An Investigation of a Reputedly Psychologically Abusive Group That Targets College Students

Michael D. Langone, Ph.D.

Executive Director, AFF

Editor, Cultic Studies Journal

Acknowledgments

This report was originally prepared for Boston University’s Danielsen Institute.  I wish to thank the Danielsen Institute, which made this study possible through honoring me as the 1995 Albert V. Danielsen Visiting Scholar.  Carole Bohn, Ed.D., Director of the Danielsen Institute, and Arthur Dole, Ph.D., Emeritus Professor of the University of Pennsylvania's Graduate School of Education, provided helpful suggestions in the original formulation of the study.  Ann Kelley, Ph.D., Assistant Director for Finance and Development of the Danielsen Institute, graciously assisted the author in a myriad of administrative details.  Dr. William Chambers, then Assistant Professor of Psychology at Mercer University's University College, analyzed the data from Component I and provided much helpful advice on the design and analysis of the study.  Dr. Steven Lynn, then Professor of Psychology at Ohio University and now Professor of Psychology at SUNY Binghamton, and then Ohio University doctoral students, Drs. Jodi Aronoff, Peter Malinoski, and Nataliya Zelikovsky, analyzed the data from Component II and, with Dr. Paul Martin, Director of Wellspring Retreat and Resource Center, had primary responsibility for the development of the test battery used in Component II.  Dr. Peter Malinoski also assisted in aspects of report writing.  Carol Giambalvo was of invaluable assistance in the development of the DDD Scale, one of the measures used in this study.  Rev. Robert Watts Thornburg, Dean of Boston University's Marsh Chapel, Rev. Harold Bussell, then Senior Pastor of the First Congregational Church of Hamilton, Massachusetts, Jeff Davis, Rev. Douglas Whallon, then New England Director of InterVarsity Christian Fellowship (IV), and IV staff members, Ming Wei, Colin Tomikawa, and Rich Lamb, helped recruit subjects.  Leanne Pellegrini, Blair Smith, and Melissa Kelley assisted in the administration of test batteries.  A special gratitude is owed the subjects who volunteered for this study; I deeply appreciate the time they gave to this research.  I also wish to thank Herbert Rosedale, Esq., President of AFF, for his continuing support through this project, and my wife, Donna, and children, Jose and Ana, for bearing with me during the weeks in which I was away.

26 April 1996 (revised 7 November 2001)

Contents

Introduction

Research Questions

Literature Review

Methods

Design/Procedures

Subjects

Instruments

Results

Discussion

References

Abstract

This study investigated former members of the International Churches of Christ (formerly and still often referred to as the Boston Church of Christ [BCC] or the Boston Movement) with regard to the nature and level of their psychological distress and their perceptions regarding the psychological abusiveness of the group.  The BCC is one of the most controversial groups on college campuses, and is often considered one of the fastest growing "cultic" groups in the world.  Psychological abusiveness was measured with the Group Psychological Abuse Scale and the DDD Scale, which inquires into concrete behaviors and practices thought to characterize the Boston Movement.  Psychological distress was measured by a battery of self-report instruments, including the SCL-90R, the Beck Depression Inventory, the Dissociative Experiences Scale, the Impact of Events Scale, and the State-Trait Anxiety and Anger Inventories.  Psychological background variables, such as a history of child sexual abuse, were also investigated.  The study consisted of two components.  In Component One, the GPA and DDD were mailed with a demographic questionnaire to 228 former Boston Movement members throughout the U.S., of whom 40 responded.  In Component Two, the psychological test battery, including the GPA but not the DDD, was administered to subjects face-to-face in the Boston area.  Component Two included 15 former members of the Boston Movement and two comparison groups, 23 graduates of InterVarsity Christian Fellowship, a mainstream campus ministry, and 19 former Roman Catholics.  Former Boston movement subjects rated their group significantly more abusive than did former Catholics or InterVarsity graduates and scored significantly higher on most measures of psychological distress.

Various groups, usually referred to as "cults," have generated considerable controversy during the past two decades (see CQ Researcher, May 7, 1993), in large part because their critics tend to believe that cults have psychologically abusive environments that harm members.  Although cults are usually thought of as religious, mental health professionals who have worked with cultists emphasize that they may also be psychological, political, or even commercial (Singer, 1987).  Some religious organizations define "cult" theologically, but mental health professionals tend to use psychological definitions.  According to this secular view, "cults" are distinguished from "new religions," "new political movements," innovative psychotherapies," and other "new" groups in that cults are characterized by extensive use of unethically manipulative techniques of persuasion and control to advance the leader's goals, often to the psychological and economic detriment of his or her followers (Langone, 1993).  It is important to note that, consistent with contemporary usage, this approach accepts the pejorative connotation "cult" has taken on in recent decades and, consequently, uses other words (e.g., "new religious movement") to describe groups, which in the past would have been deemed benign cults.

A recent factor analysis of 112 group characteristics rated by 308 former members of 101 groups that subjects deemed cultic produced a four-factor model of the varieties of psychological abuse:  Compliance, Exploitation, Mind Control, and Anxious Dependency (Chambers, Langone, Dole, & Grice, 1994).  The following empirical definition, which is consistent with clinical views, emerged from the factor analysis:

Cults are groups that often exploit members psychologically and/or financially, typically by making members comply with leadership's demands through certain types of psychological manipulation, popularly called mind control, and through the inculcation of deep-seated anxious dependency on the group and its leaders.  (Chambers, et al., 1994)

Approximately two to five million Americans have been involved with cultic groups.  This estimate derives from surveys of new religious and para-religious movements in San Francisco and Montreal (Bird & Reimer, 1982), high school students in the San Francisco Bay area (Zimbardo & Hartley, 1985), a weekly omnibus survey conducted by ICR Survey Research Group for AFF in 1993, and a Pennsylvania Medical Society survey of 1,396 primary care physicians, 2.2% of whom reported having had a family member involved with a cultic group (Lottick, 1993).

Among the problems that clinicians have observed in their work with former cultists are depression, loneliness, indecisiveness, slipping into altered states, blurring of mental acuity, uncritical passivity, guilt, fear of cult reprisals and supernatural personal catastrophes, an acute sensitivity to the "watchfulness" of parents and friends, difficulty explaining how they could have joined such a group, apprehension about their own idealism and altruism (which the cult had manipulated), a loss of the feeling of being a member of an elite, and financial difficulties (Singer, 1979).

Among groups commonly deemed cultic, the Boston Church of Christ (BCC) has been one of the most controversial (Ostling, May 18, 1992; Hill, Feb. 19, 1988).  The BCC has been especially controversial on college campuses, including Boston University.  A report by Boston University's Rev. Robert Watts Thornburg (Thornburg, 1989 [reprinted with modifications in Giambalvo & Rosedale, 1996]) provides useful background information on the BCC and a description of the psychologically abusive practices that have made it controversial.  Thornburg said that the Boston Church of Christ aroused so much concern at Boston University that for the "first time in the history of the University, leaders of a religious group have been formally and officially banned from coming onto the campus" (p.5).  He further states:  "At the September 1987 meeting of the Religious Life Council of Boston University, it was concluded by those present that the leadership of the Boston Church, despite their claims of innocence, were either unwilling or unable to change their procedures.  Our concerns can be grouped under several major categories of destructive activities" (p.6):

·         "Recruitment techniques include the duplicitous use of love and high pressure harassment, producing incredibly high levels of false guilt" (p.6).

