CHICAGO JOURNALS Change of Heart: A Test of Some Widely Held Theories About Religious Conversion Author(s): Max Heirich Source: The American Journal of Sociology, Vol. 83, No. 3 (Nov., 1977), pp. 653-680 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/2778148 Accessed: 09/04/2009 05:06 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=ucpress. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org. The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of Sociology. STOR http://www.jstor.org Change of Heart: A Test of Some Widely Held Theories about Religious Conversion1 Max Heirich University of Michigan Classical social science descriptions of religious conversion focus on psychological stress, previous socialization, and various forms of direct social influence. This paper uses data about converts to Catholic Pentecostalism and a control sample to question the conventional wisdom. It shows how data can be organized to support classical claims, then systematically watches evidence fall away when the argument is organized more carefully. The paper closes with a theoretical critique of conventional arguments and suggestions for an alternative set of questions. These probe the circumstances and procedures by which a sense of ultimate grounding is affirmed or changed, at both an individual and a social level. For at least a century a debate has raged between social scientists and religionists and among social scientists themselves about what is really going on when the phenomenon described as religious conversion occurs. Two quite different sets of issues have been involved in that debate. The first set of issues concerns the nature of religious conversion as a phenomenon, while the second set of issues involves an argument about its causes. Believers have developed arguments about the nature of divine-human encounter, while social scientists have proposed a range of social and psychological forces at work. Although more than 50 empirical studies of religious conversion have been reported in the social science literature, 11 wish to thank a number of persons and institutional sources for their help in the preparation of this study. A University of Michigan Rackham research grant, no. FRG-1474, paid for the cost of gathering the data on which this study is based and for much of the cost of preparing the data for analysis. A National Institute of Mental Health traineeship program provided two research apprentices, Frank Solomon and Jeffrey Leiter, who helped with computer analysis of the data. Michael I. Harrison and I jointly designed the questionnaire used here, basing it on one I had developed in studying the group a year previously. A number of questions used in that original survey and repeated here were suggested by sociologists John Lofland and Guy E. Swanson and by two members of the movement, Phillip Thibideau and Sister Mary Tinsley. Michael Harrison administered the questionnarie and supervised its coding. Daniel Ayers helped solve a number of computer problems that arose during the study, and Mary Scheuer proved to be an invaluable assistant in the operation of various computer programs and preparation of tables used in this paper. Robert Kahn, Emilie Schmeidler, Paul Siegel, and John Sonquist each provided suggestions which made the organization of data more illuminating; and anonymous reviewers from the American Journal of Sociology encouraged me to enlarge my earlier argument, in ways that I have found helpful. I take full responsibility for any inadequacies of the study but am grateful to these people for enriching it in the substantial ways that each has done. AJS Volume S3 Number 3 653 American Journal of Sociology evidence rarely has been organized with careful attention to the logic of the arguments being discussed. This paper offers a test of some social science arguments about the nature of religious conversion and suggests alternate ways to explore this kind of phenomenon. It compares Catholic Pentecostals who claim to have received the Baptism of the Holy Spirit and to have reoriented their lives around this encounter with the Divine with a control sample from similar backgrounds. Five questions are addressed: (1) To what extent are converts subject to the kinds of influence which social scientists have said account for religious conversion? (2) Is the presence of such influence coincidence? (When a larger population containing both converts and nonconverts is studied, are converts located disproportionately among those exposed to such influences?) (3) Just how important an influence on conversion are these social factors, singly or together? (How much of the total variance in outcome is explained by different kinds of social influence? How much is explained when all of them operate together?) (4) Under what circumstances do social influences have the most impact on susceptibility to reorienting religious experiences? (5) Are there more interesting questions to ask about this kind of phenomenon? If so, what are they and how might we proceed? In both religion and the social sciences one finds a wide range of usage for the term " con version." In some descriptions, conversion involves a dramatic turnabout—either accepting a belief system and behaviors strongly at odds with one's previous cognitive structure and actions or returning to a set of beliefs and commitments against which one has been strongly in rebellion. In other descriptions conversion involves a qualitative change in experience and in level of commitment, regardless of previous mindset (as, e.g., among Protestant Evangelicals who seek the conversion of children of the faithful, children who, while they may not be in active rebellion against the faith, have not yet shown the qualitative shift in experience which these believers take as the hallmark of true religious experience). Some Protestant sects, wishing to preserve the term "conversion" for an initial "conviction of sin and turn to Jesus," describe a second level of conversion, which they call "sanctification." This involves a dramatic, qualitative shift in understanding, commitment, and behavior which in some groups normally occurs after the first "conversion." (It typically involves speaking in unknown tongues and other practices akin to those found among Catholic Pentecostals.) New religious movements are particularly interesting phenomena in terms of this difference in understanding what religious conversion means. Such religious movements frequently recruit from among persons already oriented toward a religious quest (as also is true for the second level of 654 Religious Conversion religious conversion just described). Yet they frequently require a reorientation similar to the dramatic turnabout assumed in the first definition of conversion. Thus they fit the entire range of arguments being put forward. Our sample holds special interest for that discussion. At the time these data were gathered, Catholic Pentecostals represented a controversial position within the Catholic tradition, one involving a radical redefinition of past understandings of Christianity and Catholicism and a reaffirmation of past commitments to "Christ and His church." Since Pentecostalism had been a part of Catholic practice for only two years at the time these converts were interviewed, its status within the church was far from clear. The church watched with wary eye what was happening; a number of nuns and priests who worked actively with students interviewed in this study considered Pentecostalism a heresy. Many Catholics were shocked at Pentecostal goings-on within the church, and a number of devotees who were interviewed described their own astonishment on first encountering people within the Catholic church who claimed that God worked in this way. They told of the struggles they went through in deciding that God, and not some mass delusion, was at work. These people, moreover, often reported a number of radical changes in their own behavior—not only in terms of devotional practices, but also in places of abode, persons with whom they associated, and the like. Thus they seemed reasonably close to the first definition of "convert." It is true that they varied widely in the degree of rebellion they had expressed against the church and that many (though not all) of the rebels had begun a return before encountering Pentecostalism. But because "sanctification" (in the Protestant-sect sense) was not part of the conceptual framework of these Catholics, their Baptism in the Holy Spirit represented a more dramatic shift of position than is experienced by most second-generation "converts" among Protestant Evangelicals or by "sanctified" members of the Pentecostal sects who, after all, have been hoping that this would happen to them. For all the reasons just stated, this experience falls in the middle of the range of definitions of religious conversion found in the literature. Today, with the church's blessing and large numbers of "good Catholic" Pentecostals visible, joining the group requires a less extreme redefinition of self and is more akin to affiliating with a sectarian movement inside the church. But when these data were gathered, that was not true. Receiving the Baptism involved a radical redefinition of how God works, a major shift of behavior to accommodate that new understanding, and the acceptance of a certain amount of personal controversy by affiliation with such a definition of Christianity. A second set of issues under debate has concerned causes of religious conversion, in other words, circumstances which make individuals particu- 655 American Journal of Sociology larly susceptible to this kind of experience. Believers usually develop a series of causal arguments about how God works, whereas social scientists have tended to explain the phenomenon as a subset of psychological and/or social processes observable more generally. In the fairly extensive social science literature dealing with religious conversion, three rather different explanatory themes emerge. The first views conversion as a fantasy solution to stress, in which the threatening situation is dealt with either by making an