction, (d) quantity (can be reduced to time), (e) value (amount of activity put out to get desired reinforce- ment), (f) "norms" (related to confermity), (g) repetition (if in the past a particular stimulus--situation has been desirable) and (h) profit (favorable exchange, not accounting profit). The theory is stated in five general propositions: 1. If in the past the occurrence of a particular stimulus— situation has been the occasion on which a man's activity has 19 been rewarded, then the more similar the present stimulus- situation is to the past one, the more likely he is to emit the activity, or some similar activity, now. 2. The more often within a given period of time a man's activity rewards the activity of another, the more often the other will emit the activity. 3. The more valuable to a man a unit of the activity another gives him, the more often he will emit activity rewarded by the activity of the other. 4. The more often a man has, in the recent past, received a rewarding activity from another, the less valuable any further unit of that activity becomes to him. 5. The more to a mania disadvantage the rule of distributed justice fails of realization, the more likely he is to display the emotional behavior we call anger (Homans, 1961, pp. 53- 55, 62). Homans offers the broadest theory to be found in the literature about personal contact. Human interchange is similar to elementary economics. One man influences or leads another if he can provide a reward to the other at a price he is willing to pay. "The open secret of human exchange is to give the other man behavior that is more valuable to him than it is costly to you and to get from him behavior that is more valuable to you than it is costly to him" (Homans, 1961, p. 62). Rewards may be, of course, psychological as well as physical. Personal contact is more effective than impersonal means because the direct influence of one person to the other, i.e., the giving of reward, is performed by face-to-face encounter, not by absent persons or organiza- tions. Certain persons, because they can regularly influence others become the natural leaders in groups. Authority is proportional to the number of persons one is able to influence regularly. Felty (1965) summarized factors relating to the effectiveness of personal contacts in attitude changes from several theorists (Allport, 20 1958; Jacobson, Kumata & Gullahorn, 1960; Rosenberg, 1960; and Zetterberg, 1963) as follows: . . . frequent contact with a person or group is likely to lead to more favorable attitudes, if: (a) the contact is between status equals in pursuit of common goals (Allport, 1958, p. 267), (b) the contact is perceived as instrumental to the realization of a desired goal value (Rosenberg, 1960, p. 521), (c) contact is with members of a higher status group (Allport, 1958, pp. 254, 261-262), (d) if the contact is among status equals and the basis of status is unquestioned (Jacobson, pp g1. 1960, pp. 210-213), (6) if the contact is volitional (as reinterpreted from Zetterberg, 1963, p. 13), (f) if the contact is selected over other rewards (as reinterpreted from Zetterberg, 1963, p. 13) (Felty, 1965, pp. 31- 32) . Research on Personal Contact The slowly developing hypotheses about the nature and effectiveness of personal contact have been tested by scientific research in many areas. F. T. Smith (1943) conducted one of the earliest experiments in changing attitudes toward Negroes by using personal contact. Students of educa- tion were exposed to intellectual and social contacts with leading Negroes fbr four days on two consecutive weekends. Marked increases in favorable attitudes toward Negroes were registered, using primarily the Thurstone- Hinckley test, in the experimental group while a control group remained practically constant. A retest after 11 months showed a high persistence of gain. Of the original 46 students, 25 had kept all of their gains or increased them.and losses were concentrated in a few special cases. The modified attitudes did not seem generally to result in action of social significance, however. Cook and Selltiz (1955) analyzed more than 30 studies in changing attitudes or behavior toward ethnic groups based on personal contact methods. They found that there were no significant differences in only 21 three instances, and that 27 studies were almost equally divided between distinctly favorable gains and qualified improvements in attitudes. An experiment to change attitudes toward those of a supposedly lower social-economic group in a housing project (Festinger and Kelley, 1951) involved induced personal contact through community projects, home economic demonstrations, and recreation. For a period of eight months, 60% of the tenants in the project took part in activities which gave them new contacts with people they had formerly rejected as "low class." The results from.a series of surveys showed that favorable contacts in the community activities brought a large and steady increase in improved attitudes and invitations to homes. Those who had no contact or had unfavorable contacts showed no change. Researchers have made more descriptive than experimental studies. No study has been found which is closely identical to the present investi- gation in methodology. A high degree of similarity to the present study would require an experiment using personal contact and some impersonal medium in an attempt to modify attitudes or behavior. No such study has been found. A partially similar design was followed by Katz and Lazars- feld (1955), who surveyed the effect of various influences on the purchase decisions of women comparing different forms of personal contact with mass media for marketing shifts, motion—picture selections, and fashion changes. The data were analyzed to provide an "index of effectiveness" for each kind of influence. For marketing shifts (change from one brand to another, etc.), personal contact was rated at .39 compared to .25 for radio advertising, .07 for newspaper advertising, .07 for magazine advertising; the influence of salespersons was .18 even though their own self interest might have made women wary. For motion-picture selection, 22 personal contact was rated .33 compared to .06 for newspaper and .14 for magazines. For fashion changes, personal contact was separated into "verbal, personal" and "visual, personal." Both factors received an index of .16 compared with .14 for saleSpersons and .07 for magazines. The dominant role of personal contact now comes out fairly clearly. In two areas (marketing and movies) it is above the average on both frequency of exposure and effectiveness. In the fashion area, it is high in regard to exposure and, as we remember, relatively more frequent than any other factor; but its index of effectiveness is not pronounced. This is the only exception to the general conclusion that the impact of personal contact is greater than that of any other Sources investigated by this study (E. Katz and Lazarsfeld, 1955, pp. In a somewhat poorly controlled experiment, H. E. 0. James (1955) and two students attempted to change attitudes toward persons of African, Chinese, and Indian cultures in an English school by inviting persons from.such cultural groups to come to the school as visiting teachers and to teach lessons about their cultures. In H. E. 0. James' own experiment involving the African culture, two teachers were put in charge of two classrooms for two weeks. Interviews showed that 41 children had made favorable gains in attitudes while two remained the same. In a previous control period, before the African teachers came, 40 had net changed while one had gotten worse and two had improved. Results for Chinese and Indians were similar. Although personal contact in the studies cited above seemed to affect attitudes and/or behavior, other evidence indicates that personal- ity changes were rare. 'McGuigan (1958) compared a group of college students who went to Europe to study and travel for a year to a group who had stayed at their own school. The hypothesis tested was that intercultural personal contact would have a modifying effect on personal- ities. A large battery of personality tests showed few significant 23 results. Personality changes were not found in an experiment by Sawyier (1964). After 15 weeks in a group counseling situation, prison subjects failed to show significant differences over control groups in antisocial responses to the Thematic Apperception Test. A study in Costa Rica of attitudes toward physical disability among various occupational groupings in society, which included some aspects of the effects of personal contact, was conducted by Felty (1965). He found in general that the frequency of personal contact with physically disabled persons had little or no influence on their attitudes. Among those occupational groups directly involved with the handicapped, effects were found, but in a negative direction which was the reverse of the predicted direction (Felty, 1965, p. 103). Felty reasoned that the negative direction of changes among occupational groups immediately involved with disabled persons could be explained in two ways: (a) that there is a tendency for people to become less favorably disposed to them as they are more frequently involved with them, or (b) that the attitude instrument might have measured only a limited portion of the attitude universe re- lated to handicapped persons (Felty, 1965, pp. 169, 170). Personal Contgct by Church Workers No scientific experiment to the writer's knowledge has been under- taken to measure the experimental effects of various forms of contact on the behavior of persons in a religious context. However, on the basis of personal observation and experience, church leaders have valued personal contact. Hoh, a theological professor, warned about substituting mail, small groups, programs, or the telephone for visiting church members individually to solicit a financial pledge: 24 They can get pledges maybe; but they cannot cultivate steward- ship. The contact of member with member, the influence of one Christian personality upon another, the stimulation of visitor and visitee to greater consecration, the discovery of usable hours and talents, the closer integration of the congregational family by better understanding and larger sympathy, the joy of personal fellowshi --these cannot be had through substitutes (Hoh, 19449 P. 178 . A doctor and a hospital chaplain claimed great benefit to patients for personal contact in the sickroom: We are persuaded that the minister has a place in the sickroom, a place not that of the doctor, of the psychiatrist, of the social worker or anybody else. The minister's duty there is to rouse the great energies, certainties, and faiths of the Christian religion (Cabot & Dicks, 1945, p. 5). At the same time questions have been raised about the practicability of calling and visiting by ministers. "Perhaps no part of the work of the modern minister is more debated than the practice of pastoral calling" (Jackson, 1963, p. 185). The same writer explained that a "new approach" to pastoral calling is dictated by changes in society, the fragmentation of the family in particular. Recognizing the many demands upon the modern minister, he nevertheless advocates extensive calling as an essential part of "pastoral care." Lenski (1963) has held that the quality of contact for ministers is especially influencial with groups associated with their own churches. "Both preaching and pastoral functions bring priests and.ministers, far more than other people, into 'group-relevant' contacts with other members of their group" (Lenski, 1963, p. 286). Recent literature on the specialized ministry to colleges and universities has made similar recommendations without reporting any scientific experiments. "The 'face-to-face' ministry is still one of the greatest reaponsibilities of the university pastor" (Kemp, 1964, p. 19). 25 Eulogizing one of the best known "student workers" of all time, John R. Mott, the same writer reported: "Effective as he was from the platform and although he had a genius for organization, he became increasingly aware of the value of individual face-to-face contacts" (Kemp, 1964, p. 21). Traveling the world and Speaking to many students on many campuses, Mbtt often spent many hours in seeing individual students in 10 to 15 minute interviews. Matthews (1934, p. 151) felt that "this was, perhaps, the most significant contribution of Mott's extensive travels." Chamberlin (1963) pointed out: Discovering where a person is requires that the pastor go there too--not only geographically (as in calling in the home or dormitory), but psychologically, theologically, aesthetically, and intellectually. . . . He cannot share the world of a student if his real reason for calling upon him is to promote church attendance, to recruit a Sunday school teacher, or to enlist help with organizing a student or faculty group. Rather, he must come as a servant of the person ready to give up his prerogatives in order to meet the person. . . . The pastoral office is express- ed as a personal relationship between two individuals. The pastor may represent the whole mission and concern of the Christian Church, but when he meets a student or teacher he must let that mission be reflected through his whole personality. He does not bring some "thing" to be pronounced, or delivered, or handed out. He brings only himself (Chamberlin, 1963, pp. 178, 179). Contact by Mail The general agreement above that mass media are relatively less effective than personal contact does not prevent, as any business man or householder knows, common use of written materials sent through the mails to solicit action, participation, purchase, and to influence or assess attitudes and opinions. Sims (1941) used written communication to influence attitudes toward the Tennessee Valley Authority. Although his main interest was to study the effect of differences in the content of the various appeals, it is especially relevant to the present study that 26 all articles tested had some effect on the attitudes of his subjects. A test administered shortly after reading the materials showed mainly positive changes along with an occasional negative shift. Temporary effectiveness of written persuasion was suggested, however, by the fact that these changes decayed rapidly over time. Response to requests by mail depends on various factors: The vital point seems to be that of reasonableness. . . . if the investigation (request) is justified and reasonable, the results will be nearly the same whatever method is used in gathering the material and whatever may be the garnishing placed upon the letter itself (Boddy, 1932, pp. 84, 85, 86). Another major factor, Allgeier (1954) suggested, was personalization of letters. An experimental group of debtors was sent collection letters written in a friendly, warm style while a control group received reminder letters composed in stiff, stilted language showing little or no interest in the reader. Results showed that personalization produced few differences during the first six months, but significant differences after that period. Suggestions were made (Hodgson, 1964, pp. 727, 729, & 883) regard- ing various practical aspects. When personalization is important, an appropriate printing method, such as actual or apparently individually typed and signed letters should be chosen. When testing is underway, bias introduced by mailing on different days should be controlled by sending all letters in the same mail. Buckley (1961) informally analyzed various criteria involved in writing a successful sales letter. Ferber, Blankertz and Hollander (1964) surveyed the extensive research on advertising effectiveness. Extensive analysis was reported for newspapers, magazines, radio, and television, while neglect was indicated for research in personal salesmanship and informal personal 27 communications. They pointed out that direct mail provided its own research possibility since its message often calls for direct responses that can be accurately checked. Reference was made to a large number of controlled experiments which have led marketing researchers to believe that direct mail is an important and profitable medium of contact with consumers. Hodgson (1964) listed criteria for testing direct mail effectiveness based largely on the tabulation of responses or sales, but including some sophisticated research methods, especially in the deter- mination of sample sizes. He recommended that churches use direct mail in fund raising and the encouragement of more regular attendance by mail- ing messages such as an advance copy of the Sunday church bulletin. Fellows and Koenig (1959) applied direct mail techniques to raising money for churches. Starch (1951) tried sending an unmarked circular to proSpective customers through the mail without including a letter. Interviews made of a random sample of customers in retail stores showed that customers who received the circular bought more than those who had not. Another writer points to both hope and despair about the plight of mass media in ecclesiastical use: Mass communication, while not being individual communication, frequently bears some of the earmarks of "personal" communication in a way whose power we have not even begun to probe. . . . It is ironical that mass communication, while it exerts a tremendous influence upon individuals, yet can never achieve the kind of degree of impact that is the result of personal communication representing a relationship of persons. Mass communication may indeed move an individual, but it may not establish a personal relationship (despite all the exploited empathy up anybody's sleeve) (Boyd, 1957, pp. 110-111). 28 Measurement of Religious Involvement The empirical analysis of religion was begun more than sixty years ago by Starbuck (1899) who made a careful examination of 430 personal documents, especially on the theme of conversion, and attempted by means of descriptive categories to determine the course and results of adoles- cent religion. He used averages, modes, and percentages in his analysis. W. James (1902) in his landmark work drew heavily on Starbuck's data, added studies in depth of a few individuals, and carried forward some of Starbuck's analyses. D. Katz and F. H. Allport (1931) studied student religious atti- tudes using a questionnaireand.found that nearly two-thirds of students in all classes changed their religious beliefs in college. One of the major items of change was from belief in a personal creator toward more impersonal concepts of God. G. W. Allport, Gillespie, and Young (1948) surveyed the religious beliefs of 500 Harvard and Radcliffe students. They reported that 58% changed in college and that 32% indicated that they had become agnostic or atheistic. Their questionnaire was called an Attitude Inventory. It consisted of 17 items, some of which were cognative such as age and marital status, but in the main were composed of direct questions or attitudes which could be responded to by checking an appropriate multiple-choice statement. The same instrument has been employed in a number of subsequent studies. The editor of the Harvard Crimson (1959) in consultation with Riesman used the G. N. Allport questionnaire in connection with similar items about political beliefs in a survey of Harvard and Radcliffe students. Rossman, a leading specialist in the field of the church and higher education, accepted these measuren ments as valid descriptions of the relationship of students to their 29 religious faith: One comes to the puzzling conclusion, on the basis of the poll, that Harvard as Harvard really does not have a great influence on students' religious ideas. 