·         "The training process is a virtual cloning of one person by their prayer partner in a totally authoritarian relationship with no rights to personal choice or interpretation" (p.8).

·         "In methods (classic examples of mind control techniques) the BCC eliminates non-member association for new prospects, systematically cutting out any contacts with family, friends, or outside sources of reality checks" (p.9).

·         "Once established as the sole dispenser of salvation, this group then dominates every moment of the day, demanding attendance at every meeting of the organization" (p.10).

·         "When total submission is acknowledged in a re-baptism, then the standard of faithfulness is measured by the number of recruits which each member can bring into the organization" (p.12).

Other universities have also confronted problems with the BCC.  American University and George Washington University, for example, recently revoked the charters of student groups founded by church members, while Howard University investigated allegedly aggressive recruiting tactics by group members (Masters, April 24, 1994).  Ronald Loomis, past President of the Association of College Unions International (ACUI), has identified 24 campuses, which have banned the Boston Movement or denied it access for violations of campus regulations.  He has also identified 15 different front names which the group uses to mask its real identity (personal communication, February 21, 1996).

The problems associated with the BCC are tied to its method of discipleship (abusive discipleship, according to critics).  Flavil Yeakley, a member of the fundamentalist Church of Christ denomination from which the BCC broke off and formerly the Director of the Church Growth Institute at Abilene Christian University, said that the Boston Church of Christ was "the leading congregation among the discipling churches" (Yeakley, 1988, p.5).  There is no reason to believe that this is no longer the case.  One of the most controversial aspects of the discipling system is the discipler's giving "detailed personal guidance to the Christian being discipled.  This guidance may include instructions concerning many personal matters of a totally secular nature" (Yeakley, 1988, p.1).  When carried to extremes the discipling system insists "on changing people at all costs -- even at the cost of their personhood, autonomy, and uniqueness" (Yeakley, 1988, p.19).

Yeakley, Thornburg, and other critics contend that the highly manipulative and exploitative environments of cultic groups harm many, if not most, members and former members of these groups.  Although there is a wealth of clinical experience with former cult members (Langone, 1993, says that the contributors to his book, Recovery from Cults, have worked collectively "with at least 9,000 cultists and their families" -- p.2), the empirical literature is inconclusive.  In part, this is due to methodological shortcomings in the research.  First, there is no acceptable operational definition of "cult"; indeed, this reviewer knows of only one study that even attempted to measure the "cultishness" of a group (Adams, 1993).  Second, samples are biased, either because the subjects are selected with the cooperation of group leaders (whose motives and trustworthiness may be suspect) or come from the network of former cult members who have sought help from cult educational organizations such as the American Family Foundation (AFF) or the Cult Awareness Network (CAN).  Third, dependent measures are often not standardized and are unconnected to other research in this area.  Fourth, few studies have used a comparison group.  Fifth, because practical necessity demands that studies of harm in cultic groups be retrospective, it is extremely difficult to demonstrate causal connections.

Research Questions

This study attempted to at least partly address these methodological issues.  The study examined:  (1) former members' perception of psychological abusiveness in the group environment, and (2) former members' reports of psychological distress.  The study focused on the Boston Church of Christ movement (BCC) because it is so controversial.  Unlike other research in this area (see the next section for a literature review), this study employed a sample that does not derive largely from the network of organizations dedicated to helping former cult members and their families, used objective measures that assess the abusiveness of the group environment, and included two comparison groups:  former members of a mainstream religious denomination (Roman Catholic) and graduates of InterVarsity Christian Fellowship, a mainstream campus ministry.  (Former Roman Catholics were recruited to serve as the mainstream denomination comparison group because the Boston area has a large number of Catholics.)  The following primary questions were investigated:

·         To what extent if any do former members of the Boston Church of Christ perceive their group to be psychologically abusive (as measured by the Group Psychological Abuse Scale (Chambers et al., 1994) and how do their perceptions compare to those of InterVarsity Christian Fellowship graduates and former members of a mainstream denomination (Roman Catholics)?

·         To what extent do former members of the BCC report having had the types of specific, concrete experiences for which the BCC is often criticized -- to be measured by the DDD Scale (a measure of concrete behaviors and experiences thought to reflect deception, dependency, and dread)?

·         Do former members of the Boston Church of Christ report more psychological distress (as measured by a psychological test battery) than former members/graduates of InterVarsity Christian Fellowship or a mainstream denomination (Roman Catholics)?

The InterVarsity sample permits comparisons with a group whose former members ("graduates" is a more appropriate term) are not expected to have been unhappy with the group.  The former mainstream sample permits comparisons with people who were probably unhappy with various aspects of their former religious denomination.  The latter sample helps test the hypothesis that negative ratings toward the BCC may result primarily from ex-members' disaffection, rather than their at least partly objective evaluation of a group that is indeed more abusive than mainstream groups.

This study focused on former members of religious groups because of (a) resource limitations; (b) problems associated with attempts to secure the cooperation of current members of controversial groups; and (c) the desirability of maximizing the continuity between this study and current and past research.  In the future, however, the investigator and his colleagues also intend to examine current members of controversial and mainstream groups.

The study also examined several secondary issues.

Pre-group psychological distress.  Some have suggested that post-cult psychological distress merely reflects long-standing psychological problems that pre-date the group involvement (Maleson, 1981; Spero, 1982).  As-yet-unreported data from the research of Martin et al. (1992) found, to the investigators' surprise, an inverse relationship between pre-cult psychological distress and post-cult distress.  Pre-cult distress, however, was measured by the occurrence or nonoccurrence of pre-cult psychological counseling.  An alternate explanation of this surprising finding might be that those who had been counseled developed coping skills that enabled them to handle post-cult distress better than those who had not been counseled.  This study and a related study at Ohio University are the first to use standardized instruments to assess pre-cult psychological distress, specifically the Physical Child Victimization Scale (Briere & Runtz, 1988), the Psychological Child Maltreatment Scale (Briere & Runtz, 1988), and the Childhood Sexual Victimization Questionnaire (Finkelhor, 1979).

In-group help seeking.  Most clinical observers believe that the isolationism and elitism of cultic groups would discourage members from seeking professional mental health assistance.  This hypothesis will be tested by asking subjects if they sought help while participating in their group.

Post-group help seeking.  Another, although obviously limited, measure of post-cult psychological distress is the decision to seek professional help.  The background questionnaire asks about this issue.