alliance with supernatural forces that could change the power balance or by changing one's frame of reference so that previously distressing material no longer seems important. Much of the psychological literature on religion adopts this perspective. But it also underlies much macroanalysis of religion. Karl Marx's sense of religion as the opiate of the masses, H. Richard Niebuhr's analysis of origins of fervent religious movements among the socially disadvantaged, and a wide range of more recent studies, including Bryan Wilson's discussion of social conditions encouraging new religious movements, have made the same kind of argument but have located the stress in the social circumstances found among believers. Of all the social science arguments about religious conversion, this is the one most debunking of participants' claims and the one most likely to be resented by converts. Consequently it requires special care in measuring. A second explanatory theme sees the explanations of religious conversion less in the circumstances which produce the immediate result (e.g., stress) than in previous conditioning. Thus it looks for socialization circumstances that should leave one ripe for the plucking. It looks at parental orientations (especially as they affect eldest children, who are seen as more likely to identify with parental values), at sex-role education (since women tend to participate in religious activities more than men), and at the impact of schooling. A third and more recent explanatory theme in the social science literature focuses upon interactions that make a different understanding of one's experience possible. This kind of argument focuses upon circumstances that lead one to take a particular frame of reference seriously. Most typically, it involves analysis of patterns of interpersonal influence and what is sometimes called the process of encapsulation, whereby inputs from others become so mutually consistent and reinforcing that one begins to see things through the others' eyes.2 While many social scientists have focused their attention on a single "causal explanation," others have seen conversion as involving all three processes in interaction. For example, John Lofland, studying conversion to a deviant cult that believed Christ had returned to the earth in the 2 Examples of influential work that has presented one or another of these explanatory themes are listed in the references. 656 Religious Conversion Orient and that the end of the world was imminent, developed what he called a "value-added model" to show how all these influences work together to produce the end result. An article written with Rodney Stark describes the process they have observed: "For conversion, a person must 1) experience enduring, acutely felt tensions, 2) within a religious problem-solving perspective, 3) which leads him to define himself as a religious seeker, 4) encountering the [cult] at a turning point in his life, 5) wherein an affective bond is formed (or preexists) with one or more converts, 6) where extra-cult attachments are absent or neutralized, and 7) where if he is to become a deployable agent, he is exposed to intensive interaction" (Lofland and Stark 1965, p. 874). In my opinion, all of the foregoing explanations, whether taken singly or together, remain unsatisfying as a theoretical statement about what is occurring. They are too general an account, in that they ought to apply equally well to all forms of changed outlook or behavior, they do not explain why any particular perspective should be attractive to potential converts. Yet the explanations are not general enough, in that they do not exhaust the range of circumstances under which one might expect conversion to occur. Moreover, they largely ignore the nature of the phenomenon itself. Despite these shortcomings, it should be worthwhile to see how well they apply to concrete situations for which data are available. If they fit reasonably well, an enlargement of the argument to make it more theoretically complete would seem in order. If data do not reassure us that such components are a necessary part of an explanation of the phenomenon, we may want to start again to examine what is occurring. PREVIOUS EFFORTS AT PROOF Of the many empirical studies dealing with religious conversion in the United States and Britain since the turn of the century, few provide evidence in a form appropriate to the arguments which social scientists have generated. Between them, McKeefery (1949) and Argyle (1958) reviewed approximately 40 studies of religious conversion that were made in the first half of this century. Most were descriptive studies of odd clusters of converts, with little sense of how generalizable the description might be. Almost totally lacking, when cause-effect arguments were presented, was any examination of nonconverts to see whether they might be equally affected by the same influences. This lack of control groups made the meaning of observed correlations unclear. Unfortunately, the situation has not changed very noticeably in the two decades since they published their work. The same criticisms can be leveled against most of the empirical .studies produced even within the past 10 years. It may be worthwhile, however, to note three studies that have 657 American Journal of Sociology attempted to use control groups and five recent studies that have described participants in various expressions of the current Jesus Movement. Kildahl (1965) and Allison (1968) contrasted divinity school students who had experienced a sudden religious conversion with a control sample who had not had such experience. Using Rorschach tests, intelligence tests, and similar psychological-measurement instruments, they found little evidence of unusual or stressful personality styles among the converts. Allison found converts somewhat more likely to produce primitive emotional responses to Rorschach pictures but also more likely to have strong ego control of such material once it surfaced. Thus he asked whether conversion represented a method for handling potentially troubling emotional material for these individuals or whether, instead, their unusual ego strength had freed them to explore strongly emotional conversion situations. Both samples were small and somewhat arbitrary, so that findings are difficult to generalize. Kotre (1971) used a control sample to study what might be called the opposite of conversion—falling away from the Catholic church. He gathered 50 volunteers from the Catholic student centers at the University of Chicago and at Northwestern University and used a snowball sampling method to locate 50 friends of theirs who had similar social characteristics, including having been reared as devout Catholics, but who had since fallen away. Kotre found no significant differences in religious exposure during childhood, in reinforcement during late adoles-cense, or in social support or patterns of personal stress. Instead, he found that those who internalized loyalty to the church came from homes in which both parents were practicing Catholics and the relevant role-model parent maintained warm emotional ties with the child being studied. The nature of the sample makes generalization difficult, and its immediate extension to religious conversion is not clear; it suggests, however, that family socialization might prove a relevant variable for responsiveness to religious inputs. Streiker (1971), Harder, Richardson, and Simonds (1972), and Adams and Fox (1972) describe various groups within the Jesus Movement on the West Coast of the United States. Streiker, who provides a field observer's portrait of five organizations and their members, argues that the movement represents a "counter-counterculture/' a parody of youth styles of the late 1960s and a denial of the counterculture's most fundamental perspectives. He claims many adherents always had been fairly conservative, religiously oriented young people who often gave exaggerated testimonials about previous involvement in the drug and sex aspects of the counterculture, exaggerations stemming from their desire to have been serious sinners so that salvation would be more powerful. He does not attempt to present a statistical picture of the movement or to "explain" its adherents' openness to religious conversion of this type. The study by 658 Religious Conversion Harder et al. combines field observation with questionnaire data from a nonrepresentative sample of converts (74 men and 14 women attending a religious commune one summer) describing social-class backgrounds (upper middle), previous counterculture and social-movement experiences, political outlooks, and personality self-descriptions of respondents. The personality measures are compared with those obtained from college youth of similar age, but the meaning of observed differences is not clear. Adams and Fox present a statistical portrait, based upon a similar-sized sample. They conclude that conversion to the Jesus Movement provides an alternative for adolescents caught in a series of binds between parental pressures and peer expectations. They claim it preserves an absolutist morality of childhood, provides peer support and approval, and offers a resolution to adolescent sexual urges through denial of sexual feelings. "The Jesus trip, like drugs, appears to be used in such a way as to avoid coming to terms with the difficulties related to the identity crisis." Adams and Fox do not describe their basis for sampling or the nature of their probes concerning the process of conversion, nor do they attempt to compare the group they analyze with anything like a control sample in asserting cause-effect relations. These studies, like their predecessors, provide some interesting descriptions but do not test their conclusions in a very probing fashion. Two sociologists who have studied Catholic Pentecostals devote more attention to methodological issues and also to arguments about the nature of religious experience. Hine has published a series of papers (1969, 1970) comparing Catholic Pentecostals with three other groups that also recruit college students and demand major commitment of self: the student movement of the late 1960s, the black power movement, and the environmentalist movement. She points to similarities in mechanisms that promote commitment to the group's perspective. She also reviews a series of studies made of Pentecostals and neo-Pentecostals; these either show no relation between glossolalia and emotional disorders, as measured by a variety of standard tests, or indicate that Pentecostals score higher than others on a number of measures of psychological adjustment. Fichter's (1975) study of a national sample of Catholic Pentecostals demonstrates the appropriateness of describing this group as having undergone religious conversion. It then focuses upon the impact of this experience on belief structure and on involvement in the institutional church and in social and political issues. Fichter does not attempt to deal with previous arguments about causes of conversion, focusing, instead, upon the implications of such an experience for the ongoing social and religious life of the participants. In sum, previous empirical studies of religious conversion have differed widely in the care with which they documented the phenomenon in question, examined evidence for cause-effect arguments being put forward, 659 American Journal of Sociology concerned themselves with the generalizability of their findings, or used control groups. Some of the more sophisticated recent studies have not concerned themselves primarily with causal explanations for religious conversion but have taken conversion as a given. No single study exists that deals with the range of arguments that have been put forward over a series of decades. While this paper cannot provide that definitive test, it will offer more systematic examination of evidence relevant to these arguments and some suggestions about the direction that future research might take. NATURE OF THE DATA AVAILABLE ABOUT CATHOLIC PENTECOSTALS Over a seven-year period, with the help of a few graduate students interested in the same questions, I observed the emergence of Catholic Pentecostalism in the vicinity of Ann Arbor, Michigan. We gathered a variety of systematic data about it. In the spring of 1969, when the movement was still young enough to be a source of major controversy within the Catholic church but large enough in membership to make quantitative study practical, Michael Harrison and I gathered a snowball sample of Catholic Pentecostals in the immediate area. (We asked each respondent to name other members of the movement and tried to get a 100% universe of participants in a limited geographical vicinity, so that they could be compared with an appropriate control group.) We obtained 277 interviews with members of the movement, 67% of all persons identified through the snowball sampling procedure. Since the bulk of recruiting had occurred among Catholic-preference students at a large state university, we drew a control sample of such students, using lists gathered by that university and shared with the local Catholic student organization, the Newman Center. Names were filed alphabetically, so we used a table of random numbers to find a starting point in each drawer, then selected every fiftieth name thereafter. We acquired a control sample of 158 persons, 72% of those contacted for purposes of "control" comparisons. When the questionnaires were completed, we discovered that the movement had begun to recruit far beyond its original base. New members came from widely divergent backgrounds, and it was not clear how to enlarge the control sample to represent the new potential membership. Instead, I analyzed the data in a variety of ways, looking first only at converts whose backgrounds duplicated those of the control sample, then comparing these results with those obtained when the entire group of Pentecostal respondents was included. To my pleased surprise, the results were quite similar, regardless of sample size. For this discussion I will report patterns found when the definition of "Catholic Pentecostal convert" 660 Religious Conversion is limited to Catholics who reported receiving the Baptism of the Holy Spirit, who are then compared with the control sample of Catholics obtained "at random" from the university preference lists. With this cutting point, the total sample size becomes 310 persons, 152 converts to the movement and 158 Catholic respondents selected as a control group.3 Like all dynamic movements, Catholic Pentecostalism in this geographic area has continued to change in its patterns of outreach, its organizational style, and its respectability over time. The data used here are not intended to present a current portrait of the movement; instead, they focus upon a time at which conversion to its perspective involved both a fairly radical reorientation of religious understandings and a continuing religious quest for the majority of converts. The sample represents a point at which joining the movement often involved the full range of issues discussed by various students of religious conversion. 3 Between the time this study was designed and the time it was administered, the personnel attending the Catholic Pentecostal prayer meetings began to shift. High school students, Catholic seminarians, nuns and priests, and lay adults in the community began to join the university students who had previously made up the bulk of the membership. More ver, a number of persons who had been students at the university began to drop out, devoting themselves full time to "the service of the Lord." And Protestants, both Pentecostal and non-Pentecostal in background, were beginning to associate themselves with the group. Thus the neat convert-control sample of the study design no longer provided exact comparison groups. To meet this problem, I analyzed the data using three different definitions of "convert," to see what would happen to patterns of results if various members of the movement were included or excluded. The tightest comparison examined only converts who also were Catholic and students at the university from which the control sample had been gathered. This provided a logically tight comparison but produced a convert/nonconvert distribution that skewed the dependent variable in a way that affected the analyses reported in table 5 below. It also left out a large number of converts, raising the question of how typical of convert respondents this group was. Because of this, two other definitions of "converts" were used, with the same set of tables run each time. The second definition is the one reported in this paper: all Catholics who reported that they had received the Baptism of the Holy Spirit were defined as "converts" and compared with the control group. Finally, all persons gathered in the snowball sample were defined as "convert," on the hunch that seeking out the group might be as important a characteristic as whether the actual Baptism was claimed. This last group also had the advantage of including all persons for whom the experience had occurred; like the first sample, however, it resulted in a skewed distribution on the dependent variable, though in the opposite direction. Except for table 5 (where the skewed distribution of the dependent variable affected the R2 measure), the results were almost identical, regardless of which definition of "convert" was used. The main body of this report uses the intermediate definition for two reasons: first, the fact that converts and controls are almost equal in number leads to an intuitively clear sense of whether the influence being measured has a major impact on conversion; and second, the even distribution on the dependent variable allows clearer interpretation of the MCA results (see the section on "The Findings," below). I have prepared an appendix giving the equivalent tables when the other two definitions of "convert" are used and would be happy to share them with interested readers. 661 American Journal of Sociology CHOOSING INDICATORS FOR EACH ARGUMENT It has been possible to select a variety of indicators for each of the arguments expounded in the social science literature as a cause of religious conversion. The following paragraphs present the indicators and the reasons for choosing them. Stress.—I found four different ways of examining the impact of stress. First, the questionnaire included a variety of self-reports of stressful circumstances confronting one during the period preceding conversion: problems with members of one's family, with sex, with loneliness were probed in a variety of contexts and questions. Respondents also were asked whether they had perceived themselves to be in the midst of a "spiritual crisis" or whether they had sought counseling for personal problems during the time period. A "yes" answer to any of these probes was sufficient to credit that respondent as reporting actively felt psychological stress during the preceding two years. Second, I looked at exposure to stress-producing circumstances which the respondent might or might not have perceived as a personal problem. Wilson (1973) argues that religious movements are responses to particular kinds of socially disruptive circumstances. He describes seven types of religious movements, along with the social conditions that should encourage their development. The Catholic Pentecostal movement fits two of his seven categories, those he calls conversionist sects and thaumaturgical sects, and at least partially fits a third, which he calls introversionist sects. Conversionist sects seek a transformed self and are said to arise among populations dramatically detached from stable social contexts, so that individuals experience a highly atomized social life. Thaumaturgical sects, which emphasize special dispensations, miracles, and oracles, are believed to flourish where community organization and in-group authority structures have begun to collapse. Introversionist sects, which seek purification of self and a withdrawal from the world, occur in the early stages of the breakdown of self-sufficient communities. One could describe the Catholic church in the United States during the 1960s as facing all these circumstances. (They do not