0f the 65 who had experienced a reaction to the religious tradition in which they were raised, only 21% reacted against it while at Harvard, nearly three- quarters in their freshmen year. The majority (62%) reacted in secondary school (Rossman, 1960, p. 29). The identical questionnaire used in the Harvard Cripppp study was the instrument in a study of the relationship of religious and political attitudes of students at 12 undergraduate, liberal arts colleges through- out the country (Bell, 1963). The editors commented regarding the questionnaire that ". . . its complexity was sufficient to record the subtlest tendencies of opinion and motivation" (Bell, 1963, p. 280). Osgood, Ware, and.Morris (1961) used the semantic differential technique to measure religious attitudes. Male college students preferred the following items on a scale of 27 adjectives in a "Way of Life" survey as a description of their feelings about religion: good vs. had, strong vs. weak, active vs. passive, and stimulating vs. dull. Using a similar semantic differential method, McCann (1955) found in older subjects a preference for the "Able" vs. "Good" orientation in the religious dimension. In a summary of research methods in religion, Havens (1963) comments that most studies have used interviewing, that some of these were supplemented by objective data and observational information, and that many studies gathered data on student reports of their own feelings and attitudes. Carney and.McKeachie (1963) employed scales from the California Psychological Inventory and the Thematic Apperception Test in a study of students and found that it correlated well with religiosity. Vernon (1962) compared the direct measurement of religiosity using questions such as: "How important is religion in your day-toeday 30 living?" to the indirect approach on the Twenty Statements Test (TST) using 20 different responses to the question: "Who am I?" He found that more respondents were willing to provide positive answers on religion in response to direct questioning. Whether or not subjects identified themselves religiously on the TST, they diaplayed highly favorable attitudes toward religion. "It appears that it is enough to be 29; religion without being religious" (Vernon, 1962, p. 164). Vernon also pointed out that, "if one attempts to measure the 'deeper' aspects of what has been called religiosity, the problems multiply tremendously," and, "Not much has been done along this line" (Vernon, 1962, p. 159). Toch and Anderson (1960) found no significant relationship between the religious beliefs and denominational affiliation of 42 Muchigan State University students using a questionnaire of direct questions on beliefs in the Christian Faith. In a pioneering study, Telford (1934) used church attendance as the dependent variable and found a clearcut relationship between the fre- quency of church attendance and scores of favorable attitudes toward the church. Krech, et a1. (1962) commented regretfully: ". .‘. unfortunately church attendance was not ascertained independently; the investigator accepted the individual's own report" (Kretch et al., 1962, p. 160). Lazarwitz (1962) observed, "not surprisingly, church attendance is a fairly complex variable which depends, in addition, upon a whole range of religious feelings and attitudes" (Lazarwitz, 1962, p. 84). Buss (1964) criticized Hadden (1963) for using church membership as a dependent variable and suggested that church attendance would have been a much more meaningful measure. Doherty (1964) theorized that: 31 Religious affiliation and attendance at worship services are measures of religious involvement. But in testing for the incompatibility of scientific and religious values, one wonders, especially in the case of Catholicism, if it is not necessary to penetrate beyond mere external criteria of involvement (Doherty, 1964, p. 240). The object of the above criticism was a study by Stark (1963) which used only religion affiliation and attendance at worship as the dependent variables. Stark, incidentally, is the first known writer to use the term "religious involvement" to describe the "deeper" aspects of the religious dimension. Klausner (1964) reported on a study of data collection methods and dependent variables used in 130 studies involving religion. He listed the following methods: Selfeadministered questionnaires 44% Individual interviews 26% Examination of statistical records 19% Content analysis of mass-media or documents 10% Field observations 9% Projective techniques 7% Syntheses of previously published data 5% Interviews with informants 3% Laboratory experimental methods 2% Demographic schedules, original census data 2% The most common variables used were: Concrete behavior (church attendance) 22% Personality (ego-strength, intrinsic religious commitment) 48% Society (sect-church organization) 15% Cultural (normative system, acscetic Protestantism) 15% He also commented that there seemed to be a relationship between the choice of data collection and the profession of the researcher and recommended that the appropriateness of the method of measurement should be related only to the exigencies of the research object. Glock (1962), elaborating on ideas presented earlier (Glock, 1959), stated: Religiosity is a complex and multi-dimensional phenomenon. It may be conceptualized as consisting of five analytically discrete, 32 though inter-related, dimensions: 1. Ideological (belief) 2. Intellectual (knowledge) 3. The Ritual (religious participation, e.g., church attendance) 4. Experiential (feelings of contact with divinity) 5. Consequential (acts in everyday life which follow from the ideology) (Glock, 1962, p. 7). He found that all subjects who were orthdox attended church more and concluded that religiosity does include cultic behavior. Havens (1963) recommended that many viewpoints are needed to measure religiosity, both objective measures and phenomenological studies of a few cases in depth. Summary The review of the literature has reported evidence that the atti- tudes of college students change including the area of religion. Participation was found to be associated with affiliation, acquisitive, prestige, power, altrustic, and curiosity wants. Personal contact studies have reported induced attitude and behavior changes in marketing, voting, and intergroup relations, but changes in personality from contact have not been found. Personal contact techniques, traditionally used by religious workers, have been recently questioned by frustrated workers. The most common measuring device for religious involvement has been a questionnaire and it has been recommended that a variety of measures, including objective techniques, be used. The variables to be measured in determining religious involvement changes have been suggested to be ideo- logical, intellectual, ritualistic, experiential, and consequential criteria. CHAPTER III RESEARCH DESIGN AND METHODOLOGY The Setting This study took place at Michigan State University and in its immediate vicinity; Michigan State University is a large, mid-Western, land-grant institution of the State of Michigan, located in East Lansing near the state capital in Lansing. The university maintained cooperative relationships with religious campus workers, through the Dean of Students sponsored the Religious Advisers Association, and generally encouraged students to continue religious activities during their university careers. The members of the Religious Advisers Association had the privilege of using the campus mail system. The campus religious workers were supported by their own groups and some were quartered in special student religious centers while others were associated with local churches. Religious preference cards were collected from registering students on a voluntary basis and these cards were distributed to the appropriate members of the Religious Advisers Association. The orientation program for entering freshmen included presentations by religious workers and one evening of Orientation Week was provided for the programs of religious groups. Approximately 22 student religious groups were officially recognized by the student government. The university campus is divided from the residential and business sections of East Lansing by a wide boulevard. Residence halls were located on the relative fringes of the campus and the_distance from 33 34 student religious centers and local churches to the residence halls varies from.one block to two miles. Because of