Family background.  Several studies (Sirkin & Grellong, 1988; Wright & Piper, 1986; Marcus & Grellong, submitted for publication to Cultic Studies Journal) suggest that cultists may be more likely to have dysfunctional family backgrounds, although other studies (Maron, 1988) and clinical observations (Singer, 1986; Clark et al., 1981) suggest that cult members' families are not likely to differ from those of noncult members.  In this study (as with Sirkin & Grellong, Marcus & Grellong, and Maron), the Family Environment Scale (Moos & Moos, 1981) assesses subjects' family backgrounds.

"Seekerhood."  Some investigators suggest that "a pattern of seekership preceding conversion appears to be associated with stigmatized communal groups in which participation entails discontinuity of social roles" (Robbins, 1988, p.82).  This view predicts that those who join cultic groups are likely to have a history of religious seeking and are likely to have belonged to other unorthodox groups.  Some empirical (Greil & Rudy, 1984) and clinical (Dubrow-Eichel & Dubrow-Eichel, 1988) evidence supports this notion.  This study examined the seekerhood issue by asking subjects to list other groups to which they belonged.

Literature Review

Because the focus of this study is the relationship between psychological harm and psychological abuse in groups ("cults" in particular), I limit this review to the literature having to do with psychological harm.  (The only study that examined psychological abusiveness in groups will be briefly discussed when the GPA Scale is described in the Methods section.)

It is important to reiterate that research on cult members is fraught with methodological problems.  Studies are often based on different definitions of what constitutes a cult or "new religious movement."  Subject samples are nearly always biased in some way.  Data collected from members of groups whose willingness to deceive is well documented can be untrustworthy.  "Pen-and-pencil" personality tests and surveys do not effectively measure certain psychological states such as dissociation.  Interview data are readily influenced by the fundamental conceptual assumptions of the interviewer.  Those who observe cults may not be sensitive to the psychological subtleties that clinicians detect.  Conversely, clinicians working with one cult member at a time may not fully appreciate social factors in cults.  Statistical analyses are sometimes inappropriate to the problem at hand.  And tendencies to over generalize and make unwarranted causal inferences are common.

These technical difficulties are compounded by the fact that cultic groups are reluctant to cooperate with critics.  Therefore, with a few exceptions, most of the nonclinical studies have surveyed ex-cult members or have been conducted by researchers whom cult leaders viewed as sympathetic.  Indeed, influencing academicians is a major goal of some groups (Dole & Dubrow-Eichel, 1981).

Clinical reports tend to see dissociation as central to cult members' adaptation to a demanding and contradictory environment.  Because self-report instruments do not effectively detect dissociation, critics view studies that use instruments such as the MMPI (Minnesota Multiphasic Personality Inventory) with considerable skepticism.  Indeed, in studies using the MMPI there is evidence that cult members are not honest in their responses; their Lie Scales tend to be elevated (Ungerleider & Wellisch, 1979) and there appears to be a "moderate attempt for both men and women to `look good'" (Ross, 1983, p.418).

Given these methodological caveats, what does the literature tell us?

Some research studies suggest that the level of harm associated with religious cults may be less than clinical reports indicate, at least for some groups.  Levine and Salter (1976) and Levine (1984) found little evidence of impairment in structured interviews of over 100 cult members, although Levine and Salter did note some reservation about "the suddenness and sharpness of the change" (p.415) that was reported to them.  Ross (1983), who gave a battery of tests, including the MMPI, to 42 Hare Krishna members in Melbourne, Australia, reported that all "scores and findings were within the normal range, although members showed a slight decline in mental health (as measured on the MMPI) after 1.5 years in the movement and a slight increase in mental health after 3 years in the movement" (p.416).  Ungerleider and Wellisch (1979), who interviewed and tested 50 members or former members of cults, found "no evidence of insanity or mental illness in the legal sense" (p.279), although, as noted earlier, members showed elevated Lie Scales on the MMPI.  In studies of the Unification Church (Galanter, Rabkin, Rabkin, & Deutsch, 1979; Galanter, 1983), the investigators found improvement in well-being as reported by members, approximately one-third of whom had received mental health treatment before joining the group.

Otis (1985) examined data from a survey of 2,000 members of Transcendental Meditation in 1971.  Dropouts reported significantly fewer adverse effects than experienced meditators, and "the number and severity of complaints were positively related to duration of meditation" (p.41).  There was a consistent pattern of adverse effects, including anxiety, confusion, frustration, and depression.  The "data raise serious doubts about the innocuous nature of TM" (p.46).

The Institute for Youth and Society (1980) in Bensheim, Germany reported that TM members tended to be withdrawn from their families (57% of subjects), isolated in social relations (51%), anxious (52%), depressed (45%), tired (63%), and exhibited a variety of physical problems, such as headaches and menstrual disorder.

Former members of a psychotherapy cult (Knight, 1986) reported that they had had sex with a therapist (25% of subjects), had been assigned love mates (32%), had fewer than 6 hours sleep a night (59%), and in therapy sessions were shoved at least occasionally (82%), were hit at least occasionally (78%), and were verbally abused (97%).  These subjects, 86% of whom felt harmed by the experience, also reported depression (50%) and menses cessation (32%).

In Conway, Siegelman, Carmichael, & Coggins (1986) study, ex-members reported the following experiences during their time in the cult:  sex with leaders (5%; 60% in the Children of God), menstrual dysfunction (22%), and physical punishment (20%).  Conway and Siegelman (1982) reported that ex-members experienced floating (52% of subjects), nightmares (40%), amnesia (21%), hallucinations and delusions (14%), inability to break mental rhythms of chanting (35%), violent outbursts (14%), and suicidal or self-destructive tendencies (21%).

Galanter (1983), who studied sixty-six former Moonies, reports that "the large majority (89%) felt that they `got some positive things' out of membership, although somewhat fewer (61%) did feel that `Reverend Moon had a negative impact on members,' and only a bare majority (53%) felt that `current members should leave the Unification Church'" (p.985).  Galanter also found that "36% of the respondents indicated the emergence of `serious emotional problems' at some time after leaving the church; 24% had `sought out professional help for emotional problems' after leaving; and 3% (i.e., two respondents) had been hospitalized for such problems during this interval" (p.985).  These findings were consistent with clinical reports during the 1970s and early 1980s.  It is interesting, however, that Galanter was sometimes inclined to put a positive "spin" on the findings, e.g., his choosing to write that "only (emphasis added) a bare majority (53%) felt that `current members should leave the Unification Church.'"  This is quite a large percentage given that, according to clinical investigations and countless ex-member reports, Unification Church members are indoctrinated to assume that the Church is always right and they, when dissenting, are always wrong.  Indeed, Langone (1992) found that the suppression of dissent was one of the five most highly rated cult characteristics in a subject pool of 308 former cultists from 101 different groups.  Thus, Galanter's indices of harm, though indirect and not low, may be underestimates.