apply, however, to Episcopalians, Lutherans, and Presbyterians, among whom similar Pentecostal movements arose a few years earlier.) One could extend Wilson's argument to predict that converts to the Catholic Pentecostal movement should be previously devout churchgoers, who find themselves in atomized relationships in which previously stable senses of community and authority relations have disintegrated. For these socially generated forms of stress I chose a second set of indicators. Primarily these involve major role shifts—widowhood or divorce, decisions about leaving school or religious orders, contemplated marriage, a change of occupational plans. I also included newcomers to the university in this 662 Religious Conversion group, since they are more likely to have an atomized set of social relationships at that time than at other times. A third measure of possible stress grew out of discussions about the effects of birth order. Guy E. Swanson has argued that middle siblings may be especially inclined to seek fantasy solutions to stress. (The eldest child has power advantages in sibling struggles, he argues, and the youngest often has manipulative advantages.) Thus I looked to see whether middle children are more susceptible to conversion than others. Because the sample consisted of respondents from fairly comfortable income levels, I could not examine the impact of poverty on openness to religious conversion, other than to note that poverty obviously is not a necessary precondition. I did code parental backgrounds, however, to allow comparisons between persons coming from families in circumstances in which a college education is not traditional and those for whom there would be relatively little adjustment to college social strata. Socialization.—Similarly, four measures seemed appropriate indicators for testing the socialization argument. First, the social science literature argues that religious roles are more stressed in our society for women than for men. Second, it argues that parochial schools should reinforce a sense of religious awe and responsiveness to sacred symbols. Third, it argues that the oldest child more often identifies with parental values; consequently, I looked at eldest children of devout parents. Finally, with considerable hesitation about the legitimacy of this indicator, I examined current practices of personal piety and frequency of attendance at Mass. These could reflect previous conditioning or could be recent behaviors; they do, however, indicate an orientation to things religious prior to encountering the conversion opportunity. Immediate social influence.—From our firsthand observation of the movement's recruiting techniques, we were convinced that frequency of attendance at mass represented not only prior orientation but also accessibility to proselyters, since members of the movement attended Mass daily and there sought to interest other "serious Catholics" in their venture. Consequently, I used this item to measure exposure to social influence as well. In addition, I looked to see whether respondents were introduced to the movement by a trusted associate (e.g., a teacher, priest, nun, or spiritual adviser). I also looked for previous friendships with Catholic Pentecostals and for reports of reactions from close friends or relatives during the period of exploration. These four items indicated exposure to immediate social influence. (Encapsulation arguments will be examined more thoroughly later. This provides a first look, however, at whether immediate social influence seems to be important to the conversion process.) 663 American Journal of Sociology THE FINDINGS In order to compare this sample of Catholic Pentecostals with persons studied in earlier reports about religious converts, I first organized data in a manner similar to that reported elsewhere. Most studies, for example, lacked control groups for comparison purposes and thus simply reported how often a circumstance supposedly important for religious conversion was found among the converts being studied. When the present data are organized in that way, the results look strikingly similar to those reported elsewhere. For example, Catholic Pentecostals show high exposure to stress: 83% report having been under stress personally, and over 50% were involved in major role shifts. A surprisingly high proportion are middle siblings, and the same is true of persons from noncollege backgrounds. The socialization argument, however, came off less well, despite the fact that the turn to Pentecostalism tends to be a second-stage conversion for persons already religiously oriented: there is a moderate preponderance of female converts, and many converts report a high degree of previous piety or attendance at Mass; but relatively few report devout upbringing, parochial education, or being the eldest children in devout households. In contrast, immediate social influence appears to play a bigger role: except for the proportion of persons having previous friendships with Catholic Pentecostals, all measures of immediate influence received "yes" responses from almost two-thirds of the respondents. Looked at uncritically, these findings seem to fit the classic arguments rather well. Yet the results mean little. Until we know the extent to which the general population shares the same circumstances, we cannot assume that these factors account for the phenomena observed. When the distribution of the same factors among the control sample is examined, in fact, the picture becomes far less clear. Two-thirds of the controls also report personal stress (a proportion that might have gone higher had they been probed with as wide a variety of questions as were used on the converts themselves). A higher proportion of controls are involved in major role shifts, and they are at least as likely to be middle children or to be involved in upward social mobility. In terms of upbringing, converts and controls are similar, except that converts include a somewhat larger proportion of women. Converts, however, differ strikingly from the controls in reported religious practice (Mass attendance and personal piety) and in availability for social influence by the Pentecostals. Actually, the organization of evidence in this form is specious. To show that something influences the likelihood of conversion, one should group respondents in terms of the presence or absence of that influence on them. Then one should note the proportion of converts present within each group. One need not argue that the influence will lead everyone to convert. (There could be a variety of ways to respond to that influence.) Nor need 664 Religious Conversion TABLE 1 Percentage of Converts among Persons Exposed to Different Stress Situations Present (N) Absent (N) Somers's "D" Stress situation: Actively felt psychological stress... 55 (230) 33 (78) .2145 Potential stress: major role shifts. . 45 (190) 56 (120) -. 1081 "Style of coping" stress: middle sibling.......................... 55 (125) 45 (185) .0936 Social-mobility change............ 46 (116) 50 (180) -. 0487 Expected by chance if stress influence is irrelevant................... 49 49 .0000 Note.—Sample (N = 310) includes all "Spirit-baptized" Catholic Pentecostals + sample of Catholic university students. Table reads as follows: e.g., among the 230 respondents who reported actively felt psychological stress, 55% were converts. Among the 78 respondents who did not report this, 33% were converts. (Two respondents did not answer this set of questions.) If this factor accounted for all conversions, Somers's "D" would be 1.0000. If it were irrelevant, Somers's "D" would be approximately .0000. one argue that no one lacking that influence will convert. (There could be several routes to conversion.) But if the influence works as claimed, the proportion of converts should be noticeably higher when the influence is present than when it is absent. Tables 1-3 use the information just described but regroup it to organize the evidence in a form consistent with the arguments being made. Table 1, for example, examines the stress argument by grouping respondents according to the presence or absence of a particular indicator of stress. Since 49% of the respondents are converts, for any measure we could expect about half of the persons subject to its influence to be converts even if the measure is irrelevant to what is going on. A statistical measure, Somers's "D," shows how well each social influence predicts which respondents will prove to be "Spirit-baptized" Catholic.4 Once the data are organized in a form consistent with the argument being made the stress argument becomes far less convincing. The Somers's UD" scores hover around .00, and the only indicator which seems worth considering (actively felt personal stress) obtains its predictive value almost entirely from the fact that converts are less likely than others to deny all stress rather than more likely to report some stress. (Compare, e.g., the percentages and Somers's "D" scores for the first and third measure of table 1). Since converts were pressed harder than controls 4 Somers's "D" is peculiarly appropriate to the kind of sample involved in this study, requiring no statistical assumptions other than those met by our various sampling procedures. It allows a comparison of how much relative help different measures provide in predicting the likelihood of conversion. For obvious reasons, I have not reported significance-test scores. One part of the sample is selected from an attempted universe, while the other part is a sample in the normal sense. Thus the question of whether these differences would appear by chance is an odd one to use on these data. To me, the more crucial question is not whether these are real results but, rather, how important they are. For this question, Somers's "D" is the more appropriate measure. 665 American Journal of Sociology TABLE 2 Percentage of Converts among Persons Exposed to Different Socialization Circumstances Present (N) Absent (N) Somers's "D" Socialization circumstance: Sex: female..................... 58 (163) 40 (147) . 1753 Parochial education.............. 