Sociological Analysis 1991, 52:3 227-240
Religious Conversion of
Adolescents: Testing the
Lofland and Stark Model
of Religious Conversion*
Willem Kox
Wim Meeus
Harm 't Hart
Univeriity of Utrecht
In a study among 92 Dutch adolescents we tested the Lofland and Stark model of religious
conversion. Our results show that the model offers a fairly adequate set of conditions of
conversion, but is inadequate as a model for the process of conversion. Young people are attracted
to religious groups for two reasons: First, the group offers them a new perspective on life and
so liberates them from entrapment in their own problems. Second, the group provides them
with a satisfying personal network.
The model offered by John Lofland and Rodney Stark has been studied more
than most other models of religious conversion. These authors define conversion as
the process by which "a person gives up one . . . perspective or ordered view of the
world for another" (1965:862). Conversion implies that the convert renounces his or
her former life pattern, joins a given religious movement or sect, and becomes an
adherent of this group's teachings.
Lofland and Stark developed their model in a study of the Divine Precepts. The
model consists of seven conditions for conversion which the authors formulate as follows:
For conversion a person must:
1. Experience enduring, acutely felt tensions
2. Within a religious problem-solving perspective
3. Which leads him (her] to define him [her)self as a religious seeker
4. Encountering the D.P. (Divine Precepts] at a turning point in . . . life
5. Wherein an affective bond is formed (or pre-exists) with one or more converts
6. Where extra-cult attachments are absent or neutralized
7. And, where, if he (she] is to become a deployable agent, he [she] is expKDsed
to intensive interaction (1965:874).
*The authors arc associated with the Department of Development and S<Kializati(m of the Univeruty of
Utrecht. Willem Kox and Wim Meeus are members of the Section Youthstudies, and Harm 't Hart of the Section
Methodology. This article is based on the thesis of the first author (Kox, 19H9). Reprints may ht requested from
Wim Meeus. Section Youthstudies, Poxtbox 80.140, 3508 TC Utrecht, The Netherlands.
Ill
228 SOCIOLOGICAL ANALYSIS
Lofland and Stark's model contains both dispositional or "predisposing" and situational
conditions. The dispositional conditions (1, 2. and 3) are personality attributes
which make people potential converts (Lofland, 1966:33; cf. Lofland, 1977,1978). The
situational conditions (4 through 7) turn a potential convert into a factual one.
Accordingly, the model may be characterized as markedly interactionistic: conversion
is understood as a function of the interaction between (characteristics of) the individual
and the social environment.^
Lofland and Stark hold that these seven steps represent the necessary and sufficient
conditions for conversion. The conditions of their model, they say, have a "tunnellike"
effect. For every subsequent condition met by a potential convert, the factual
number of pxjtential converts becomes smaller, but for those remaining, the possibility
that they will convert increases. That is, the more conditions are met, the smaller
the number of potential converts left, and the greater their chance of actually converting.
As we said, the model has been studied a great deal. To assess it, certain problems
peculiar to conversion research must be taken into account. One of these is conceptual,
the others are methodological (see also Hoenkamp-Bisschops, 1982).
The conceptual problem concerns the definition of conversion. Conversion implies
both an objective and a subjective break in one's life. Tlie occurrence of an objective
break is not difficult to ascertain; the convert joins a group or movement, and because
of this his or her situation (home, environment) changes. It is harder to determine
a subjective break: at what point can one justifiably speak of a radical change in a
person's perspective on life?
Richard Travisano (1970) has made a rather elaborate attempt to solve the problem.
He defines conversion as a change of someone's general or "pervasive" identity. A
general identity, that of a priest for example, defines (1) a person's world view or life
perspective, (2) many aspects of one's personality, and (3) for a large number of situations,
prescribes how a person should relate to others and vice versa. Conversion, then,
is a revision of one's world view such that one's self-image and notions as to how
to deal with others undergo radical change. Travisano is prepared to speak of "radical
change" when the new identity contrasts with the old on the three points mentioned.