In an in-process report of a survey of 308 former cult members, Langone paints an even more negative picture of the cult experience.  Eighty-eight percent of 308 subjects from 101 groups saw their groups as harmful (37%) or very harmful (51%).  During an average time of membership of 6.7 years, 11% of the subjects reported being sexually abused.  Sixty-eight percent of the subjects each knew an average of 28 former members who had not contacted helping resources.  Thus, approximately 5,500 persons known to these subjects had not sought help.  Yet 30% of the subjects estimated that "all or nearly all" of their friends and acquaintances had difficulty adjusting to post-group life, 21% felt that "most" had difficulty, 4% "about half," 13% "some," 6% "hardly any," and 25% were unsure.  Martin et al. (1992) used a variety of instruments, including the Millon Clinical Multiaxal Inventory (MCMI) to assess the psychological status of 111 former cultists.  Martin says:

This sample of ex-cultists can be characterized as having abnormal levels of distress in several of the personality and clinical symptom scales.  Of those subjects completing the MCMI-I, 89% had BR's ["Base Rates" -- indicates presence of a disorder] of 75 or better on at least one of the first eight scales.  Furthermore, 106 out of the 111 subjects (95%) who completed the MCMI at Time I had at least one BR score on one of the MCMI scales.  The contention that this population of former cultists is indeed distressed is further buttressed by their mean score of 102 on the HSCL (Hopkins Symptom Check List).  Typically, scores of over 100 are indicative of the need for psychiatric care.  Moreover, these ex-cultists had a mean of 72 on the SBS-HP burnout scale , which is suggestive of burnout and more than one standard deviation above the mean from Martin's (1983) sample of para-church workers.  (Martin et al., 1992, pp.231, 234)

Yeakley (1988) gave 835 members of the Boston Church of Christ (BCC) the Myers-Briggs Type Indicator (MBTI), a psychological instrument that classifies people according to Carl Jung's type system.  Individuals may differ in the way in which they tend to perceive (some being more sense oriented, others more intuition oriented), the way they judge (thinking oriented versus feeling oriented) and their basic attitudes (extraversion versus introversion).  Isabel Myers and Katherine Briggs, the developers of the MBTI, added a dimension to Jung's typology:  the person's preferred way of orienting himself to the outside world.  This orientation may be judging or perceiving.  The MBTI thus produces 16 personality types based upon the permutations of these variables.  Yeakley asked subjects to answer the questions in the MBTI as they think they would have answered before their conversion, as they felt at the time of testing, and as they think they will answer after five more years of discipling in the BCC.  He found that "a great majority of the members of the Boston Church of Christ changed psychological type scores in the past, present, and future versions of the MBTI" (p.34) and that "the observed changes in psychological type scores were not random since there was a clear convergence in a single type" (p.35).  The type toward which members converged was that of the group's leader.  Comparisons with members of mainstream denominations showed no convergence, but members of other cultic groups did show convergence, although toward different types than that on which the BCC members converged.  Yeakley concludes that "there is a group dynamic operating in that congregation that influences members to change their personalities to conform to the group norm" (p.37).  Yeakley's study does not tell us whether or not the changes occur, but it does provide compelling evidence that members perceive themselves as changing in the direction of a common personality type.  Although Yeakley's study did not directly examine harm, it does indirectly support clinical observations, which contend that the personalities of cult members are bent, so to speak, to fit the group.

Conclusions

Clinical observations (Ash, 1985; Clark, 1979; Langone, 1991) and research studies (Galanter, 1989; Langone, 1992) suggest that people join cults during periods of stress or transition, when they are most open to what the group has to say.  Approximately one-third appear to have been psychologically disturbed before joining, as evidenced by having participated in pre-cult psychotherapy or counseling (with figures varying from 7% to 62% of subjects among eight studies -- Barker, 1984; Galanter et al., 1979; Galanter & Buckley, 1978; Knight, 1986; Spero, 1982; Schwartz, 1985; Sirkin & Grellong, 1988).  The majority, however, appear to have been relatively normal individuals before joining a cult.

Certain studies cited earlier (Levine, 1984; Ross, 1983; Ungerleider & Wellisch, 1979) found that cult members score within the normal range on psychological tests or psychiatric interviews.  Galanter (1983) found some improvement in the general well-being of cult joiners, which he attributed to a psychobiologically grounded "relief effect" of charismatic groups.

Wright (1987) and Skonovd (1983) found that leaving cultic groups was very difficult because of the psychological pressure, a finding consistent with clinical observations.

There is much evidence, reviewed earlier, of psychological distress when people leave cultic groups.

And yet, they do leave.  Why?  If they were unhappy before they joined, became happier after they joined, were pressured to remain, left anyway, and were more distressed than ever after leaving, what could have impelled them to leave?

The inescapable conclusion seems to be that the cult experience is not what it appears to be (at least for those groups that deem it important to put on a "happy face"), either to observers or to members under the psychological influence of the group.  As Wright (1987) found, when members are separated from the group for a period of time, have an opportunity to share doubts with an intimate, witness the failures of the group, or learn about or observe the hypocrisies of the group's leadership, the group's psychological influence over the individual wanes.  Clinical observers, beginning with Clark (1979) and Singer (1978), appear to be correct in their contention that dissociative defenses help cult members adapt to the contradictory and intense demands of the cult environment.  So long as members are not rebelling against the group's psychological controls, they can appear to be "normal."  However, this normal appearing personality, as West and Martin (1994) maintain, is a pseudo-identity.  When cult members leave their groups, the flood gates open and they suffer.  But they don't generally return because the suffering they experience after leaving the cult is more genuine than the "happiness" they experienced while in it.  A painful truth is better than a pleasant lie.

If this analysis is correct, ex-members may indeed provide more accurate information about cults than would current members, although the responses of the former certainly cannot be treated as the last word on the issue.  Understanding the dynamics and effects of cultic groups is a difficult task.  Clinical observations and analyses have been very useful.  But the empirical testing of these observations and analyses will require many coordinated studies conducted over a period of years. 

This study is an early step in that it examines the hypotheses that an allegedly cultic/abusive environment can be distinguished from a noncultic environment and that former members of a cultic group experience more psychological distress than former members of noncultic groups.

Methods

Design/Procedures

This study had two components.

In Component One 228 former members of the BCC (available through a national mailing list from an ex-member organization) received the Group Psychological Abuse Scale, the DDD Scale, a background questionnaire, a consent form, a cover letter, a list of readings and resources, and two Boston University business reply envelopes (one for returning the instruments, the other for the consent form).  This component collected information on possible group psychological abuse from a national sample of former members of the BCC.

In Component Two subjects were administered the background questionnaire, the GPA Scale, and a battery of instruments assessing psychological background and distress.  This battery was developed by a team of researchers at Ohio University and Wellspring Retreat and Resource Center.  The test battery took most subjects 1.5 to 2.0 hours to complete.  Subjects from three groups were recruited: former members of the BCC; graduates of InterVarsity Christian Fellowship (a mainstream campus organization); and former members of a mainstream denomination (Roman Catholics).  This component enabled the investigator to compare the responses of former BCC members to two comparison groups on the dimensions of perceived psychological abusiveness and self-reported, post-group psychological distress.  (The DDD Scale was not given to subjects in Component Two in order to keep the test-taking time down to about 1.5 hours.)