42 (84) 52 (226) -. 1032 Parents devout.................. 59 (99) 44 (209) . 1457 Oldest child from a religious family 55 (45) 47 (263) .0803 (?) High previous personal piety. . . 64 (166) 32 (143) .3229 (?) Frequent Massattendance. . ... S3 (115) 29 (193) .5446 Expected by chance if socialization influence is irrelevant............ 49 49 .0000 Note.—Samples (N = 310) includes all "Spirit-baptized" Catholic Pentecostals + Sample of Catholic University students. Table reads as follows: e.g., among the 163 female respondents, 58% were converts. In contrast, 40% of the 147 male respondents were converts. for information in these areas (because of the design of the interview instrument used), it is unclear that this reflects anything other than measurement technique. Even if it is taken at face value, the size of the Somers's "D" score leaves one aware that stress, at least as measured here, is insufficient to account for what is going on. Table 2 presents the data about effects of socialization on conversion to Catholic Pentecostalism. Again the results are unimpressive, except for two measures whose validity as indicators of prior socialization is not yet established. It is clear that current religious orientation, as reflected in measures of previous personal piety and frequency of attendance at Mass, makes a difference. It is not clear whether respondents exhibiting these behaviors come from backgrounds that should produce it. How clearly is Mass attendance related to prior socialization? A Multiple Classification Analysis was run to see how closely frequent Mass attendance relates to the childhood socialization influences just described, to childhood contacts with priests or nuns, to the various stress indicators seen previously, or to the mutual influence of all these factors working together.5 Findings were not encouraging for proponents of either the stress or childhood-socialization theories. For the persons studied, variation in how frequently one attended Mass was almost totally unrelated to any of the stress or childhood-influence measures described earlier. Taken together, they could account at most for 10% of the variation in Mass attendance found among these respondents. Thus I must conclude that, while Mass attendance here reflects current religious orientation, it does not necessarily result from previous conditioning. These findings point up the costs of previous failures to use control 5 For a description of MCA, see Andrews, Morgan, and Sonquist (1967). This is a statistical test similar to analysis of variance but more appropriate to the form of data available here. 666 Religious Conversion TABLE 3 Percentage of Converts among Persons Exposed to Different Kinds of Immediate Social Influence Present (N) Absent (N) Somers's "D" Social influence: Frequent attendance at Mass...... 83 (155) 29 (193) . 5446 Introduced to movement bv trusted associates.............'........ 75 (118) 31 (187) .4387 Previous friendships with Catholic Pentecostals................... 87 (63) 39 (247) .4779 Positive inputs from close friends or relatives during exploration..... 94 (108) 25 (202) .6974 Expected by chance if immediate social influence is irrelevant to conversion....................... 49 49 .0000 groups and to organize data in a form consistent with the argument that is being made. Table 3 shows the variables representing social influence arguments to be much more powerful predictors than what has gone before. In terms both of proportions of converts among those subject to immediate social influence and of Somers's "D" measures of how well this information predicts outcomes, the results look promising. It is clear that members of the movement, when recruiting, turn to previous friends and to persons they meet at daily Mass. It is also clear that introduction to Catholic Pentecostalism by a trusted person, together with positive inputs from others while exploring its claims produces fairly positive outcomes. This evidence, however, does not really deal with the heart of the social-influence argument. Few converts will be surprised to learn that a religious movement grows by the contacts it makes through friendship networks in which the converts participate. Nor will many observers be surprised to learn that people who receive positive feedback about a movement explore it more seriously than those who do not. This argument becomes interesting when it goes beyond those demonstrations to show how a person comes to shift his or her understanding of the world under the influence of others. Our data allow a first exploration of evidence concerning what social scientists call the encapsulation process, but the results are only suggestive. The snowball sampling method used to contact converts for this study involves a built-in bias against locating people who explored the movement and then rejected it. (They may be less likely to be remembered and reported by believers once they have dropped from sight.) Official reports of the movement during the period studied estimate that about one-fourth of the 1,089 recorded visitors to their meetings eventually sought the Baptism of the Holy Spirit (Word of God Community 1969, p. 3). Our 667 American Journal of Sociology snowball sample plus the control sample produced only 17 persons reporting extensive contact with the group who did not plan to receive the Baptism. An additional 18 persons described themselves as actively seeking this Baptism, and 25 others indicated contact with the group but no Baptism in the Spirit, leaving their "seeking" status unclear. Despite these sampling problems, we can determine to what extent an encapsulation process occurs before converts join the group. This, of course, will not prove that such a process always leads to conversion. If converts lack encapsulation experiences, however, it should be possible to reject encapsulation as an explanation, whether or not an adequate control group is available for comparison. After encountering the Pentecostals, about a third of the seekers turned to teachers, religious advisers, or other persons they trusted for reactions. Almost half turned to members of their family, and over four-fifths discussed the matter with close friends. For this feedback a large minority of seekers turned to, persons already positively disposed toward the claims of the movement. For example, 40% of the friends and advisers contacted were themselves Pentecostals. Among the respondents 70% remembered their close friends as giving either positive or neutral advice about the Pentecostal experience, and over half reported that family members responded similarly. Thus the vast majority of seekers were not receiving discordant information from family or friends (or at least were not remembering any). About a fifth of the converts reported that they spent less time with regular companions during the time they were exploring the movement but had not yet received the Baptism. All of the eventually baptized spent time during the exploration period with others who had received the Pentecostal signs. Fifty-six percent reported that the amount of time spent with persons who had received the Baptism increased noticeably during this period. About half this many reported that they spent more time with other seekers who had not yet received the Baptism. One final bit of evidence lends additional support to an encapsulation argument: 39% of the currently baptized reported that they now spend more time than previously with persons who are seeking but who have not yet received the Baptism of the Holy Spirit. Table 4 divides seekers into groups in which encapsulation might be occurring and groups for which there is little evidence of it. It is clear from this table that encapsulation aids conversion (92% of encapsulated seekers in the study received the Baptism, compared with 70% of seekers not under such influence). But the high proportion of converts among persons for whom encapsulation did not occur shows that, while it may be sufficient, it is not a necessary condition for conversion. Indeed only about a fifth of all seekers and less than a quarter of those who actually receive 668 Religious Conversion TABLE 4 Preliminary Test of Encapsulation Arguments Proportion of Converts among Seekers Yes (%) (N) No (%) (N) Somers's "D" Turned to already known Pentecos- tals for advice during exploration 100 (55) 75 (123) .2460 Had Pentecostal friends and began spending more time with them or else had no Pentecostal friends and began spending less time with former companions................ 86 (38) 78 (140) . 0865 Spent increased amount of time with other seekers................... 89 (39) 77 (139) .1100 Spent increased amount of time with baptized Pentecostals............ 88 (101) 69 (77) .1842 Note.—Summary: proportion o f converts by degree of encapsulation experienced by informant (as combination of the four measures): full = 92% (N = 36); partial = 85% (N = 81); none = 70% (N = 57); Somers's "D" = .1808. Expected by chance if encapsulation is irrelevant to conversion: full =78%; partial =78%; none = 78%; Somers's "D" = .0000. the Baptism reported encapsulation experiences. Not surprisingly under these circumstances, the Somers's "D" measure of how often one can predict conversion by knowing encapsulation status is rather low. This preliminary examination of the encapsulation argument suggests that it would be erroneous to assume that it is social influence rather than simply social contact that accounts for most of the conversions that occurred. HOW IMPORTANT ARE THESE INFLUENCES IN COMBINATION? Some of the more sophisticated social science arguments about religious conversion, of course, insist that it is not a single influence but the mutual interaction of various forces that makes a person susceptible to conversion. Do these social influences together have a stronger impact on conversion than any might by itself? When all is said and done, how much of the variance in conversion and nonconversion has actually been explained? Table 5 (an MCA test)6 suggests that social influences could account statistically for almost half of the findings about who is a convert. (The R2 score, estimating the proportion of variation statistically accounted for by the joint effects of all these social influences, is .43.) Yet the findings depend heavily on the impact of a single measure, frequency of previous 6 MCA was designed for samples of considerably greater size than we have, and it assumes a fairly even division of people on the dependent variable. (In our case, this would mean about the same number of converts and nonconverts, a condition met when "convert" is defined to include all Spirit-baptized Catholics but not met by the other definitions of "convert" discussed earlier.) It is not entirely clear how MCA scores are affected when either of these assumptions is violated. Consequently these results are not a definitive statement of how well the social science arguments have explained susceptibility to conversion. Instead, they give a first estimate of whether the arguments should be taken seriously at all in accounting for what has happened. 