According to Travisano, an objective and a subjective break need not occur
simultaneously. A person can join a religious group without radically changing his
or her perspective on life — for instance, when transfering from one Christian group
to another. Travisano holds that use of the term "conversion" is warranted only if
the two kinds of breaks occur together. To join a group without undergoing radical
change in world view is to "alternate" rather than to convert.
The methodological problems have to do with the consequences that the process
character of conversion has for research and with the use of control groups. As the
model by Lofland and Stark indicates, the process of conversion can extend over a
long period. Ideally, conversion research should keep track of potential converts
throughout the process. But in practice this is hard to do, since intensive tracing of
'In the Netherlands this group is known as the Verenigingskerk. Internationally it has gained some
renown as the Unification Church. Its members are also referred as "Moonies" (after their leader, the
Rev. Moon).
RELIGIOUS CONVERSION OF ADOLESCENTS 229
people for periods of six months or more is almost invariably precluded by temporal
and financial constraints, the more so since propx)rtionally few of those who become
acquainted with a group actually end up joining it. This is why conversion research
is nearly always retrospective and yields only data concerning people who actually
went through with conversion. There is little wrong with this — as long as a control
group is questioned as well. After all, what we want to know is whether the findings
apply to converts exclusively. Many studies, however, do not seem to meet this rather
self-evident requirement.
These considerations provide us with a number of criteria by which research in
terms of the Lofland and Stark model may be evaluated (see Table 1):
1. Definition of conversion as subjective break. Does the study explicitly define conversion as
a radical change of life perspective?
2. Selection. Are the sole respondents researched those who by this definition may be
designated as "converts"?
3. Retrospective data gathering. Were the respondents interviewed during or after their
conversion?
4. Control group. Did the researcher(s) have recourse to a control group?
For each of Lofland's conditions we record the average percentage. Condition 6
is divided into two: 6a stands for absence of positive social ties with parents, friends,
and so on, while 6b represents neutral reactions in the personal network to the
respondent's contact with the group.
The table shows that most studies do not cover the whole Lofland and Stark
model; certain conditions only are selected. This raises the question as to what
conclusions may be drawn from the table.
It is evident that the model's conditions apply to a substantial part of the
respondents in the various studies. This may mean that Lofland and Stark have indeed
isolated a number of conditions relevant to conversion. There are two reasons, however,
why this conclusion cannot be drawn. First, we do not know whether the figures relate
to converts or to persons who become members of a group without radically changing
their perspective. In Travisano's terminology, some may be "alternators" only.
Conversion is defined in terms of an altered world view in no more than 12 of the
25 studies, and in fact just one study (Austin, 1977) used this definition in the selection
of respondents. Second, only three studies make use of a control group.^ Consequently,
it remains entirely possible that Lofland and Stark's conditions would apply
equally well to persons who did not join a religious group or movement.
The conclusion is inescapable: some twenty years of research have failed to yield
adequate empirical evidence for the conversion model developed by Lofland and Stark.
In the present study an attempt is made to procure such evidence. To that end we
implement three principles.
I. We adopt Travisano's definition of conversion. TTvat is to say, when a respondent joined
the three studies that did make use of a control group, reporting concerning the differences between
converts and the control group is unclear to the point that no conclusions can be drawn from it.
230 SOCIOLOGICAL ANALYSIS
a religious group a radical change in life perspective or world view (Travisano: "universe
of discourse") must have occurred as well.
2. The selection of converts is based on this. We looked for respondents who renounced
a nonreligious world view when they joined a religious group.
3. A control group was formed which matches the group of converts on a number of
background characteristics.
TABLE 1
Research Characteristics and Results in Terms of the Lofland and Stark Model
(Total number of studies: 25)
Mean Percentage of
Number of Relevant Respondents for Whom
Studies Study^ These Conditions Obtain
Research Characteristics
1. Definition of conversion
as subjective break
2. Selection according to this
definition
3. Data gathering retrospective
4. Control group
Results Relative to Loflaixd Conditions
1. Prior tensions
2. Religious perspective
3. Self-definition as seeker
4. Turning point
5. Affective bonds in group
6a. Social ties absent
6b. Social ties neutral
7. Intensive interaction
•Clarification: the numbers given under "relevant study" refer to the studies as given below. Studies
7 and 8 are reported in one article, as are studies 15 and 16.