Former Roman Catholics were recruited through advertisements in the Boston University newspaper.  InterVarsity subjects were recruited through lists of 56 graduates provided by InterVarsity leadership and from volunteers at an IV graduation weekend.  Former BCC subjects were recruited from a list of 152 Boston-area ex-BCC members on the national list mentioned above and from the First Congregational Church of Hamilton, Massachusetts.

Subjects received the test battery either at the Danielsen Institute or the First Congregational Church of Hamilton.  After completing the instruments, subjects were asked if they had any questions, were troubled about anything, or needed any information or assistance.  A reading/ resource list was given to each subject.

When subjects came to their appointment the investigator or an assistant briefed them about the study, reviewed consent forms with them, and administered all instruments.  After completing the self-report measures, subjects were seen again by the investigator or an assistant in order to answer questions and to address any concerns that subjects may have had.

Subjects

Component 1 (mailed survey).  A total of 40 former Boston movement subjects returned usable questionnaires, a response rate of 18%.  Sixty percent of the subjects were female.  Subjects' average age at the time of completing the questionnaire was 37.5 years.  Sixty percent were married, 30% single, and the other 10% equally divided among separated, divorced, divorced and remarried, and widowed.  Seventeen subjects had an average of 2.4 children; 3 subjects had an average of 2.3 children who were born while the subjects were in the group.  Eighty-seven and one-half percent of the subjects were white, 7.5% Asian, 2.5% black, and 2.5% Hispanic.  Current average annual household gross income was $42,162.

Religious upbringing more or less reflected national averages:  25% Protestant, fundamental; 20% Protestant, liberal; 30% Roman Catholic; 2.5% Jewish; 2.5% Eastern; 15% other; and 5% none.  Current religious preferences, however, indicated a marked falling away from mainstream religion, except for "Protestant, fundamental," which probably reflects the mainstream Church of Christ's energetic outreach to the Boston movement population:  32.5% Protestant, fundamental; 2.5% Protestant, liberal; 2.5%, Roman Catholic; 2.5% Jewish; 2.5% Islamic; 32.5% other; 22.5% none; 2.5% missing data.

Educational level was high, with 16.2 (SD=2.2) being the average number of years of education.  Subjects reported having earned the following degrees:  5% Associate; 42.5% Bachelor; 12.5% Masters; 2.5% Ph.D.; 12.5% other professional; 22.5% no degree; 2.5% missing data.  Mean income was $42,162 (SD = 4098).

Subjects had belonged to the Boston movement an average of 4.5 years and had been out of the group an average of 6.5 years.  Unlike early research studies (e.g., Conway et al., 1986), which relied upon samples derived from the network of cult educational organizations, only a small percentage of subjects from this study left the group because of deprogramming (2.5%) or exit counseling (10%).  Fifty-seven and one-half percent said they had "walked away"; 12.5% reported having been ejected from their group; 15% responded "other"; and 2.5% had missing data for this question.

Although this sample included a higher percentage of subjects familiar with cult educational organizations than was initially expected, the percentage is smaller than most studies of former members.  Limiting data only to the American Family Foundation (AFF) and Cult Awareness Network (CAN) because these are the only organizations with which more than a few subjects were familiar, we find that 9 of 39 responding subjects were familiar with AFF, 19 of 38 with CAN.  Seven of 40 responding subjects had received information from AFF, 15 of 40 from CAN.  Three of 40 subjects had attended AFF conferences or workshops, 6 of 40 had attended CAN conferences.  Thus, given the likelihood of overlap in these subject groups, it is reasonable to conclude that only about half of the subjects had some familiarity with cult educational organizations and even fewer had received information or had attended events sponsored by cult educational organizations.

Component 2  (test battery).  A total of 57 subjects (15 former BCC members, 23 former IV members, former IV members, 19 former Catholics) completed the test battery.  Females predominated in all groups.  (Percentages that follow in parentheses may not always add up to 100% because figures are rounded.)  Overall, 38 of 56 subjects were female (68%): 11 of 15 Boston movement (73%); 14 of 23 InterVarsity (61%); and 13 of 19 Roman Catholic (68%).  The average age of BCC subjects at the time of the interview was 33.35 years, that of IV subjects 23.64 years, and that of Roman Catholics 22.24 years.

Not surprisingly, given that subjects from two of the groups were either in college or recently graduated, 50 were single (89%).  Only 1 of 19 Catholics and 2 of 23 InterVarsity were married.  Two of 15 former Boston movement subjects were married, 1 was separated, and 1 was divorced. One former BCC subject had 3 children, 2 born while the subject was a member of the BCC.  One IV subject also reported having 3 children, all born while the subject was a member of IV.  None of the former Catholics reported having any children.

Regarding race, 36 subjects were white (64%), 4 black (7%), 1 Hispanic (2%), 13 Asian (23%; all in the InterVarsity sample), and 3 other (5%).  Thirteen BCC subjects were white (87%) and two black (13%).  Nine InterVarsity subjects were white (39%), 13 Asian (57%), and 1 other (4%).  Fourteen Catholic subjects were white (74%), 2 black (11%), 1 Hispanic (5%), and 2 other (10%).

Average annual household income of the groups was:  $44,767 (all groups); $38, 857 (BCC); $56, 389 (IV); $32,214 (RC).  The income figure for most subjects in the latter two groups probably referred to parents' income.

Former BCC subjects had belonged to the group an average of 4 years and had been out an average of 5.4 years.  IV subjects had belonged to InterVarsity for 2.4 years and had been out 1.8 years.  Former Catholics said they had been Catholics for 16.2 years and had left the church an average of 5.7 years before completing the survey.

Religious upbringing of BCC subjects was 1 Protestant, fundamental (7%); 6 Protestant, liberal (40%); 4 Roman Catholic (27%); 1 Eastern (7%); 1 other (7%); and 2 none (13%).  InterVarsity subjects reported the following:  5 Protestant, fundamental (22%); 1 Protestant, liberal (4%); 3 Roman Catholic (13%); 7 other (30%); 7 none (30%).  All former Roman Catholics had been raised in that religion.

Current religious preferences demonstrated substantial change from religious upbringing.  Obviously, none of the former Roman Catholics were still Catholics.  Fourteen of the 19 former Catholics (74%) declared "none" as their religious preference; 4 (21%) checked "other"; and 1 (5%) preferred Protestant, liberal.  Seven of 15 former Boston movement subjects (47%), again reflecting the influence of the mainstream Church of Christ on this movement, selected "Protestant, fundamental" as their current religious preference.  One identified with "Protestant, liberal" (7%); 1 with "other" (7%); and 6 with "none" (40%).  Ten of 23 InterVarsity graduates considered themselves to be Protestant, fundamental (43%); 6 Protestant, liberal (26%); 2 Roman Catholic (9%); 4 other (17%); and 1 none (4%).