669 American Journal of Sociology TABLE 5 Comparative Explanatory Power of Various Kinds of Arguments (R2 = .43) Stress Socializa tion Mass Attendance Social Influence 2/32 scores..... Xß2/R2........ .02 .05 .01 .03 .21 .48 .07 .17 Mass attendance, which has a rather ambiguous relation to the social science arguments discussed earlier. (The ß2 scores in table 5 sum the contribution made by each set of social-influence measures—for stress, for socialization experiences, for immediate social influences, and for previous Mass attendance. In the second row of table 5, the ß2 scores are divided by R2, to suggest their relative importance vis-ä-vis each other in explaining outcomes. Here Mass attendance is almost twice as powerful an influence as are all the others combined.) Attendance at Mass offers a contact point for social influence to work. Given its lack of relation to the other social influences measured in the MCA test described previously, it is not clear that Mass attendance represents a distinctively social influence in itself. Therefore I ran an additional MCA test in which the Mass-attendance variable was omitted. The R2 score declined from .43 to .26. It is clear that the social influences measured here are at work. It is equally clear that they are not the primary explanation for what is happening. WHEN DO THESE SOCIAL INFLUENCES HAVE MOST IMPACT? How these social influences work to encourage conversion becomes intuitively clear if one examines figure 1, which presents the result of an Automatic Interaction Detection (AID) program used to prepare variables for the MCA. The AID is a sorting device that ignores all theoretical arguments being put forward; it simply divides a data set into ever smaller groups of respondents in terms of the single additional variable that provides the greatest contrast within the dependent variable. The AID continues this sorting process indefinitely, until it either runs out of cases or finds no additional variable that affects distribution of the dependent variable. It then prints out a "tree" showing the sequence of groups it has created and the mean score on the dependent variable for each of the groups shown. Its primary use is for discovering whether interaction affects the direction of the impact which any single variable has. The results in this case, however, become illuminating in their own right. The dependent variable here is con ver sion/noncon version, with converts 670 Religious Conversion CONVERSION LEAST LIKELY |(2)Attended mass weekly or less ofterj/ before encountering \ Jmovement ■■ 1.278_________ (3)Attended mass daily! Nor almost daily before! encountering movement f |y = 1.828__________ (4)No special source of [ Introduction y - i 141 1(5) Introduction] by a teacher, ]spiritual 'adviser, etc. [ý = 1.646 I(8)Only one /,parent devout = 1.607 ! (9)Both parents devout or neither parent devout y = 1.915 | (6)No Pentecostal! friends y = 1.129_________ (7)Pentecostal friends y = 1.500_______ (10)An only f an eldest childj | ý = 1.400 J(11)A middle or jyoungest child ~ 1.821 (12)No Pentecostal] friends [y = 1.522 !(13)Pentecostal I |friends 2.000 ______( |(16)Religious or political frames of reference used for (problem solving in recent pastl ' 1.222_________________ j(17)Psychological frame of rence used for problem (solving in recent past L.545 |(18)No Pentecostal ^friends (y - 1.666 s. 1(19) Pentecostal 'friends jy = 2.000_______ J (14) Female] n? "1.385 j \|(15)Male I (y = 1.700 [ CONVERSION MOST LIKELY Fig. 1.—AID (Automatic Interaction Detection program) tree showing proportion of respondents who became Pentecostals under differing combinations of circumstance. If no respondents in a group became Pentecostals, y = 1.0; if all became Pentecostals, y = 2.0. receiving a score of 2 and nonconverts a score of 1. Since there are almost equal numbers of converts and nonconverts, the mean scores for conversion (y) give an intuitively clear sense of what is happening. For the sample as a whole, the y score is 1.476. Among the 176 persons who attended Mass infrequently before encountering the Pentecostals, the y score is 1.278. (This is group 2, in which not quite 28% became converts.) In contrast, group 3, consisting of the 99 persons who were frequent Mass attenders before meeting the Pentecostals, has a y score of 1.828 (72% of them converted). In the AID tree, group 2 and group 3 are each progressively subdivided, so that one can see the impact created by the addition of each new cluster of circumstances. At the top of the tree one finds the persons least susceptible to conversion (group 6). These 124 persons did not attend Mass frequently before encountering the Pentecostals, were not introduced to the movement by a teacher or spiritual adviser, and had no close friends who were Pentecostals. Their y score is 1.129, for less than a quarter of them received the Baptism 671 American Journal of Sociology TABLE 6 N Respondents and SD Found within Each Group in AID Analysis Group N SD 275 .499 176 .448 99 .377 128 .348 48 .478 124 .335 4 .500 28 .488 71 .278 20 .490 28 .383 23 .152 5 .000 13 .487 10 .458 9 .416 11 .498 15 .471 13 .000 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. of the Holy Spirit. In contrast, group 19 has a y score of 2.0, showing that all who faced the following circumstances converted: not previously frequent in their attendance at Mass, they were introduced to the Pente-costals by a teacher or spiritual adviser, were middle or youngest children, and had Pentecostal friends. Where one or more of the influences is lacking, the proportion of converts declines. (See table 6 for the number of respondents and the standard deviation of each group in the tree). Social influences also have some impact on the group already embarked on a religious quest, as indicated by their previous frequency of attendance at Mass (group 3). But the contrast is much less pronounced for each of the groups created by the AID subdivisions of these original seekers. Such results suggest that the social influences studied here have greatest impact on persons not already involved in an active religious quest. They also suggest that immediate personal influences have more impact than does one's psychological state or prior socialization. SOME IMPLICATIONS OF THE FINDINGS Taken overall, however, these data suggest that the conventional wisdom about what is going on in religious conversion may need reexamination. Social factors can account statistically for enough of the variation in outcome to make us take them seriously. Yet for these converts, at least, these social influences do not seem to work in the ways our literature would suggest. 672 Religious Conversion The findings of this study look similar to those of much earlier research when data are organized in traditional but logically incorrect relation to the central arguments. However, when the argument is drawn more tightly, support drops away for theories that explain religious conversion entirely in terms of psychological stress or previous socialization. Moreover, the contribution of stress and socialization measures to an explanation that assumes a more sophisticated set of interactions remains quite small. The impact of social networks is striking indeed—for those already oriented toward a religious quest. But these arguments from the social sciences have been unable to account for the basic orientation involved. The AID analysis informs us about a combination of circumstances that seem to generate susceptibility to religious interest and conversion among those previously lacking it. But such a combination occurs infrequently among the persons studied here and seems inappropriate to the circumstances that had produced religious interest previously in the bulk of converts studied. It seems clear that the process of conversion occurs through use of available social networks. The means by which such use comes about, however, is not clear. If one is not already a religious seeker, such contact is insufficient in most cases to produce a "change of heart.'' Among those seekers studied here, network contact, coupled with social encapsulation, seems to guarantee the outcome. Yet conversion occurs often enough in the absence of encapsulation to show that this is not a necessary condition. In short, we can account for the route that conversion takes within a population; but we cannot, with these arguments, explain what lies behind the religious quest and response to it. It is not clear, of course, whether these findings would apply to previously studied groups of converts, although the similarity of our first descriptions to their reported findings is suggestive. Rather than urge a widespread replication of these testing methods in a variety of settings, however, I prefer to use these results to suggest that a more fundamental shift of analytic focus may be in order. I suspect that the inability of classic arguments from the social sciences to account statistically for religious conversion stems from a fundamental misconception of the process involved. Rather than argue about which truth is more accurate (i.e., sacred or social science conceptions of what has occurred), we might more fruitfully include both as complementary examples of a process that seems fundamental to human existence. I refer to the assertion of a sense of ultimate grounding—one that provides a clear basis for understanding reality, that provides meaning and orientation for understanding one's situation and acting in relation to it. The arguments as now formulated could apply equally well to many action choices; they tell us little about conversion, the process of changing 673 American Journal of Sociology a sense of root reality. As I see it, the interesting questions about conversion are not who will change (as socialization and stress arguments imply) or when they will change (as the arguments about immediate social influence imply). Rather, they are the following. 