12
1
19
3
9
13
5
3
13
8
3
4
1,2.3.4.5.10,11,
12,19,20.21.22
1
1.2.3,5,6,7,8,14,
15,16,17,18.19,
20,21,22,23.24,25
6,12,13
1,2,13.15,16.17.20.
22.23
1.2,4,13.15,16.17.
19,20,21.23,25
1,12.15,16,23
1.13,23
1,2,7,8.11.13,14.
15.16.17,18.22,23
1,15,16,17.18.19.
20.22
1,15,16
1,13,19.20
—
—
67
65
31
67
67
32
44
74
(1) Austin. 1977
(2) Baer. 1978
(3) Balch and Taylor. 1977
(4) Bromley and Shupe, 1979
(5) Christensen, 1965
(6) Downton, 1979
(7) Ebaugh and Vaughn 1, 1984
(8) Ebaugh and Vaughn 2, 1984
(9) Gerlach and Hine, 1970
(10) Gordon, 1974
(11) Harrison, 1974a
(12) Harrison,' 1974b
(13) Heirich. 1977
(14) Hoenkamp-Bisschops, 1979
(15) Lans and Dahlmans 1, 1982
(16) Lans and Dahlmans 2, 1982
(17) Lynch. 1978
(18) Nordquisc. 1978
(19) Richardson, Harder, and Simmonds,
(20) Richardson. Stewart, and Simmonds.
(21) Roberts, 1965
(22) Seggar and Kunz. 1972
(23) Snow and Philips. 1980
(24) Straus. 1976
(25) Wimbcrly et al.. N75
RELIGIOUS CONVERSION OF ADOLESCENTS 231
METHOD
92 young people participated in the survey, 46 of them converts and 46 in the
control group. Their ages ranged from 15 to 25 years. Most of the converts joined
a charismatic Christian denomination over the past two or three years: they became
members of the Unification Church or the Pentecostal Church.^ Typical of adherents
of these groups is their complete surrender of life and person to a higher, religious
power experienced in a daily, continuous, and intensely personal relationship. We
recruited these respondents with the assistance of youth leaders and clergy, and by
written invitation sent to persons whose names occurred on a mailing list of a gospel
music publisher.
To recruit a matched control group, we first asked converts to give us the name
of a p)erson whose background was similar to their own in terms of age, gender, training,
occupational level of father, and degree of urbanization. We asked such persons to
participate in the survey — 48 percent agreed to do so. The other 52 percent were
recruited via schools, community centers, and the university.
The average age in both groups was 21 years; 54 percent were male, 46 percent
female; 57 percent were trained at lower secondary or vocational schools (low-level
training), 43 percent at O/A level secondary schools or colleges (high-level training).
The two groups did not differ significantly on these points. Degree of urbanization
and parental occupational level were the same as well.^
The data were gathered in face-to-face interviews of at least 90 minutes each. T^e
respondents were asked to answer many written questions, to fill out some questionnaires,
and to respond to a number of open questions. Answers to the open questions
were categorized.
One special problem in the interviews was their retrospective character. The
interviewer kept having to lead the respondent to recall a period of up to three years
prior to conversion. For conditions 5 through 7, he asked about the period when
the respondent first came into contact with the religious group. For the variables used
to measure the presumed effects of conversion, the questions covered a span of 1.5
years subsequent to conversion. The interviewer asked the control group members
^The remaining 30 percent experienced conversion on average 4.5 years
^Background characteristics of converts (n
age
gender m/f
training
low
high
occupational level of male parent
low
intermediate
high
urbanization
rural
urbanized area
urban
= 46) and the control group
Converts
21.1
57/43%
59%
41
26%
33
41
15%
43
41
ago.
(n = 46):
Control Group
21.2
52/48%
54%
46
28%
33
39
20%
41
38
232 SOCIOLOGICAL ANALYSIS
to go back in time for a period of three to five years relative to the Lofland and Stark
conditions; to check for the presumed effects of conversion, he had them recall the
past two years. To avoid errors of recollection as much as possible, we invaribly asked
about tangible factual events and actual behavior. It is much easier for respondents
to report these then to recall aspirations and convictions in an unbiased way.