Average number of years of education was also high for Component 2 subjects.  Former Roman Catholics, most of whom were still students at Boston University, had an average of 14.6 years of education (SD=2.0): 3 had a bachelor's degree; 2 a master's.  Former BCC subjects had an average of 16.5 years of education (SD=2.5): 2 had an associate's degree; 9 a bachelor's; 3 a master's; 2 another degree; and 3 no degree.  IV graduates had an average of 16.3 years of education (SD=.92): 23 had bachelor's degrees; 1 a master's (presumably one subject checked both categories).

Results regarding method of departure for the former BCC subjects were similar to those for Component 1:  10 said they walked away (67%); 2 were ejected (13%); 3 were exit counseled (20%).  One InterVarsity graduate and 2 former Catholics checked "other" for method of departure, perhaps reflecting some ambiguity in the listed choices with respect to these populations.

Instruments

Background.  A background questionnaire collected background information, information on the subject's experience with the group in question, and information on the subject's psychological history.

Abusive environment.  The psychological abusiveness of the group environment was assessed by the following instruments (only former BCC subjects in Component One completed the second instrument):

·         Group Psychological Abuse Scale (GPA Scale).  The GPA Scale is a factor-analytically derived, 28-item scale based upon the responses of 308 former members of 101 cultic groups to 112 questions inquiring into the nature of the group environment (Chambers et al., 1994).  The GPA includes four distinct factors, labeled Compliance, Exploitation, Mind Control, and Anxious Dependency.  The range for the composite score is 28-140.  The range for subscales is 7-35.  Scores above the midpoint (84 for the composite; 21 for subscales) indicate the subject is rating the group as in the abusive range.  A master's thesis reported that this scale successfully differentiated former members of the BCC in Cincinnati from former members of InterVarsity Christian Fellowship (Adams, 1993).  Former BCC members scored higher than IV graduates on all subscales.  This study replicated and expanded Adams's study, but added a former mainstream denomination comparison group.

·         DDD Scale.  This scale was developed by Carol Giambalvo and me in order to assess the degree to which former members of the Boston movement experienced concrete events related to deception, induced dependency, and dread, which Singer, Temerlin, & Langone (1990) contend are central to the behavioral changes observed in people who join cults.  The DDD Scale consists of a list of concrete behaviors and events exemplifying the DDD process as it is alleged to occur in the BCC.

Part 1 of the DDD Scale includes one question asking what name the group used when the person was first recruited and 18 questions that ask subjects to rate the accuracy (1=definitely true; 2=probably true; 3=don't know; 4=probably not true; 5=definitely not true) of statements that refer to the period of time when they were relatively new to the group, that is, when they first joined or the first few months after joining.  These statements were selected because they are believed to reflect common types of deceptions practiced by the BCC.  Part 2 asks subjects to rate the accuracy of statements that refer to any time during which they were members.  These questions were selected because they are believed to reflect practices and beliefs that contribute to deception, dependency, and dread.  This section consists of 48 questions that ask subjects to rate the degree to which the statement was accurate in their personal experience (e.g., "You were personally told this at least once") and 48 paired questions that ask subjects to rate the degree to which the statement applies to the BCC movement (e.g., "The statement above characterizes the group").  Three additional questions ask subjects how many disciplers they had, how many people they discipled, and the degree to which their happiness was a function of the qualities of their disciplers.  One question asked if the person was encouraged to date or discouraged from dating a particular person and one question asked if the person had been told to break up a dating relationship with a nonmember.

Five composite scores were constructed for the DDD Scale, with some questions being reversed scored because a "definitely true" rating would reflect a positive characteristic of the group.  Non-rated items were not included in the composite scores.  Thus, lower scores indicate a more negative evaluation of the group.  DDD consists of the sum of all ratings and has a range from 120 to 600.  DDDA consists of the sum of ratings in Part 1 and has a range from 18 to 90.  DDDB consists of ratings in Part 2 and has a range from 102 to 510.  DBSELF consists of those questions in Part 2 that refer specifically to subjects' personal experience (range: 48-240), and DBGROUP consists of those questions in Part 2 that refer to subjects' evaluations of the group in general (range: 48-240).  Assuming that subjects are more likely to be accurate in their ratings of personal experiences than their ratings of the group, DBSELF provides a more conservative and reliable means of assessing subjects' experience of practices contributing to deception, dependency, and dread.

·         Questions on the demographic questionnaire that inquire into perceptions of how much pressure was put on subjects when they left and their global evaluations of the group's harmfulness.

Social desirability.  Tendencies toward social desirability were measured by the Marlowe-Crown Social Desirability Scale (Crowne & Marlowe, 1960).  The Marlowe-Crowne is a 33-item true-false questionnaire designed to assess social desirability on self-report measures.  Correlations with MMPI scales are as follows:  L scale = 0.54; K scale = 0.40; Pd scale = -0.41; and Sc scale = -0.40.  Crino, Svoboda, Rubenfield, and White (1983) report an internal consistency coefficient for the scale, using the Kuder-Richardson formula 20, of 0.88 (n=39 undergraduates), test-retest reliability of 0.89 (n=31 undergraduates), and a correlation with the Edwards SDS of 0.35 (n=120 undergraduates).

Psychological distress.  The psychological distress and psychological background of subjects were measured by the following instruments:

·         SCL-90-R.  The SCL-90-R (Derogatis, 1977; Derogatis, Lipman, & Covi, 1973) is a 90-item, multidimensional, self-report measure of symptoms, especially those seen in psychiatric and medical outpatients.  The measure, which is designed to reflect psychological symptom patterns, is scored and interpreted according to 9 primary symptom dimensions and 3 global indices of distress.  The symptom dimensions are:  Somatization, Obsessive-Compulsive, Interpersonal Sensitivity, Depression, Anxiety, Hostility, Phobic Anxiety, Paranoid Ideation, and Psychoticism.  Global indices include the Global Severity Index (GSI), the Positive Symptom Distress Index (PSDI), and the Positive Symptom Total (PST).  Internal consistency and test-retest reliability coefficients for the SCL-90 range from .77 to .90 on the various symptom dimensions.  Criterion validity studies have been performed with several scales of the MMPI and the Middlesex Hospital Questionnaire, among other instruments.  The SCL-90 has also been used productively in studies with cancer patients, substance abusers, and sexual disorders.

·         Beck Depression Inventory (BDI).  The Beck Depression Inventory (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) is one of the most commonly used self-report methods of assessing depression.  Written in a multiple-choice format, the BDI is symptom focused, with each item reflecting behavioral manifestations and symptoms of depression, such as depressed mood, negative attitude, psychomotor retardation, and somatic complaints.  The four choices within each item are rank-ordered and weighted to reflect severity (0-3) of depression.  The scores across all items are summed, with the total score ranging from 0 to 63.  A score of 10 or more is considered to be beyond the normal range and a score of 17 or more suggestive of depressive disorder.