1. What circumstances destroy clarity about root reality (both for individuals and for collectivities) ? 2. How is an alternative sense of grounding asserted in ways that lead various observers to take it seriously? What ingredients must it have? What must it be able to do? How is it brought to their attention? Under what circumstances will an alternative sense of root reality (or grounding) become widely shared? An attempt to answer these questions would tie the study of conversion more closely to the growing body of literature that treats religion less as "systems of truth" than as efforts to discover a ground of being that orients and orders experience more generally. Conversion becomes especially interesting because it involves a conscious shift in one's sense of grounding. Whatever the outcome, it involves examination of core senses of reality, identifying aspects which must be responded to with the whole being and which presumably will affect action choices for the convert thereafter. Participants in many normal, ongoing activities take the ultimate sense of reality underlying their actions largely for granted. They may participate in activities for a variety of reasons and with a wide range of commitment to the underlying basis for that activity. New converts, in contrast, usually have made a major examination of the claims—or encounters—which underlie the symbols and activities involved. Whether or not we share their sense of what is ultimate, we can learn much about the process by which we ourselves acquire a sense of grounding. I will close this discussion by suggesting a theoretical stance that might prove useful for empirical studies of conversion. I will not try to formulate specific operational measures, nor will I suggest strategies for analyzing data that become available. These methodological issues deserve discussion in greater depth than would be appropriate at this time. Rather, using the questions formulated above, I would like to offer a few suggestions about how we might enlarge the way we think about problems of this kind before gathering empirical data. 1. What circumstances destroy clarity about root reality (for individuals or collectivities)? The conventional social science wisdom turns immediately to arguments concerning individual or collective stress, but religious tradition suggests a wider range of circumstances that might be at work. It sees a person's sense of ultimate reality shifting when one of two or possibly three circumstances arises. If experiences or encounters take place that cannot be encompassed within current explanatory schemes yet cannot be ignored, present understandings of root reality may come into question. 674 Religious Conversion Alternatively, when quite unacceptable outcomes appear imminent and inevitable, if current understandings of root reality are correct, many persons begin to reexamine their most basic assumptions. (One need not abandon one's sense of basic order to avoid undesired outcomes; but one might be more likely to question it under such circumstances and to look seriously at alternative claims that provide more options.) In the classic religious literature, experiences or encounters that shatter understandings from the past are referred to as mystical experience or religious ecstasy. The sense of impending doom that leads to reexamination of one's basic position is referred to as religious judgment.7 Many social scientists would not wish to ascribe a sacred source to such events, but this enlarged range of circumstances that destroys the hold of previous understandings opens new areas for exploration. One also might add a third circumstance encouraging reexamination of root reality: if respected leaders publicly abandon some part of past grounding assumptions, that step should either weaken their authority or encourage basic reexamination by others. The outcome should depend on the extent to which experiences underlying the leaders' shift of public position are shared by their followers. This reformulation of argument allows stress to play a role in explanation but forces a crisper depiction of its nature and how it should work. In addition to enlarging the range of factors that might be relevant to understanding what has happened, the perspective just presented suggests criteria for judging the potential value of explanations that are put forward: the new "reality" used by converts should speak directly to the problem they have encountered and should explain it more successfully than its earlier competitor. Thus the content of a new vision should provide one clear criterion for judging the theoretical value of a causal account of conversion. If the new vision does not offer a solution for the particular experiences that have broken the hold of past explanations, why should it be any more attractive than alternative possibilities for responding to stress? To be convincing, any causal argument should have to show links between content and experience. Mere correlational data would not in itself be convincing. 2. How is an alternative sense of grounding asserted in ways that lead observers to take it seriously? The socialization literature assumes that previous experience affects assessment of new claims. The encapsulation argument suggests that the range of perspectives used by close associates affects the persuasiveness of a new view of reality. An additional body of literature from the social sciences might be useful here, because it deals with groups of people who self-consciously shift their organizing assumptions. I refer to the history and philosophy of science and particularly to 7 For a discussion of these points, see Heirich (1976). 675 American Journal of Sociology the recent spate of writings concerning the adoption of "new paradigms" in science.8 This literature tends to emphasize two kinds of explanations of the acceptance of new claims. One involves a methodical application of common sense (as understood in western European-based cultures) in terms of the "scientific method"—a fairly simple set of procedures for assessing new claims in terms of past experience and reorganizing one's sense of past order in terms of new assumptions and social and experiential bases for accepting claims being put forward. The other kind of explanation concerns the social circumstances that generate doubt-producing encounters with old assumptions as a necessary part of carrying out standard procedures. It also examines the "political" processes by which doubts are held in check or given free reign and by which alternative grounding assumptions are hidden from view or made accessible to various potential publics. Many of these arguments deserve rewording in terms of religious-conversion arguments, for they suggest fairly simple and reasonable procedures by which people approach new phenomena and new kinds of claims about them. Yet we should remember that some senses of ultimate grounding assert an order that is unknowable and nonrational. The process involved in accepting such perspectives and the circumstances which encourage doing so may be far more intriguing than those involved in reordering a portion of one's grounding assumptions, as is the case with new-paradigm assumptions in science or in a shift from one version of the traditional religious world view to another. There are few suggestions in the literature about what this kind of radical shift involves; indeed most discussions assume that the process is a one-way flow moving from nonrationality to a commitment to rational perspectives. Many religious movements as well as 20th-century developments in the arts and other areas of intellectual life call this assumption into question. The problem is an intriguing one which deserves attention. An explanation of the spread of a new sense of grounding must include some sense not only of the circumstances that make its perspective viable to onlookers but also of the mechanism used to find adherents. The social science literature cited earlier and our own data as well suggest the importance of access points and networks for dissemination of new perspectives. The kinds of roles promoters of the new perspective occupy and the extent to which they reach accommodation or conflict with persons representing other organized interests within a population seem important for understanding how a viewpoint spreads or declines. Such detail becomes far more interesting, however, if set within a larger context that asks what 8 The most widely cited argument of this type is put forward by Kuhn (1970). 