A large number of questions were asked relative to the conditions as specified
by Lofland and Stark and regarding each of various consequences of conversion.
Through factor analysis and scaling, the relevant variables were categorized into one
TABLE 2
Operation alizations of the Lofland and Stark Conditions
and the Presumed Effects of Conversion
Coruiition Operationalization
Type of
Measurement
Variance
Explained* Cronbadi's
Alpha
1. Tensions
2.
3.
4.
5.
Religious
perspective
Self-definition
as seeker
Actively
searching
Turning point
Affective
bonds
6a. Social ties
absent
6b. Social ties
neutral
7. Intensive
interactions
Changes,
presumably
effected through
conversion
Personal problems
Social problems
Personal problems relative
to one's person
Problems with one's situation
"Religious reflection"
factor score 1
factor score 2
nominal
nominal
nominal
38
24
.73
.71
Self-definition as seeker
Actively searching
Moving house
Change of jobs
Life-event
Frequency of contact with
friends in the group
Spare-time contact with
parents
Spare-time contact with peers
Emotional support by peers
Emotional support by parents
Material support by peers
Material support by parents
Neutral reaction by
parents
Interaction frequency with
groupmembers
nominal
scale score
nominal
nominal
nominal
scale score
factor score 1
factor score 2
factor score 1
factor score 2
factor score 1
factor score 2
scale score
interval
Change of general ideas
Reduced contact with
(close) relatives
Reduced contact with peers
factor score 1
factor score 2
factor score 2
29
20
44
24
41
30
71
22
.89
.72
.66
.53
.66
.68
.76
.73
.85
.64
.66
•"Variance explained" concerns the corresponding factor in a factor analysis with varimax rotation;
cf. note. 5.
RELIGIOUS CONVERSION OF ADOLESCENTS 233
or more measures.^ Most of these measures were newly designed for this study (see
Table 2). We discuss the Lofland and Stark conditions first.
1. To measure prior tensions a questionnaire was used and open questions were
asked. The list inquired about tensions on account of personal problems (loneliness,
feelings of inadequacy, etc.), and tensions due to concern with social problems
(unemployment, peace and security, natural environment, etc.). Factor analysis yielded
a two-factor structure: personal problems and socUd problems. By way of an open question,
we obtained information on various aspects of personal problems. The answers
were put into two categories: personal problems relative to one's person (e.g., lack of selfconfidence
or inadequate self-acceptance) and problems with one's situation (e.g., feelings
of meaninglessness, boredom, or lack of challenge).
2. The religious problem-solving perspective was traced by putting an open question
to the respondent concerning the manner in which he or she dealt with important
problems. The answers indicate a number of approaches. One of these, *Weligious reflection,"
was taken to define operationally the prior perspective of religious problem solving.
3. The definition of self as "seeker" was determined by an open question of whether
the respondent would describe him or herself as "a person in search of a group with
similar ideas." In addition, we checked the degree to which a respondent in fact engaged
in this search (in terms of the scale "actively searching"), by noting attendance at informational
meetings, courses taken, or participation in action groups. A respondent
whose answer for condition 2 was "religious reflection" and for the self-definition and
the scale replied positively to the designations "seeker" or "actively searching," was
taken to be a religious seeker or active religious seeker respectively. Since this occurred
in but a few cases (on account of the small number of positive responses to "religious
problem-solving perspective"), the designations "seeker" and "actively searching" were
used in the analysis as separate variables. These variables may be seen as secularized
versions of the variable "religious seeker" introduced by Lofland and Stark.
4. Operational turning points are moving house and change ofjob. In addition, an
open question was asked concerning incisive happenings immediately preceding contact
with the group {life-events such as the death of a parent, parents' divorce, running
away from home).
5. Friendship with members of the group: the scale ''frequency of contact voith frieruis
in the group" records how often these friends were contacted.
6a. To measure (lack of) other social ties, the personal network was demarcated
'In constructing our measurements we assumed a number of limiting conditions. For factor analysis
we took note of the Kaiser-Meyer-Olkin measure of sample adequacy and Barlett's test of the identitymatrix.