Internal consistency is high, with a mean alpha coefficient of .86 for studies involving psychiatric populations and .81 for nonpsychiatric populations.  The test-retest stability correlations, though troublesome because of the variability of a person's experience of depression, are adequate, ranging from .48 to .86.  Concurrent validity has been demonstrated with diverse measures, including the Hamilton Psychiatric Rating Scale for Depression, the Zung Self-Reported Depression Scale, the MMPI depression scale, and the Multiple Affect Adjective Checklist Depression Scale.  The mean correlations for the concurrent validity studies ranged from .60 to .76 (Conoley, 1990).

·         Dissociative Experiences Scale (DES).  The DES (Bernstein & Putnam, 1986) is a 28-item, self-report inventory used as a screening instrument for dissociative experiences and disorders.  The subjects are asked to indicate the percentage of time that they experience feelings and behaviors described in each statement, by making a mark on a line labeled 0 to 100.  The subjects are also asked to write down in a space provided, a percentage score corresponding to their mark.  The total score is calculated by summing across the indicated percentages.  The mean scores can range from 0 to 100.  The reliability coefficients (Spearman-Brown) of the items ranged from .19 to .75.  Test-retest reliability is reported to be r=.84, and the criterion-referenced validity is good as well, while the construct validity is adequate, ranging from .50 to .79 (Bernstein & Putnam, 1986).

·         Draijer & Boon (1993) and Steinberg, Rounsaville, & Cicchetti (1991) reported that the DES distinguished control patients from patients with dissociative disorders, as determined by scores on the Structured Clinical Interview for The Diagnostic and Statistical Manual of Mental Disorders Dissociative Disorders (SCID-D).  Ross, Joshi, and Currie (1991), who gave the DES to 1055 respondents in the general population of Winnipeg, identified three factors accounting for 47.1% of the combined variance of the scores:  absorption-imaginative involvement; activities of dissociative states; depersonalization/derealization.  Carlson and Putnam (1993) reviewed studies relating to norms, reliability, and validity of the DES and concluded that, though useful, the DES needs refinement and additional validation research.

·         Hopkins Symptom Checklist, subscales for dissociation.   Briere & Runtz (1990) took one item found in the standard HSCL (Hopkins Symptom Checkl List) and SCL-90 (Symptom Check List) which appears to tap dissociative symptomatology ("Your mind going blank") and added 13 items based on clinical experience and congruent with the style and comprehension level of the SCL-90.  These questions were then embedded in the HSCL or the SCL-90.  Reliability analysis of the Dissociation scale for Samples 1 and 2 showed internal consistency for both (alpha = .85 and .90, respectively).  For Sample 1, the mean, using the HSCL scoring format, was 20.95 (SD - 5.80).  For sample 2, using the SCL-90 format, the mean was 11.41 (SD - 10.11).  The Sample 2 scale showed a correlation between dissociation scores and histories of sexual (r = .14, p = .007) and physical (r = .23, p < .001) abuse in childhood.  Two brief dissociation measures are used in order to compensate for psychometric deficiencies in each.

·         Stait-Trait Anxiety Inventory (STAI).  The STAI (Spielberger, Gorsuch, & Lushene, 1970) consists of 40 items measuring two distinct forms of anxiety.  The state form of anxiety is transitory feelings of fear and worry, which most people occasionally experience.  As emotional states and subjective feelings vary in intensity, subjects are asked to indicate how well the statements describe them at the present time.  The 20 state-anxiety scale items are each rated on a four-point intensity scale, labeled "Not At All," "Somewhat," "Moderately So," and "Very Much So."  The trait form is a stable tendency for an individual to respond anxiously to a stressful situation.  Individuals are asked to indicate how they "generally feel."  The 20 trait-anxiety scale items are rated on a four-point frequency scale that is labeled "Almost Never," "Sometimes," "Often," and "Almost Always."  The scores of each of the two forms of anxiety range from 20 to 80.  High scores on their respective scales mean more state or trait anxiety.

The internal consistency for the state-anxiety scale ranges from .86 to .95.  The coefficient alpha for the trait-anxiety scale ranges from .89 to .91.  The number of significant correlations suggests that individual items have a good discriminating ability.  Test-retest intervals ranged from one hour to 104 days.  For the trait-anxiety scale the coefficients ranged from .65 to .86, whereas the range for the state-anxiety scale was, not surprisingly, .16 to .62.  The state-anxiety scale has good construct validity and a high level of face validity.  The trait-anxiety scale correlates with the Taylor Manifest Anxiety Scale (r=.80), the IPAT Anxiety Scale (r=.75), and the Multiple Affect Adjective Check List (r=.52).

·         Stait Trait Anger Scale (STAS).  The STAS, developed by Spielberger, Jacobs, Russell, and Crane (1983), consists of 20 items emphasizing the distinction between state and trait aspects of anger.  For the construction of this measure, state anger was defined as an emotional state or condition that consists of subjective feelings of tension, annoyance, irritation, fury, and rage.  The 10 state-anger scale items are each rated on a four-point intensity scale, labeled "Not at All," "Somewhat," "Moderately So," and "Very Much So."  Trait anger was defined in terms of individual differences in the frequency with which anger was experienced over time.  The trait-anger scale is divided into two subscales:  angry temperament and angry reaction.  The 10 trait-anger scale items are rated on a four-point frequency scale that is labeled "Almost Never," "Sometimes," "Often," and "Almost Always."  The scores on the state and trait scales range from 10 to 40.

The norms for the STAS are based on large samples of high school students, military recruits, and college students.  The alpha coefficient for college students for the state-anger scale is .95 for males and females.  The alpha coefficient for the trait-anger scale is .89 for males and .91 for females.  In a study of male inmates (Kroner & Reddon, 1992), the coefficient alpha was .94 for the state-anger scale and .88 for the trait-anger scale.  In this sample, the stability coefficient for the state-anger scale was .70 at one-week follow-up and .88 at one-month follow-up.  The test-retest coefficients for the trait-anger scale were .57 and .64, and one-week and one-month intervals, respectively.  The state scale's stronger stability than the trait scale is surprising.

The STAS is reported to have good construct validity.  In a sample of college students, the correlations of the STAS with the Buss-Durkee Hostility Inventory were .71 and .66, for males and females respectively.  The correlations of .59 and .43 (males and females, respectively) with the Hostility Scale, although statistically significant, are somewhat lower (Spielberger et al., 1983).

·         Impact of Events Scale (IES).  The Impact of Events Scale (Horowitz, Wilner, & Alvarez, 1979) contains 15 items that assess the experience of posttraumatic stress for specific life events and their context (e.g., death of a loved one).  The IES measures intrusive experiences and the recognized avoidance of certain ideas.  IES subscales show internal consistency coefficients ranging from .79 to .92.  The IES differentiated outpatients seeking treatment for bereavement and three field samples.  Normative data are available on a sample of 35 outpatients and 37 adult volunteers who had a recently deceased parent.