676 Religious Conversion difference various perspectives make in the range of options available to those who consider them. Converts may always be a small proportion of the population. Rather than treat them as deviants, to be explained in terms of variations from the status quo, we might learn more about both their experience and social processes generally if we approach them as offering a unique vantage point for examining the establishment and disestablishment of root senses of reality. In summary, religious conversion holds far more interest than most social scientists have allowed themselves to pursue. By treating it as an odd experience (rather than as one form of a fundamental human encounter) and attempting to explain it in social-psychological terms that ignore how its content relates to the structure of larger patterns of social interaction, we have neglected much of its potential for enriching our understanding of social life more generally. If more careful replications of past findings lead to their rejection, this may be a good thing: it may free us to formulate alternative questions that touch experience at a more basic level. REFERENCES Adams, Robert, and R. Fox. 1972. "Mainlining Jesus: The New Trip." Society 9 (February): 50-56. Allison, Joel. 1968. "Adaptive Regression and Intense Religious Experiences." Journal of Nervous and Mental Disease 145:452-63. (Reprinted [pp. 308-38] in Research in Religious Behavior, edited by Benjamin Beit-Hallahmi. Monterey, Calif.: Brooks-Cole, 1973.) Altus, William D. 1966. "Birth Order and Its Sequelae." Science 151 (January): 44^9. Ames, Michael M. 1957. "Reactions to Stress: A Comparative Study of Nativism." Davidson Journal of Anthropology 3 (Summer): 17-30. Andrews, Frank, James Morgan, and John Sonquist. 1967. Multiple Classification Analysis. Ann Arbor: Survey Research Center, Institute for Social Research, University of Michigan. Argyle, Michael. 1958. Religious Behavior. Glencoe, 111.: Free Press. Barber, Bernard. 1941a. "Acculturation and Messianic Movements." American Sociological Review 6 (October): 663-68. ---------. 19416. "A Socio-Cultural Interpretation of the Peyote Cult." American Anthropologist 43 (October-December): 673-75. Boisen, Anton T. 1955. Religion in Crisis and Custom. New York: Harper. Borhek, J. T. 1965. "Role Orientations and Organizational Stability." Human Organization 24 (Winter): 332-38. Cantril, H. 1941. The Psychology of Social Movements. London: Chapman & Hall. Clark, E. T. 1949. The Small Sects in America. Nashville, Tenn.: Abingdon-Cokesbury. Enroth, Ronald, Edward E. Ericson, Jr., and C. Breckenridge Peters. 1972. The Jesus People: Old Time Religion in the Age of Aquarius. Grand Rapids, Mich.: Eerdmans. Färber, I. E., Harry F. Harlow, and L. J. West. 1957. "Brainwashing, Conditioning and DDD." Sociometry 20 (December): 271-85. Festinger, Leon. 1955. "A Theory of Social Comparison Process." Human Relations 7 (2): 117-^0. 677 American Journal of Sociology Festinger, Leon, H. W. Rieksa, and Stanley Schachter. 1952. When Prophecy Fails. Minneapolis: University of Minnesota Press. Fichter, Joseph. 1975. The Catholic Cult of the Paraclete. New York: Sheed & Ward. Flowers, J. D. 1927. An Approach to the Psychology of Religion. London: Kegan Paul. Flügel, J. C. 1945. Man, Morals and Society. London: Duckworth. Ford, Josephine. 1970. The Pentecostal Experience. New York: Paulist. Freud, Sigmund. 1927. The Future of an Illusion. London: Hogarth. ---------. 1930. Civilization and Its Discontents. London: Hogarth. Gerlach, Luther P., and Virginia M. Hine. 1968. "Five Factors Crucial to the Growth and Spread of a Modern Religious Movement." Journal for the Scientific Study of Religion 7 (1): 23-40. ---------. 1970. People, Power and Change: Movements of Social Transformation. Indianapolis: Bobbs-Merrill. Gerrard, Nathan L. 1968. "The Serpent-handling Religions of West Virginia." Trans-Action 5 (May): 22-28. Gerrard, Nathan L., and Louise B. Gerrard. 1966. "Scrabble Creek Folk: Mental Health, Part II." Mimeographed. Charleston, W.Va.: Department of Sociology, Morris Harvey College. Goodman, Felicitas D. 1972. Speaking in Tongues: A Cross-cultural Study of Glosso- lalia. Chicago: University of Chicago Press. Goody, Jack. 1957. "Anomie in Ashanti." Africa 27:356-62. Gordon, David F. 1974. "The Jesus People." Urban Life and Culture 3 (July): 159-79. Greeley, Andrew. 1972. The Denominational Society. Glenview, 111.: Scott, Foresman. Harder, Mary, James T. Richardson, and Robert B. Simonds. 1972. "Jesus People." Psychology Today 6 (December): 45. Harrison, Michael I. 1971. "The Organization of Commitment in the Catholic Pentecostal Movement." Ph.D. dissertation, University of Michigan. ---------. 1972. "Sources of Recruitment to Catholic Pentecostalism." Journal for the Scientific Study of Religion 13 (1): 48-64. Hawley, Florence. 1948. "The Kersean Holy Rollers: An Adaptation to American Individualism." Social Forces 26 (September): 272-80. Heirich, Max. 1976. "Cultural Breakthroughs." American Behavioral Scientist 19 (July-August): 685-702. Hine, Virginia H. 1969. "Pentecostal Glossolalia: Toward a Functional Interpretation." Journal for the Scientific Study of Religion 8 (2): 211-26. ---------. 1970. "Bridge Burners: Commitment and Participation in a Religious Movement." Sociological Analysis 31 (Summer): 61-66. Hodge, William H. 1964. "Navajo Pentecostalism." Anthropological Quarterly 37 (July): 79-93. Kammeyer, Kenneth. 1967. "Birth Order as a Research Variable." Social Forces 46 (September): 71-80. Kelsey, Morton T. 1968. Tongue Speaking: An Experiment in Spiritual Experience. Garden City, N.Y.: Doubleday. Kildahl, J. P. 1965. "The Personalities of Sudden Religious Converts." Pastoral Psychology 16 (156): 37-45. Kotre, John N. 1971. The View from the Border: A Social-Psychological Study of Current Catholicism. Chicago: Aldine-Atherton. Kuhn, Thomas. 1970. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lanternari, Vittorio. 1963. Religions of the Oppressed: A Study of Modern Messianic Cults. New York: Knopf. Lenski, Gerhard. 1963. The Religious Factor: A Sociologist's Inquiry. Garden City, N.Y.: Doubleday. Leuba, J. H. 1925. The Psychology of Religious Mysticism. London: Kegan Paul. Lifton, Robert J. 1961. Thought Reform. New York: Norton. 678 Religious Conversion Lofland, John. 1963. Doomsday Cult: A Study in Conversion, Proselytization, and Maintenance of Faith. Englewood Cliffs, N.J.: Prentice-Hall. ---------. 1970. "The Youth Ghetto." Pp. 756-78 in The Logic of Social Hierarchies, edited by Edward O. Laumann, Paul M. Siegel, and Robert W. Hodge. Chicago; Markham. Lofland, John, and Lynette Lofland. 1969. Deviance and Identity. Englewood Cliffs, N.J.: Prentice-Hall. Lofland, John, and Rodney Stark. 1965. "Becoming a World Saver: A Theory of Conversion to a Deviant Perspective." American Sociological Review 30 (December) : 862-75. Lynd, Robert S., and Helen M. Lynd. 1937. Middletown in Transition: A Study in Cultural Conflicts. New York: Harcourt, Brace & World. McGuire, Meredith. 1974. "An Interpretative Comparison of Elements of the Pentecostal and Underground Church Movements in American Catholicism." Sociological Analysis 35 (Spring): 57. ---------. 1975. "Toward a Sociological Interpretation of the Catholic Pentecostal Movement." Review of Religious Research 16 (Winter): 94-104. McKeefery, W. J. 1949. "A Critical Analysis of Quantitative Studies of Religious Awakening." Ph.D. dissertation, Union Theological Seminary and Columbia Teachers College. Marx, Karl. 1959. "Toward a Critique of Hegel's Philosophy of Right." Pp. 262-66 in Marx and Engels: Basic Writings on Politics and Philosophy, edited by Lewis S. Feuer. Garden City, N.Y.: Doubleday-Anchor. Mayer-Gross, W., E. Slater, and M. Roth. 1954. Clinical Psychiatry. London: Cassell. Niebuhr, H. Richard. (1929) 1954. The Social Sources of Denominationalism. Hamden, Conn.: Shoestring. Nolan, James. 1971. "Jesus Now: Hogwash and Holy Water." Ramparts Magazine 10 (August): 20-26. Oates, Wayne E. 1967. "A Socio-Psychological Study of Glossolalia." Pp. 76-99 in Glossolalia, edited by Frank Staff, E. Glenn Hinson, and Wayne E. Oates. New York: Abingdon. O'Connor, Edward. 1971. The Pentecostal Movement in the Catholic Church. Notre Dame, Ind.: Ave Maria. Parsons, Anne. 1965. "The Pentecostal Immigrants." Journal for the Scientific Study of Religion 4 (no. 2): 183-97. Pasley, Virginia. 1955. 21 Stayed. New York: Farrar, Strauss & Cudahy. Pfister, O. 1948. Christianity and Fear. New York: Macmillan. Ranaghan, Kevin, and D. Ranaghan. 1967. Catholic Pentecostals. New York: Paulist. ---------. 1971. As the Spirit Leads Us. New York: Paulist. Read, W. R. 1965. New Patterns of Church Growth in Brazil. Grand Rapids, Mich.: Eerdmans. Richardson, James T. 1973. "Psychological Interpretations of Glossolalia: A Reexamination of Research." Journal for the Scientific Study of Religion 12 (June): 199-207. Saler, Benson. 1965. "Religious Conversion and Self-Aggrandizement: A Guatemalan Case." Practical Anthropology 12 (May-June): 107-14. Samarin, William J. 1973. Tongues of Men and Angels. New York: Macmillan. Sargent, William. 1957. Battle for the Mind: A Physiology of Conversion and Brainwashing. Baltimore: Penguin. Schein, Edgar. 1962. Coercive Persuasion. New York: Norton. Sonquist, John A., and James N. Morgan. 1964. The Detection of Interaction Effects. Survey Research Center Monograph no. 35. Ann Arbor: Institute for Social Research, University of Michigan. Starbuck, Edwin Diller. 1912. The Psychology of Religion. New York: Scribners. Streiker, Lowell D. 1971. The Jesus Trip: Advent of the Jesus Freaks. Nashville, Tenn.: Abingdon. 679 American Journal of Sociology Sutton-Smith, Brian, John M. Roberts, and B. G. Rosenberg. 1964. "Sibling Associations and Role Involvement." Merrill-Palmer Quarterly 10 (January). Thouless, R. H. 1924. An Introduction to the Psychology of Religion. Cambridge: Cambridge University Press. Travisano, R. 1970. "Alienation and Conversion as Qualitatively Different Transformations." Pp. 594-606 in Social Psychology through Symbolic Interaction, edited by G. P. Stone and M. Faberman. Waltham, Mass.: Ginn-Blaisdell. Vidich, Arthur J., and Joseph Bensman. 1960. Small Town in Mass Society. Garden City, N.Y.: Doubleday. Wallace, Ruth. 1968. "Some Social Determinants of Change of Religious Affiliation." Ph.D. dissertation, University of California, Berkeley. Wilson, Bryan. 1973. Magic and the Millennium. London: Heinemann. Word of God Community. 1969. "A Pastoral Report on the Catholic Pentecostal Community in Ann Arbor." Mimeographed (April). Ann Arbor, Mich.: Word of God Community. Yinger, Milton. 1963. "Religion and Social Change: Functions and Dysfunctions of Sects and Cults among the Disprivileged." Review of Religious Research 4:65-84. Zetterberg, Hans. 1952. "Religious Conversion as a Change of Social Roles." Sociology and Social Research 36 (January-February): 159-66. 680