Cultic Studies Journal, Vol. 2 No. 1 1985, Page 62
dozen teenagers who generated appropriate terms and language for the semantic
differential items used. A pilot sample of 150 students from local Catholic and Jewish youth
centers was then used to pre-test the wording of questions and other features of the
survey.
The 50 items of information requested in this self-report survey were organized around four
general classes: pre-contact variables, contact variables, mediational variables, and
outcome measures. Questionnaire format varied within each of these classes of items.
Students responded using check lists, open-ended write-ins, Likert-type response scales,
and semantic differential scales. Some questions had as many as 30 subparts.
Pre-Contact variables
Pre-contact variables relate to predispositions of the respondents and to their knowledge
structure. Fixed predispositions were sex, age, race, religion, and family composition.
Modifiable predispositions included background questions about GPA, geographical mobility,
nature and extent of media exposure, spiritual practices, type of classroom preferred,
shyness, views about authority, and the value of exposure to opposition ideas.
Knowledge structure regarding cults was measured by items that asked the students to
identify if they had heard of any or over 30 listed groups, and whether or not each was a
cult. Descriptions of prototypical cult members were also elicited, as were general
perceptions of groups referred to as cults. Semantic differential subsets were used (with
nine-point equal intervals) to collect these reactions. Thus cults might be: charity
minded/profit minded, non-religious/religious, and so forth. Typical cult members were
described on fifteen semantic differential subsets including: foolish/wise,
responsible/irresponsible, weak/strong, lonely/not lonely, and others. Check lists of the
primary purposes of these groups included both multiple response items, such as: ―to raise
money,‖ ―to improve the world,‖ ―to destroy individual freedom,‖ ―to strengthen the
American family,‖ and eleven others, and also open-ended items for free responses to this,
and other check-list items.
Contact variables. Contact variables were those related to the recruiter‘s strategies and
tactics. The fundamental question that separated our participants into those with and
without contact was: ―Have you been approached by a member of a group that you
thought was a cult?‖ For those who had been contacted, the survey elicited information
about the situation in which that contact had taken place, e.g., public or private, whether
alone or with others. Descriptions of the recruiter and the verbal and non-verbal behaviors
she or he or she used during the contact were also noted by the respondents.
Mediational processes. Mediational processes were indirectly measured by trying to
determine the degree of perceived congruence between the student-respondent and the cult
recruiter, as well as the extent of similarity perceived between the student and prototypical
cult members. A series of questions asked the student‘s reaction to the person who had
made the contact. Some reactions were of a general type: friendly/unfriendly,
afraid/unafraid, interested/disinterested, angry/pleasant. Other items focused on the
recruiter‘s specific verbal and non-verbal aspects that the student felt affected him/her
(either favorably or unfavorably). Among them were: eye contact, enthusiasm, affection,
hand gestures, dress, tone of voice, product being sold, and so forth. Affective reactions
were also elicited to a series of eight items asking about imagines transactions within cults,
such as: if a relative dated a cult member, if parents attended a cult meeting, or if parents
objected to the student‘s joining the group.
Outcome measures. Outcome measures were of three kinds. The most basic question is
that which established whether there had been any contact with a cult recruiter: ―Have you

Cultic Studies Journal, Vol. 2 No. 1 1985, Page 63
been approached by a member of a group that you thought was a cult? Answers to this
item divided our sample into two broad groupings of those contacted and those never
contacted. Number of contacts and consequences of that contact (what it led to) were also
assessed. Another major outcome measure we developed probed the student‘s interest in
attending a cult group meeting, if invited. Those who ―definitely would not go‖ (non-
negotiators) were categorized separately from all the rest of the respondents who were
accessible to further exposure to a cult, namely, all who answered ―definitely would go,‖
―might go out of curiosity,‖ ―might go with a friend,‖ and ―not sure‖ (negotiators).
Survey Administration
During the spring of 1980, one of our staff administered the eleven-page survey to
classroom groups of 15-30 students. A standard introduction provided general guidelines
for taking the survey, along with explanations of those items with a potential difficulty of
interpretation. Most students completed the task in about 30 to 40 minutes. Contingency
(or filtered) questions made it easy for students to answer only relevant items by skipping
sections if they had not heard of the word ―cult,‖ or if they had never been approached by a
cult member.
For classes with low reading levels, the survey was guided by a local radio announcer‘s
tape-recording of instructions and questions. When requested by a teacher, a lengthy
question-and-answer session followed the survey administration.
Data analysis
The questionnaire responses were sorted into categories, assigned code numbers, and then
manually transcribed onto optical scanning forms. These data were verified by independent
judges who visually compared the scanning form numbers with the original survey
responses and assigned codes. Statistical analyses were conducted at the Stanford
University Computer Center with the aid of various SPSS software programs (Nie, Hall,
Henkins, Steubrent, &Bent, 1980).
In order to provide a general portrait of the respondents, descriptive statistics of selected
characteristics of the entire sample will be presented first. Percentages for any item
reported have been adjusted for valid codes only, that is, excluding missing or faulty data.
Significance tests of relationships among measures were computed using Chi Square
(corrected for continuity). With more than two categories within variables, an additional
test for ordered categories (Goodman and Kruskal‘s Gamma) was applied. Following this
preliminary data presentation, we will use our multivariate analyses to address the primary
questions of this investigation: is it possible to identify a set of characteristics that predict
which individuals will be contacted by a cult recruiter, and if contracted, which of these
students will reject or be receptive to further inducements for affiliation with a cult? Chi
square analyses and discriminant function analyses were used to assess the predictive
utility of components of our theoretical model for each of these two outcomes.
Results
Our results will be presented in four sections: (1) data related to the two outcome
measures (b) data that provide information about the students‘ conceptions of cults and
the descriptions of their contact with cult recruiters (c) the associations between selected
independent variables from the students‘ pre-conversion background (predispositions,
knowledge, and attitudes) and the dependent outcome variable and (d) multivariate
analyses of a composite of attributes that predict which type of students are most likely to
have contact with a recruiter, and which are likely to be receptive to future affiliation with
cult groups.