housing regulations requiring most freshmen to live in the halls and regulations prohibiting them from operating automobiles, freshmen experienced difficulty in reaching their respective churches and centers of religious activity. No public trans- portation was available and walking was the most usual means of trans- portation. The "Centrex" system provided for a telephone in each residence hall room. As many as 60% of the students left the campus on a typical weekend. University Lutheran Church was the center of religious activity involved in the off-campus aspects of the study. Of traditional church architecture, its building houses a worship chamber, the offices of the religious workers who participated in this study, a student lounge, and other facilities especially designed for a campus ministry of the Lutheran Church. It is a congregation of the American Lutheran Church which, together with several other national Lutheran Churches, conducted a united campus ministry. Both the cost of the original structure and the current budget were heavily subsidized by these churches separately and jointly through the Division of College and University Work of the National Lutheran Council. There were threeprofessional religious workers on the staff and two had extensive duties in the campus ministry. One of these workers was an ordained clergyman, the other was an unordained lay woman. Students made up half of the attendance at worship on a typical Sunday and Services were held at 9:00 a.m., 10:15 a.m., and 11:30 a.m. Ten to 15 other events of special interest to students were offered during a typical week. Almost half of the governing board of the congregation, the Church Council, and a part of all committees were composed of students. 35 They were given the privilege of Student Membership if they desired, even the right to vote in congregational meetings, without changing their permanent membership away from their home-town congregation. The experimental treatments were administered as though they were parts of the regular campus ministry of this congregation and its staff. The Sample The subjects for this study consisted of all the single entering Lutheran freshmen for whom University Lutheran Church received religious preference cards in the Fall Term of 1961. Almost all were within the usual age range for freshmen--l7 or 18 years old. They were approximate- ly evenly divided between men and women. The homes from which they came were almost all located in Michigan and varied from small farm to large metropolitan settings. All indicated, of course, a preference for the Lutheran Church on their cards, but about 20% stated that they were not members of any church. They represented about 4% of the total freshmen class. All but a few whose residences were in East Lansing or Lansing lived in university residence halls, the closest of which was four blocks and the farthest, almost two miles away from the church. Freshmen in their first term were chosen in order to keep experimental contamination coming from.previous contact with campus religious workers or some other university Lutheran church to a minimum. Lutherans only were chosen in order to be compatible with the setting and workers involved and because the study was designed to show the effect of a denominational ministry on students of that particular denomination. 36 Grouping The stratified randomized design called for 180 students to be assigned to six groups of 30 each. Three groups each were assigned to each worker fer the administration of three different kinds of treat- ment as indicated in Table 1. TABLE 1.--Assignment of groups by type of treatment and workers. Group Worker Treatment Sample Size A1 Campus pastor Face-to-face con- 30 tact (3 attempts) A2 Lay campus worker Face-to-face con- 30 tact (3 attempts) B3 Campus pastor Regular personalized 3O mail B4 Lay campus worker Regular personalized 30 mail C5 None One printed brochure 30 C6 None One printed brochure 30 Since there were no essential differences in groups C5 and C6, they were usually combined for computational and analytical purposes. Specifically, subjects were assigned to groups by using informa- tion contained on the IBM religious preference cards. They were stratified for sex, church membership, campus residence location, and size of home town. Each card was coded for these variables by markings made on the backs of the cards which were blank except for IBM punches meaningless to the experimenter. Similarly coded cards were placed in piles face down, shuffled, numbered consecutively, and then drawn for the 37 six groups by the use of a random numbers table. The total number of subjects was 180. Since the available subjects were limited in supply, it was not possible to provide a perfect stratification of all classi- fication variables. Priority was given in such cases to the sex and church membership variables. The groups contained a relatively small number of subjects randomly mixed across treatment groups for the size of home town and area of campus residence variables. The specific make-up of each group of 30 is shown in Table 2. An equal number of men and women was Table 2.--Composition of each treatment group of 30 individuals by sex, church membership, type of home town, and area of campus residence. Classification variable . Males Females L V' r—V Church.Memberphip Member of a Lutheran church . 12 13 Non-member ~ 3 ' 2 . _ Type of home tOwn. ' Metropolitan " 3 4 Town and country 11 7 Randomly mixed across treatment groups 1 4 Area of campus residence (from.University Lutheran Church) About 1/2 mile , 2 11 l - 2 miles 9 3 Off-campus l 0 Randomly mixed across treatment groups 3 1 in each group, although the pattern of the other variables was different for men and for women. A small bias for greater religious activity for females may have existed since each male group contained three non-church members while each female group contained only two. No bias existed in the selection of the members of the groups. Code numbers were assigned to each subject and thereafter the use of names was avoided whenever possible. 38 The Workers Two workers were involved in the administration of experimental treatments. One was an ordained Lutheran clergyman who served as campus pastor at Michigan State University and, with another clergyman, as pastor to University Lutheran Church. He was in his early forties and his twelfth year of university ministry. The other worker was an un- ordained, lay woman employed by University Lutheran Church and assigned exclusively to campus work. She was in her late twenties and in her sixth year of university ministry experience. The campus pastor generally were a clerical collar in his personal contacts and sometimes transported students in his car. The lay worker wore no distinctive churchly garb and had no car. Both had offices at University Lutheran Church and lived no further than two blocks from it. The Instrument No instrument to measure the full scope of religious involvement has been published nor exists to the knowledge of the experimenter. Chapter II contains a discussion relative to measuring the variables and pointed out that suggestions were made by various writers that many variables should be included in a measuring device. It was necessary, therefore, to design an original questionnaire (Appendix A). Every effort was made to use elements of existing instruments from previous research. Approximately half of the items used in the questionnaire were obtained from these sources and the other half were original. In addition to the questionnaire, registration cards (Appendix A) were used at University Lutheran Church for 17 Sunday mornings during the experi- ment. In order to conceal the fact that an experiment was being 39 conducted, the instrument was designed to appear to be an evaluation of the campus ministry of University Lutheran Church. Specifically, the items in Part II of the questionnaire were from the Senior-Year Experience Inventory of the Cooperative Study on Attitudes and Values of College Students (Lehmann & Dressel, 1962, p. 296). These were scrambled into a pattern differing from the original form. This portion of the instrument was designed to measure the religious belief (ideological) variable. The items in Part III were from the Differential Values Inventory of the same study (Lehmann & Dressel, 1962, p. 22). The Puritan Morality and Relativism scales were extracted and arranged in a forced-choice pattern similar to the original inventory. It was intended that this section measure the religious attitude (intellectual) variable. Items 19 and 20 of Part IV and 8a, 8b, 8c, and 8d of Part V were from the Detroit Area Study, Survey Research Center, University of Michigan, Project 855 (Lenski, 1963, p. 15). Questions 6 and 7 of Part IV and all the items in Part V except 8, 11, 12 were from G. W. Allport et a1. (1948). The last item was open-ended and gave subjects the opportunity to express whatever they wished. Some of these plus the original items were design- ed to measure the action (ritual), and participation (experiential and consequential) variables. Eyperimental Treatment Treatment A consisted of three attempts to make a personal contact with the 30 subjects assigned to each worker. Early in the Fall Term each subject was called upon in his own residence hall and engaged in conversation for 15 to 20 minutes centering