The Kaiser-Meyer-Olkin varies from .50 to .68. This is average. The values for Bartlett's test
vary from 17.48 through 273.71. p varies from < .00001 to .00003. In view of this very small p we
may safely a.ssume that the correlation matrixes used are not identity matrixes. Factor analyses were
carried out using principal factor solution. Rotation followed the Varimax method. Variables with a
factor loading < .35 were deleted from further analysis. Factor scores were subsequently dichotomized
at the zero point. Nominal variables were dichotomized at the median. When constructing scales based
on dichotomous variables it would seem obvious that a Guttman analysis be used. But this yielded no
useful results. We turned to Cronbach's alpha next, finding values varying from .53 to .91. This means
that the reliability of the scales is adequate to good. Scales were dichotomized at the median. Results
concerning simple variables were always related to receded dichotomous variables. Dichotomizing occurred
at the median.
234 SOCIOLOGICAL ANALYSIS
with the aid of the role-relation method (Meeus, 1989). In this method the respondent
indicates to what extent he or she relates to a standard set of persons, each of whom
has a distinct position and corresponding role with respect to him or her. This standard
set consists of one's partner, close friend, father, mother, brother(s), sister(s), other
relatives, friends, and acquaintances. For each of these nine reference persons the
respondents indicated how much of their spare time was spent with them and how
much material or emotional support they received from them. A seven-p)oint scale
was used for this, ranging from 1 (very much) to 7 (almost none). As in a previous
survey (Meeus, 1989), a two-dimensional factor structure was found for each of the
three domains: spare-time contact, material help, and emotional support from parents and
from peers respectively.
6b. Neutral reactions on the part of other social ties are measured by means of
the scale ''neutral reactions of parents." A scale "neutral reactions of peers" proved impossible
to construct.
7. Intensive interaction with group members was measured by one Likert-type
item: "interaction-frequency with group members."
Changes, presumably effected through conversion, were defined operationally as
follows:
1. Via a scale "change of general ideas" we established whether or not converts
had altered their ideas regarding school, work, and relationships.
2. Via the role-relation method we investigated whether contacts with the various
members of the personal network had diminished. As in the case of condition 6a,
factor analysis of the contact scores with the different members of the network resulted
in a two-factor structure: factor 1 stands for reduced contact with family and relatives;
factor 2 refers to diminished contact with peers.
CHARACTERISTICS OF COl^VEKTS AS A GROUP
What sort of young people convert firom a nonreligious background to a sectarian
Christian group? As a group they differ from the control group in a number of attributes
(see Table 3).
Converts have experienced p)ersonal problems far more often: 67 percent, compared
to 20 percent who had no such experience. Preoccupation with social problems is less:
7 percent compared to 26 percent. Both seekership and active searching is more common
among converts than among the control group members: the ratios are 54:24 percent
and 91:54 percent respectively. Converts fijrther experience a turning point in their
lives more often: they move house more frequently (80:57 percent), and they are exposed
to a life event more often (67:46 percent). The personal network of converts is of lesser
quality. They spend less spare time with peers (a ratio of 62 percent for converts compared
to 30 percent in the control group), and more converts feel that peers provide
poor emotional support (67:28 percent). Also, they receive less emotional support from
their parents (79:37) and less material help from peers (63:28 percent).