Psychological background.  The following instruments provided data on various aspects of psychological background:

·         Family Environment Scale (FES).  The Family Environment Scale (Moos & Moos, 1981) is a multidimensional measure of the perceived family environment.  It assesses the interpersonal relationships among family members (Relationship Dimension), the directions of personal growth stressed by the family (Personal Dimension), and the organizational structure of the family (System Maintenance Dimension).  The scale consists of 90 true-false statements divided equally among 10 subscales which constitute these three major dimensions.  Internal consistencies for the subscales range from .61 to .78  Test-retest reliabilities range from .68 to .86 (8 weeks) and .52 to .89 (12 months) (Busch-Rossnagel, 1991)

·         Physical Child Victimization Scale (PCVS).  The PCVS was adapted from Briere and Runtz (1988).  The scale consists of five items describing various parental behaviors that might be considered physically abusive.  The items are answered on a six-point scale ranging from 0 (never) to 6 (more than 20 times a year).  Subjects report the frequency of occurrence for each item before the age of 16, separately for each parent (mother or step mother and father or step father).  The scores on the PCVS may range from 0 to 30 for each parent, with the highest possible score being a 60.  The internal consistency of this scale is acceptable (alpha = .78 for mother and .75 for father).

·         Psychological Child Maltreatment Scale (PCMS).  The PCMS was also adapted from Brier and Runtz (1988) and follows the same format as the physical abuse scale.  It consists of 7 items that reflect verbal reactions, rather than physical behaviors, that might be psychologically damaging to the individual.  The items are scored on the same six-point scale as the ones for physical abuse, ranging from 0 (never) to 6 (more than 20 times a year).  The scores on the PCMS have a possible range of 0 to 42 for each parent, with a highest possible score of 84.  The internal consistency of this scale is also acceptable (alpha = .87) (Briere & Runtz, 1988).

·         Childhood Sexual Victimization Questionnaire (CSVQ).  This measure, comprised of eight questions, was adapted from Finkelhor (1979).  The subjects are asked to read each statement and respond whether the described sexual experiences occurred to them before the age of 16.  The items increase in severity ranging from "Another person showing his/her sex organs to you" to "Another person had intercourse with you."  The subjects are also asked the approximate age of the other person(s) and how old they were when they had the sexual experience corresponding to the last number to which they answer "yes."  In this study, individuals who endorse any of the sexual experiences described in the questionnaire with a person five or more years older than the respondent will be considered sexually victimized.

Results

Question One:  Perceived Abusiveness

Table 1 shows the means and standard deviations of the GPA composite scores of all subjects in both components of the study.

Table 1

GPA Composite Scores

A one-way ANOVA was conducted on the three groups in component two on the overall GPA score.  The ANOVA was significant, F(2,52) = 112.62, p < .0001.  Tukey post-hoc tests showed that the former BCC members scored significantly higher on the GPA (M = 108.5, SD = 11.28) than the former Catholics (M = 65.26, SD = 15.90) and the former InterVarsity members (M = 46.91, SD = 8.10).  The former Catholics scored significantly higher than the former InterVarsity members.  Neither the former Catholic, nor the IV group's average score was above the 84 neutral score separating nonabusive from abusive ratings.  Indeed, the mean for the ex-Catholic group was nearly 1.2 standard deviations below 84., while the IV mean was 4.6 standard deviations below 84.  The BCC mean, on the other hand, was 2.2 standard deviations above 84.

The GPA also evaluates the degree of psychological abuse on four subscales: anxiety, compliance, mind control, and exploitation.  A MANOVA was conducted on the four subscales for the three groups in Component 2.  The MANOVA was significant, Wilks'   = .108, F(8,98) = 25.00, p < .001.  Follow-up ANOVAs on each of the variables revealed that all were significant.  Tukey's post-hoc tests indicated that the former BCC members scored higher than either of the other two groups on all four subscales (see Table 2).  Former Catholics scored significantly higher than Intervarsity graduates on Anxious Dependency, Mind Control, and Exploitation.  Table 2 gives the results for subjects in both components.

Table 2

GPA Subscale Scores

A discriminant analysis was conducted to see if the composite GPA score alone could correctly classify individual subjects in their respective groups.  The discriminant function analysis revealed that the GPA score alone could predict group membership better than chance, Wilk's   = .188, F(2,52) = 112.62, p < .001.  The classification results in Table 3 revealed that 13 of the 14 former BCC members 92.9%) were correctly classified on the basis of their GPA scores.  None of the former IV members and only 2 of 19 former Catholics were incorrectly classified as former BCC members.  Overall, the discriminant function correctly classified 76.4% of the cases.

It must be noted that discriminant functions derived from a sample and used to classify cases on the same sample capitalize on chance variation of the sample and are very likely to generate more accurate classification results than would be achieved if the canonical coefficients were used to classify cases on a separate sample.

Table 3

Discriminant Analysis

Two questions in the demographic questionnaire asked subjects to give global ratings that reflect how abusive they perceived their group environment to be.  One question said: "Overall, how beneficial/harmful was the group experience for you?  Ratings were: 1 = "very beneficial," 2 = "beneficial," 3 = "neutral," 4 = "harmful," 5 = "very harmful."  The other question asked:  "To what extent did group pressures make it difficult for you to leave?"  Ratings were 1 = "very difficult," 2 = "difficult," 3 = "not sure," 4 = "mildly difficult," 5 = "no difficulty."

Table 4 contains the mean ratings for these two questions for all groups.  Regarding the question on harm/benefit, an ANOVA found that the groups differed significantly in their responses, F(2,54) = 80.22, p < .001.  Follow-up Tukey post-hoc tests revealed that the BCC scored higher than both the former Catholics and the IV graduates.  Former Catholics also scored significantly higher than IV graduates.  An ANOVA on the group pressure question was also significant: F(2,53) = 32.26, p < .001.  Follow-up Tukey post-hoc tests revealed that the BCC scored lower than both the former Catholics and IV graduates, and former Catholics scored lower than IV graduates (all ps < .05).  Lower scores reflect higher group pressure.

Table 4

Ratings of Group Harmfulness and Group Pressure

*Lower score implies greater group pressure

Question Two: Concrete Experiences

Table 5 contains the means and standard deviations of subjects' scores on the DDD Scale.  DDD refers to the sum of ratings for all items (range: 120 to 600; 120 corresponds to an average rating of 1; 240 to 2).  DDDA refers to the sum of ratings for Part 1 (range: 18-90; 18 corresponds to an average rating of 1; 36 to 2).  DDDB refers to the sum of ratings for Part 2 (range: 102 to 510; 102 refers to an average rating of 1; 204 to 2).  DBSELF refers to the sum of ratings for items in Part 2 that refer specifically to subjects' personal experience (range: 48-240; 48 corresponds to

an average rating of 1; 96 to 2).  DBGROUP refers to the sum of ratings for items in Part 2 that refer to subjects' evaluations of the group (range: 48-240; 48 corresponds to an average rating of 1; 96 to 2).  Lower scores reflect more negative ratings, with 1 = definitely true, 2= probably true, 3= don't know or can't decide, 4 = probably not true, and 5 = definitely not true.

Table 5

DDD Scale Scores (n=39)

Factor analysis was not performed on the DDD Scale because there were not enough subjects.  A correlational analysis of the DDD scale