about the subject's religious welfare. Appointments were made by telephone three or four days in 4O advance. It was arranged that the worker would be at the residence hall at a given time, call subject in his room from the entrance foyer by telephone, and wait for the subject to come to the main lounge. The conversations were as non-directive as possible. The workers sometimes proposed questions like, "How does God seem to you after a few weeks away from home and your home church?", or "Is there anything that I can do to help you in your relationship to God and the church?". No precise wording nor common approach was pre-determined. The workers tried to conduct these visits according to the apparent needs of the subjects. No attempt was made to promote attendance at events nor joining an organiza- tion except to give the regular term brochure to each subject at or near the end of the conversation. Subjects, however, often asked questions about activities and these were anewered. All subjects in these groups were visited even though considerable difficulty in making contact was experienced in two or three instances in each group. A few subjects were rarely or never in their rooms and were found only after considerable "detective" work. The calls were well received in every case, although a few students were ill at ease. In the second phase of Treatment A, the same students were invited for an office interview with the worker at University Lutheran Church. The appointment was made by telephone several days in advance and a particular time was agreed upon. A few students declined the invitation for a variety of reasons, the most fre- quent being that they did not have time. This was probably only an excuse for some, but others seemed to be anxious about their academic work. About three-fourths of the subjects actually came for an office interview of approximately 20 minutes and all interviews were completed by the Thanksgiving recess. The conversations were more non-directive than in 41 the first phase, largely because the subjects were more at ease, knew what to expect from the worker, and had matured in their academic and personal experiences enough to have developed needs that could be verbalized. Subjects seemed to react positively to this interview come pared to the first contact. The workers' offices were adjacent to the University Lounge and all student interviewees had the opportunity to see it. Those who wished were shown other parts of the church building. If the other worker was present at the time, introductions were made. The third phase of Treatment A began between Thanksgiving and Christmas. As many subjects as possible were invited for hospitality to the home of the worker in groups of five or six. The ordained worker offered transporta- tion in his car. The lay worker offered a meal while the campus pastor offered dessert. The lay worker entertained alone and the campus pastor's family was present, a factor which was apparently mutually appreciated. One-half to two-thirds of the subjects accepted these invitations. Two weeks of the Winter Term were needed to complete this part of the treat- ment. The reaction of the subjects was positive and verbal and non- verbal expressions of friendship were exchanged. Treatment B consisted of personalized mail (Appendix C) from each worker to each of the subjects in the groups. Early in the first term the general publicity brochure was sent to each subject along with a personally signed letter. During the course of the Fall Term, about six mailings were made. In every case some message, sometimes as short as "Hi1", was written on the printed material or on an attached note. In each instance, the name of the worker was signed, often only the first names: "Don" or "Tecla." The mailings were individually planned by each worker, but each student in each group assigned to this treatment 42 received the same materials at approximately the same time. Care was taken in the use of campus mail to make sure that the mailings were actually placed in mail boxes in the residence halls. Names and room numbers were checked on the posted lists in the lobby in order to keep addresses up-to-date. The contents of the mailings included devotional materials, the Lutheran student magazine, printed pamphlets and brochures, and personal letters. The term brochurel (Appendix D) for the Winter Term was sent as part of this treatment during the first week of the term. Groups given Treatment 0 were as close to being control groups as possible. The Fall Term publicity brochure (Appendix D) was sent to each student by campus mail. Inasmuch as every student on record at Univer- sity Lutheran Church was sent such a piece and it was the common practice of other churches in East Lansing to send publicity material to students, it could be said that nothing in the nature of experimental treatment was administered. In a strict scientific sense, however, the sending of the brochure must be considered a treatment. In order to keep the groups as close to being control groups as possible, however, no leaflet was sent for the Winter or Spring Terms. The three treatments were selected to represent realistic patterns of campus ministry contact that could be performed in actual local campus ministries. Conceivably, with substantial increases in workers, programs similar to Treatment A could be employed. Personalized mail operations 1The brochure used in the actual study was in short supply and all copies were exhausted in the study and the regular program. A copy of the brochure published by University Lutheran Church for the Spring Term of the same year is shown in Appendix D. It is similar to the one used in the study except that the original was commercially printed and included pictures of the workers. \ d 43 could be performed economically in local situations, even by one worker, since the messages and signatures could be written by paid or voluntary clerical help. Any local church, even without special financial subsidy or campus workers, could send out a printed brochure at least once a year. To the extent possible, the sample was excluded from.the regular outreach program of University Lutheran Church. Except for that provided in the treatments, no initiative was taken by the workers toward the students in the sample. Random contact was avoided as much as possible, but if a subject took the initiative, the worker responded in his usual manner. Unavoidably, subjects who came to worship and other activities on their own initiative saw or met the workers. A certain amount of contamination occurred in this way. Such contamination tended to reduce the probability of achieving statistically significant results. The Gathering of the Data "Registration Cards" (Appendix A) were requested from all who attended Sunday morning at University Lutheran Church. All worshippers were asked to complete these cards and almost everyone did so. It was not generally known that the workers were gathering data for an experi- ment. Many churches ask for "Record of Worship" cards every Sunday as a regular part of their program. University Lutheran Church had used registration cards once a month for communion services and the only change was to use them on other Sundays. So far as the subjects knew, the registration procedure was normal. It was planned to gather these cards during the Fall and Winter Terms, but some pressure from the church members prompted the congregation to return to the practice of one registration per month for February and March. This data was used in the 44 analysis. Registrations from the Christmas recess when students were normally absent from East Lansing were not used. The use of this church attendance information fulfilled some recommendations for measuring reli— gious involvement found in the literature and provided a means of check- ing the accuracy of the selfLreports from the questionnaires. The questionnaire (Appendix A) was sent to the subjects by campus mail accompanied by a letter (Appendix B) from the chairman of the Campus Work Committee of University Lutheran Church at the end of the Winter Term. The experiment was disguised as a study of the effectiveness of the campus ministry and may also have given the impression that it was an evaluation of the workers. The letter read in part: Your help is needed in a study we are making of our campus ministry. As the local representative of all of the churches of the National Lutheran Council, it is our responsibility to decide how personnel and funds are to be used and what form the ministry should take. Whether or not you have had any contact with this ministry, will you please help us by filling out the enclosed questionnaire? Since we plan to use the period between terms for the completion of our study, it is essential that you return the questionnaire by the end of the classes this term. A return envelope addressed to the chairman at her home was enclosed. No names were requested on the questionnaire itself, but each person was asked to put his name on a separate sheet which, it was promised, would be removed upon receipt and used only to exclude the person from follow- up procedures for those who did not return their questionnaire. The chairman of the Campus Wbrk Committee actually did remove these signature sheets so that the questionnaire could be scored by the use of numbers only. In anticipation of the possibility that some subjects would not sign their names, a concealed code that would indicate the group to which the subject belonged was placed on the return envelope. It was possible 45 to identify the returns from the few students who did not sign their names by the use of the code and, by the process of elimination, even the individuals were made known. After the beginning of the Spring Term, two reminders were sent to each person who had not returned the questionnaire. After another period had passed, a telephone call was made in behalf of the Campus Work Committee to each subject who had not yet responded. Any further pressure seemed to be more than the pretense for the study warranted and no further steps were taken. Statistical Procedures Wherever the assumptions could be met, analysis of variance was chosen as the statistical method. To gain the greatest amount of informa- tion from.the data, a form of analysis of variance using the means of groups and producing a number of subtle contrasts was used (Snedecor, 1956, p. 385) (Cochran & Cox, 1957, p. 62). The statistical design is shown in Table 3. TABLE 3. --Statistical design for analysis of variance computations by treatment groups, sex, and workers. Treatment Worker Group Code Male Female Personal Ordained Al (Mean) (Mean) Contact (A) Lay A2 (Mean) (Mean) Personalized Ordained B3 (Mean) (Mean) Mail (B) Lay B4 (Mean) (Mean) Brochure (0) None 05 8: 06 (Mean) (Mean) The information available from this procedure produced the following effects and interactions: 46 Main effects: 1. Differences from sex of subjects. 2. Treatments A and B compared to Treatment 0. 3. Treatment A compared to Treatment 0. 4. Differences between workers' effectiveness. Interactions: 1. Between workers and sex of subjects. 2. Differences from sex of subjects in C groups. 3. Between sex of subjects and Treatment A and B. 4. Between workers and Treatments A and B. Whenever there was a suggestion that differences arising from the sex of subjects might arise, a separate analysis was made for males and females. In this case, the main effects and interaction obtainable were: Main effects: 1. Treatments A and B compared to Treatment C. 2. Treatment A compared to Treatment B. 3. Differences between workers' effectiveness. Interaction: 1. Between workers and Treatments A and B. Inasmuch as the degrees of freedom for each effect was one and the sum of squares and the mean square was identical, the sum of squares was omitted from the tables in showing results (American Psychological Association, Council of Editors, 1957, p. 32). When the assumptions of analysis of variance could not be met, chi square was chosen as an appropriate statistic. The statistical design was the same as shown in Table 3 except that actual frequencies were used instead of means. When necessary to obtain a sufficiently large expected frequency in each cell, adjacent categories were combined. Whenever appropriate, Fisher's exact tables for chi square (Owen, 1962, p. 479-488) were used--otherwise the standard computation for chi square was employed. 47 A form of chi square analysis was used to test medians (Dixon & Massey, 1957, p. 296) for the one analysis of ranks. In a few instances, the responses called for were ordered (Tables 18 and 19). Since the treat- ments also could be arranged in the order of the amount of personal contact subjects had received, the requirements of the Goodman--Kruskal correlation analysis (Goodman & Kruskal, 1963) were met and the comput- ations were made with the aid of the Univac computer at the University of North Carolina at Chapel Hill. Data were discarded when flaws, most often produced by a large number of subjects failing to answer a particular question, made them questionable. Procedures for handling "missing data" in the computer program would have improved the analysis and are recommended in future studies. The minimum level of acceptable significance was predetermined at the .05 level. Summary The hypotheses of the study were tested in a stratified random sample design whereby each of six groups was subjected to one of three contact treatments by one of two religious workers. A questionnaire was constructed consisting of original items and selections from.previous related research. A registration card was used to gather objective data about church attendance. Statistical analysis involved a form of analysis of variance, chi square, and Goodman--Kruskal correlation analysis. CHAPTER IV RESULTS OF THE STUDY In this study, action, attitudes, participation, and beliefs as operationally defined in Chapter I are treated as dependent variables. Data associated with each of these variables were analyzed to test the following statistical null hypotheses: 1. Personal contact and personalized mail groups show no differences in religious involvement when compared to the impersonal mail group. 2. Personal contact groups show no differences in religious involvement when compared to the personalized mail groups. 3. There is no interaction between the type of contact and the sex of subjects. 4. There is no interaction between the type of contact and the sex and ecclesiastical status of the worker. All data were analyzed when possible. Results which were not significant were not included in this chapter except in cases of special interest or as an illustration of negative findings. Sample Size after Treatment The treatments were administered into the middle of the Winter Term and data were gathered at the end of that term. Followaup procedures were conducted in the first month cf the Spring Term in order to insure an adequate return of the questionnaire. A shrinkage in the sample took place as subjects left the university for various reasons. Those who were not registered in the Spring Term were considered to have left the experiment. No analysis could be performed on the questionnaire 48 49 data from those who left the university since no data could be gathered. This limitation was not binding on data from the registration cards, since they were collected Sunday by Sunday and not made inaccessible by subjects who left the university. The remaining sample is shown in Table 4 as well as questionnaire return information. As the last element TABLE 4.--Size of sample and groups at end of treatment and number and percentage of questionnaires returned. Treatmegt Remaining size Question- Percent- Group of group naires returned age returned Al 25 23 92% A2 26 22 85% B3 26 18 69% B4 28 20 72% cs 28 21 75% 06 28 21 75% Total 161 125 78% aA = Personal contact B = Personalized mail 0 = Brochure in the follow-up procedures, each subject who had not returned the ques- tionnaire was called by telephone and urged to complete and return it. A notation of each conversation was made and they are summarized in Table 5. TABLE 5.--Types of responses given by telephone by subjects who did not return questionnaire. Types of reaponses Number of subjects Refused to complete questionnaire . . . . . . . . . . 14 Gave excuse, but agreed to respond . . . . . . . . . 18 Inconclusive . . . . . . . . . . . . . . . . . . . . 14 Action Data The hypotheses stated at the beginning of this chapter were test- ed for the data gathered from the registration cards for attendance at University Lutheran Church and from the questionnaire action items. 50 Attendance at University Lutheran Church A larger number (F = 6.34, P <2.025) of males in the personal contact and personalized mail groups completed worship registration cards than males in the brochure groups. The analysis is shown in Table 6 and the mean group attendance is diaplayed in Table 7. The orthogonal TABLE 6.-—Analysis of variance for attendance at University Lutheran Church as determined from registration cards by treatment groups, sex, and workers. Source ofVariationb d.f. Mean Square F Entire Sample Sex 1 .67 -- Treatment A.& B vs. C l 2.16 2.00 Treatmenth vs. B 1 4.00 3.70 Workers 1 .97 -- Workers by sex 1 .00 -- Sex within C l 3.49 3.23 Sex by treatment 1 .02 -- Workers by treatment A.& B 1 .70 -- Error 153 1.08 Males . Treatments A.& B vs. C l 5.58 6-34* Treatment A vs. B l 2.28 2.59 Workers 1 .05 -- Workers by treatment 1 .005 -- Error 77 .88 8N = 163 for entire sample; N a 82 for males bA = Personal contact B = Personalized mail 0 = Brochure *P < .025 = 5.29 requirement of the analysis, however, made it necessary that the "Sex in C" item include some material from.the means of all of the groups. A significant result for the "Sex in C" test would enable a rejection of the null hypothesis that there is no interaction between treatment methods and the sex of subjects. It is somewhat surprising that 51 TABLE 7.-€Means of treatment groups for attendance at University Lutheran Church as determined by registration cards by sex and workers. Group Treatment Worker Male Female Al Personal Contact Campus Pastor 4.23 3.54 A2 Personal Contact Lay Campus Wbrker 4.08 4.43 B3 Personalized Mail Campus Pastor 2.79 3.33 B4 Personalized Mail Lay Campus Worker 2.50 2.00 C Brochure None .76 3.64 differences in the means (.76 and 3.64) of the male and female brochure groups were not found. It is possible that the null hypothesis is being accepted when it is false and that a more powerful test would have indicated differences. Figuresl.ahd.2 graphically present the registra- tion card information over time. The graphs show the differences in the patterns of the groups from Sunday to Sunday. There, is some indication that males and females respond to treatments in different ways or, more specifically, that females will attend church services without the personal treatment as well as with it. It has been seen above that the males in the brochure groups made a significantly different reSponse than those who had received personal contact or personalized mail. An examina- tion of Figure 2 suggests, however, that the female groups also responded to personal contact since their performance after Thanksgiving, when office interviews and home hospitality were taking place, was superior to the brochure groups and eventually equalled them in spite of an initially better start made by the brochure groups early in the experiment when ‘treatments had just begun. The_analysis of.attendance data from questionnaire self-reports essentially agrees with the analysis from the registration card data. The tabulation is shown in Table 8 and the means of groups appear in 52 110 ~- 100 - F///p///p Personal Contact (A) Q) ,r 'o / 8 " 3:3; 80 .4. //.F( t I“ W U 70 .JL fl 0/ °H , E P nor to " ’0’ so .. / ’0’ ‘H l ’0’ 0 H )Y’ 3 // g 50 ~~ / ,c’ a i/ I F‘ _ 3 4° "‘ 7/ {7 53 p// f” “euPersonalized Mail (B) 30 l. / f’ / 20 d“ /" ’ {JO /;/ ,nn—Ovv'O/Ar -- "'O" 10 ~L/ / °“°°_ (r V If olra»~.o” "cBrochure (C) («o-“0"" 4 4 4 4 4 4 4 4 —— 2 4. 6 8 10 12 14 16 Sundays on which attendance was taken Figure 1 - Sunday morning church attendance by males at University Lutheran Church as recorded on registration cards. 53 110-- 100" 90.). 'U 8 8 u 80-- u 6 co ‘3 7o~~ g 60$. “-0 O H a 50.1... I! a 1+0» 4.: 30-- ,P ,' Personalized Mail (B) I ,1 20- 10- is. J J I l J_ 7 ' 1 1 T 1 1 l 2 4 6 8 10 12 14 16 Sundays on which attendance was taken Figure 2 - Sunday morning church attendance by females at University Lutheran Church as recorded on registration cards. TABLE 8.--Analysis of variance for attendance at University Lutheran 54 Church a determined by the questionnaire by treatment groups, sex, and workers. Source of Variationb 1‘ d.f. Mean Square F Entire Sample Sex 1 4.7 1.40 Treatment A.& B vs. C l 9.57 2 82 Treatment A vs. B 1 .00 -— Workers 1 .16 -- Workers by sex 1 .44 -— Sex within C 1 7.18 2.12 Sex by treatment 1 .54 -- Workers by treatment A & B 1 .57 -- Error 115 3-39 Males Treatment A.& B vs. C 1 16.67 5.46* Treatment A vs. B 1 .30 -- Workers 1 .56 -- Workers by treatment 1 2.72 -- Error 56 3.05 'Females Treatment A.& B vs. C l .09 -- Treatment A vs. B 1 .24 —- Workers 1 .03 -- Workers by treatment 1 7.37 1.99 Error 59 3.79 an .- 125 for entire sample; N = 61 for males; N = 64 for females b *P < .025 '-' 5.29 A = Personal contact B = Personalized Mail 0 = Brochure Table 9. The attendance of the males in the personal contact and person- alized mail groups exceeds (F = 5.46, P <:.025) the brochure groups. The mean attendances of all groups obtained from self-reports all exceed the mean attendances of groups obtained from the registration cards. A come parison permits an estimate of faking in self reports and it seems that the selfbreports contain a degree of exaggeration. 55 TABLE 9.-4Means of treatment groups for attendance at University Lutheran Church as determined by the questionnaire by sex and workers. Group Treatment Worker Male Female Al Personal contact Campus Pastor 7.00 5.30 A2 Personal contact Lay Campus Worker 6.10 7.83 BB Personalized7mail Campus Pastor 4.80 8.50 B4 Personalized mail Lay Campus worker 7.20 5.60 C Brochure None 1.71 6.48 Other Action Qgtg The analysis for attendance at the student's home church found no differences associated with treatments. The results are shown in Table 10 and the mean attendances of groups in Table 11. Male and female TABLE 10.--Analysis of variance for attendance at students' home church by treatment groups, sex, and workers. Source of Variationb d.f. Mean Square F Sex 1 .02 ~- Treatment A & B vs. C 1 1.10 -- Treatment A vs. B 1 .15 -- workers 1 6.39 3.12 workers by sex 1 3.45 1.68 Sex within C l 9.15 4.46* Sex by treatment 1 4.98 2.42 Wbrkers by treatment A & B 1 .04 -- Error 115 2.05 8‘N = 125 for entire sample hA = Personal contact B = Personalized mail C = Brochure *P < .05 = 3.92 brochure groups responded differently (F = 4.46, P <:.05) and it is possible to reject the hull hypothesis that there is no interaction between treatment and sex. 56 TABLE 11.-éMeans of treatment groups for attendance at students“ home church by sex and workers. Group Treatment worker Male ‘- Female A1 Personal contact Campus Pastor 6.80 4.62 A2 Personal contact Lay Campus worker 3.40 4.17 B3 Personalized mail Campus Pastor 5.20 6.50 .B4 Personalized mail Lay Campus worker 2.40 6.00 C Brochure None 7.67 3.76 v—v- It is interesting to note that the analysis of total church .attendance data found no differences associated with treatment. The results and means of the analysis are shown in Tables 12 and 13. TABLE l2.-Ana1ysis of variance for total church attendance by treatment groups, sex, and workers. Source of Variationb d.f. iMean Square F Sex 1 19.46 3.57 Treatment A & B vs. C 1 1.31 -- Treatment A vs. B 1 1.12 -- Nbrkers l .26 -- Hbrkers by sex 1 .19 -- Sex within C 1 9.22 1.69 Sex by treatment 1 1.40 -- workers by treatments A & B 1 1.67 -- Error 115 5.45 aN=125 b A = Personal contact B = Personalized mail C = Brochure TABLE 13.-€Means of treatment groups fer total church attendance by sex and workers. Group Treatment worker Male Female Al Personal contact Campus Pastor 13.00 12.23 A2 Personal contact Lay Campus werker 12.00 14.33 B3 Personalized mail Campus Pastor 11.80 16.75 B4 Personalized mail Lay Campus worker 10.40 15.60 C Brochure None 11.24 13.48 57 Likewise, no differences associated with treatment were found for the number of times communion was received. A11 female groups received communion more frequently than male groups (F = 4.64, P .05 by Fisher's exact test Treatment A vs. B: P :>.05 by Fisher's exact test bTreatment A.& B vs. C: P <1.05 by Fisher's exact test Treatment A vs. B: P <:.05 by Fisher's exact test initiated toward the campus pastor. Table 17 contains the frequencies TABLE 17.--Joint frequency of visits to a local church or student center by treatment groups. A l Treatment Group 0 or 1 visit 2 or more visits Personal contact (A) 20 24 Personalized mail (B) . 27 11 Brochure (C) 28 14 *Y eTreatment A a B vs. c: x2 = 1.02, P >.05 Treatment A vs. B: x2 = 5.46, P < .025 = 5.02 59 which are associated with visits to a local church or student center. A disproportionately larger number of students in the personal contact groups made visits to a local church or student center than the personal— ized mail groups (X2 = 5.46, Pdf .025 = 5.02). Attitude Data The assumptions and techniques of the Goodman--Kruska1 analysis were such that it was not possible to test the null hypotheses regard- ing treatment effects separately. Technically Speaking, positive findings in this test indicate that differences exist in the data large enough to be considered significant and in the assumed order. The first two null hypotheses were combined for the Goodman--Kruska1 analyses only into one as follows: Null hypotheses: Personal contact groups show no differences in religious involvement when compared to personalized mail groups and personalized mail groups show no differences when com— pared to the brochure groups. Table 18 contains the analysis of reaponses to the question concerning TABLE l8.--Goodman--Kruskal correlation for reaponses to the question about the extent of the conflict between science and religion. Treatment Group "Support "Negli- "Consid— "Very "Irrec- One gible" erable" Consid- concil— 2 Another" erable" able" Personal contact (A) 8 10 13 12 2 Personalized mail (B) 18 8 6 3 2 1.90* Brochure (0) 14 9 12 5 2 *P < .03 = 1.90 60 the extent of the conflict between science and religion. Differences between treatment groups were found (z = 1.90, P'< .03 = 1.88) and personal contact group frequencies were larger in desired categories than personalized mail groups and personalized mail groups outdid brochure groups. Other Goodman--Kruskal analyses are shown in Table 19 for TABLE 19.--Goodman—-Kruskal correlation for responses to the question asking what difference coming to believe there were no God would make in various opinions. .Treatment Group "Lots "Somewhat "No z different" different" difference" flips things a good husband or wife would do." Personal contact (A) 7 8 30 Personalized mail (B) 7 5 25 .85 Brochure (C) 8 10 24 "The way the government is run." Personal contact (A) 4 8 33 Personalized mail (B) 3 4 30 1.34 Brochure (C) 6 ll 25 "The way I should study and do my work." Personal contact (A) 1 9 35 .Personalized mail (B) 2 6 30 2.01* Brochure (C) 4 13 25 "The help I should give to relatives and friends." Personal contact (A) 3 7 35 Personalized mail (B) 5 10 22 3.39** Brochure (C) 8 14 19 "Suicide" Personal contact (A) 5 10 30 Personalized mail (B) 10 ll 16 4.13*** Brochure (C) 18 10 14 *P < .02 = 2.00 **P < .001 = 3.09 ***P < .0001: 3.72 61 responsesto the question, "Suppose you came to believe there were no God, would this make a difference in some of your opinions?" For three of the hypotheses tested, significant differences were found (P .05 Treatment A vs. B: P > .05 by Fisher's exact test bTreatment A 8: B vs. C: P > .05 by Fisher's exact test Treatment A vs. B: P < .05 by Fisher's exact test c:Treatment A & B vs. C: X2< 1, P > .05 Treatment A vs. B: P > .05 by Fisher's exact test dTreatment A 8: B vs. C: P > .05 by Fisher's exact test Treatment A vs. B: P > .05 by Fisher's exact test 63 personal contact female groups were compared to personalized mail female groups (P‘<.05). A diaprOportionately smaller number of females ranked the campus pastor above the median in the personal contact group than in the personalized mail group. No significant differences were feund among treatment groups for the Puritan Morality--Relativistic scale. The analysis is shown in Table 22 and the means are exhibited in Table 23. A high score on the TABLE 22.--Analysis of variance fer scores on the Puritan Morality-- Relativistic scale by treatment groups, sex, and workers. Source of‘Variationb d.f. Mean Square 'F= Sex 1 .018 -- Treatment A.& B vs. C 1 2.09 2.84 Treatment A vs. B l .002 -— workers 1 .340 -- workers by sex 1 .160 -- Sex within C 1 .059 -- Sex by treatment 1 .00002 -- workers by treatments A.& B 1 1.256 1.71 Error 113 .735 ELN=123 b A = Personal contact B = Personalized mail C = Brochure TABLE 23.--Means of treatment groups for scores on the Puritan Morality-- Relativistic scale. Group Treatment worker Male Female A1 Personal contact Campus Pastor 7.30 7.69 A2 Personal contact Lay Campus worker 8.90 8.50 B3 Personalized mail Campus Pastor 8.56 7.62 B4 Personalized mail Lay Campus worker 7.22 8.20 C Brochure None 6.66 7.05 scale indicated more choices associated with Puritan Mbrality. 64 The responses to an open-ended question inviting comments about the subjects' religious needs and how they had been met while at Michigan State University were classified and placed in appropriate categories. No statistical analysis was appropriate, but it can be observed that personal contact groups showed a tendency to make complimentary statements about the personnel of University Lutheran Church while brochure groups reported the greatest number of confessions of religious inactivity.‘ The tabulation of the reaponses is shown in Table 24. TABLE 24.--Joint frequency of types of essay responses about the adequacy of the campus ministry and fulfilling students' religious needs by treat- ment groups. Type of Comment Treatment Groups Personal Contact Personalized Mail Brochure A1 A2 B3 B4 C Favorable comment about University Lutheran Church or its staff. 9 2 l l 4 General, but positive comment. 0 2 3 3 7 Confession of religious inactivity 0 1 l 2 9 Mixed feelings toward religion 2 3 3 3 3 Criticism of question- naire. 2 0 0 0 0 General, but negative comment. 0 0 0 0 2 Critical of University Lutheran Church 0 1 1 1 0 The following items from the questionnaire in the attitude data section are of special interest in that they suggest that the personalized mail groups are lowest in response to treatment, although at the signifi- cant level in only one instance. It should be remembered in this 65 connection that the personalized mail groups were least responsive in the return of the questionnaire which was solicited by mail. It is possible that a saturation point for the effectiveness of personalized mail had been reached for certain purposes and that a sharpening of procedures to .measure this possibility might have been useful. Responses to the question about the degree of religious influence in the students' up- bringing were combined into "very" and "moderate" and "slight" and "none." Male personal contact groups exceeded male personalized mail groups (P«< .025) as presented ianahle 25. Responses to the question about how TABLE 25:1;Joint frequency of responses to the question about the degree of religious influence in students' upbringing by treatment groups. Treatment Group "Very" "Mbderate" "Slight" "None" Mesa Personal contact (A) 7 13 E 0 0 Personalized mail (B) 7 7 : 5 1 Brochure (C) 7 12 : 2 0 Females6 Personal contact (A) 10 12 E 3 0 Personalized mail (B) 8 6 j 2 l Brochure (C) 11 8 : 2 0 a‘l‘reatment A a B vs. c: P > .05 by Fisher's exact test Treatment A vs. B: P <1.025 by Fisher's exact test bTreatment A a B vs. c: P > .05 by Fisher's exact test Treatment A vs. B: P > .05 by Fisher's exact test desireable students thought it was to maintain campus workers and religious centers were compressed into "very" and'"desireable," "little," and "none“ and no significant differences were found for personal contact and personalized mail groups compared to the brochure groups (X2 = 2.78, P :>.05) nor for the personal contact groups compared to the brochure 66 groups (X2 = 2.74, P :>.05). The frequencies and the results are shown in Table 26. TABLE 26. --Joint frequency of responses to the question about maintaining religious workers by treatment groups.3 Treatment Group "Very" "Desireable" ""Little" "None" Personal contact (A) 17 § 18 10 0 Personalized mail (B) 8 g 21 4 5 Brochure (C) 19 E 16 5 2 aTreatment A a B vs. a, x2 - 2. '78, P > .05 Treatment A vs. B: =.2 74, P > .05 Pagticipation and_Belief Data No significant differences were found for any hypothesis tested for data associated with the participation and belief variables. A theoretical basis for the possibility that the deeper aspects of person- ality would not be affected by contact was found in the review of the literature. Although all the null hypotheses listed at the beginning of the chapter must be accepted, a selection of analyses and tables will be presented here as an example of the types of data which were collected. No statistical tests were performed for the means of books and articles on religious themes by treatment groups. The means of the number of pages and books read are presented in Table 27. A difficulty in the analysis was found because a number of persons reported that they did some reading, but did not specify the number of pages. In such instances one-half of the mean number of pages for the group was used as the appropriate number. If the data have any value, they suggest a large amount of interaction between treatment and sex. Personal contact may 67 TABLE 27.--Means of books and articles read on religious themes by treat- ment groups. Treatment Group Means of pages of articles Means of number of books Females Males Females Males Personal ' - ~ * contact (A) 22.1 4.4 .5 .1 Personalized mail (B) 16.2 7.0 .3 .3 Brochure (C) 3.7 12.4 .8 .6 actually reduced the religious literature reading of males while increas- ing that of females. A tabulation of the number of persons who read and did not read is shown in Table 28. Table 29 contains the frequencies of TABLE 28.--Joint frequency of reading books and articles on religious themes by treatment groups. w—_ Treatment Group Book'sa Articlesb Yes No Yes No Personal contact (A) 7 38 23 .22 Personalized mail (B) 7 31 22 15 Brochure (C) 13 29 24 17 6Treatment A 8: B vs. C: = 3.27, P > .05 Treatment A vs. B: x2 =< 1, P > .05 bTreatment A a B vs. cé =<1, P > .05 Treatment A vs. B: X = < l, P > .05 TABLE 29.--Joint frequency of responses to question about the Nature of Christ by treatment groups.8 Treatmenthroup "Divine" "Great "Mythical "None of Prophet" Figure" These" Personal contact (A) 28 6 2 9 Personalized mail (B) 22 5 0 11 Brochure (C) 28 3 0 ll axz = 4.67, P > .05 68 responses to the question about the Nature of Christ. The lack of any differences in reaponses given by treatment groups, let alone significant differences, should be noted. Tables 30 and 31 contain the statistical TABLE 30.--Ana1ysis of variance for religous beliefs test by treatment groups, sex, and workers. Source of Variationb 23.76 d.f Mean Square F Sex 1 .19 -- Treatment A & B vs. C l .55 -- Treatment A vs. B l 1.61 1.01 .Norkers l .29 -- Norkers by sex 1 1.37 -- Sex within C l .58 -- Sex by treatment 1 1.43 -- workers by treatment A.& B 1 2.73 1.72 ‘Error 115 1.59 3N = 125 bA.= Personal contact B = Personalized mail C = Brochure TABLE 31.--Means of treatment 'groups for religious beliefs test. Group Treatment worker Mble Female A1 Personal contact Campus Pastor 24.40 22.54 A2 Personal contact Lay Campus worker 23.20 26.83 B3 Personalized mail Campus Pastor 23.60 23.88 B4 Personalized mail Lay Campus worker 23.90 22.00 C quphure None 25.00 analysis and means of treatment groups for the religious beliefs test. The homogeneity of the treatment groups means, as in the responses to the question about the Nature of Christ, should be observed. imam It was possible to reject the null hypothesis that personal con- tact and personalized mail groups show no difference in religious 69 involvement when compared to the impersonal mail group.for the action variables (a) attendance by males at University Lutheran Church by regism traticn cards (P <1.025) and by self reports (P < .025), and (b; contacts initiated toward the lay worker by females (P <..05); and the attitude variable: (a) religion and science conflict (P <1.03), (b) effect of non-belief in God on Opinions about study and work (P < .02), helping others (P < .001), and suicide (P‘< .0001). and (c) eValuaticn of contact with lay worker {P <1.005). The null hypothesis that personal contact grotps show no differences § in religious involvement when compared to the personalized mail groups was rejected for the action variables (a) contacts initiated toward lay worker (P < .05) and (b) visits to a local church or student center (P‘< .025); and the attitude variable: (a) religion and science conflict (P'< .03), (b) effect of nonwbelief in God on opinions about study and work (P'< .02 , helping others (P