The converts' ideas regarding school, work, and other p)eople have changed far
more than those of nonconverts. Converts distanced themselves firom peers who
belonged to their old personal network more than nonconverts did. Accordingly,
RELIGIOUS CONVERSION OF ADOLESCENTS 235
TABLE 3
Converts and Nonconverts Compared in Terms of the
Lofland and Stark Conditions, and for Effects of Conversion
Conditions of Lofland and Stark
1 Personal problems
Social problems
Personal problems relative to one's person
Problems with one's situation
2 "Religious reflection"
3 Self-definition as seeker
Actively searching
4 Moving house
Change of jobs
Life-event
5 Frequency of contact with
friends in the group
6a Absence of:
Spare-time contact with parents
Spare-time contact with peers
Emotional support by peers
Emotional sup]x>rt by parents
Material support by peers
Material support by parents
6b Neutral reaction by parents
7 Interaction frequency with group members
Changes, presumably effected through conversion
Change of general ideas
Reduced contact with (close) relatives
Reduced contact with peers
Converts (%)
67
7
7
59
22
54
91
80
33
67
80
64
62
67
79
71
59
27
67
78
43
60
f^onconverts (%)
20
26
4
50
7
24
54
57
52
46
—
46
32
28
37
38
33
—
—
17
39
34
Significance
< .01
< .05
< .01
< .01
< .05
< .05
< .01
< .01
< .01
< .01
< .05
< .01
< .05
The results are presented in conformity with Lofland and Stark's formulations. That is to say that for
negatively formulated conditions, such as 6a, "absence of other social conditions," the percentages indicate
the degree to which this absence is the case.
converts may be described as follows:
1. They have relatively many personal problems, which they seek to solve. Hence
they are not likely to be much interested in social problems.
2. Their life is rather unstable. Shifts and turning points are relatively frequent
occurrences.
3. They have at their disposal a less well-functioning network. From both their
parents and their peers they receive relatively little social support.
4. Their notions regarding "secular" matters such as school, work, and relating
to other persons have changed. With regard to relations with other people the change
is not only with reference to the ideas, but actual contact with the old personal network
decreases as well.
236 SOCIOLOGICAL ANALYSIS
THE CONDITIONS OF LOFLAND AND STARK
Table 3 shows the differences between converts and nonconverts relative to the
Lofland and Stark conditions. Except for the condition "social problems," a percentage
of converts greater than that of nonconverts meets these conditions. Also shown is
that to arrive at conversion one need not meet all of Lofland and Stark's conditions.
There are no differences between converts and nonconverts on the matter of a prior
religious problem-solving perspective.
To see to what extent these conditions together can predict conversion and nonconversion,
a distriminant analysis was applied to the five conditions that relate to
both groups of respondents, i.e., "prior tensions" (1), "religious perspective" (2), "selfdefinition
as seeker" (3), 'turning pjoint" (4), and "social ties absent" (6a).^ The analysis
indicates which of the variables contribute factually to optimal discriminability of
converts and nonconverts. For each predictor in the discriminant analysis we indicate
successively the name of the variable, the Lofland and Stark condition referred to,
and the disciminant function coefficient. The figure of the discriminant function
coefficient represents the relative contribution of the variable in the distinction between
converts and nonconverts (see Table 4).
TABLE 4
Predictive Variables in the Discriminant Analysis Relative
to the Lofland and Stark Conditions
Label
Personal problems
Actively searching
Self-definition as seeker
Emotional support by peers
Moving house
Emotional support by parents
Change of jobs
Conditions
ofL&S
1
3
3
6a
4
6a
4
Type of
Measurement
factor score
scale score
nominal
factor score
nominal
factor score
nominal
Disaimijumt
Function
Coefficient
.50
.44
.36
.38
.47
.32
.30
The analysis shows that with the aid of five of the Lofland and Stark conditions
it is possible to distinguish converts from nonconverts in 85 percent of the cases (see
Table 5).7
In a discriminant analysis it is possible to combine the recoded dichotomous variables at the nominal
or ordinal level. The selection method used was to minimize the nonexplained variance between the
groups. To do so the analyzed variables were first dichotomized. The dependent variable was recoded
to 0 or 1. In this case, a discriminant analysis takes the form of a regression analysis. The order of
importance of the independent variables is determined by way of the Pearson correlation coefficient
with the discriminant function.
The table shows that for nine of the 84 respondents the prediction regarding the group to which
these people belong is incorrect. Additional characteristics that would distinguish these nine respondents
from the others yielded no result. Perhaps this should be ascribed to the contingency that attends conversion
processes no less than any other process. Untraceable, uniquely incidental factors have caused
these nine to remain outside of the pattern that does allow the other 75 resfx>ndents to be categorized.
RELIGIOUS CONVERSION OF ADOLESCENTS 237
TABLE 5
Predictive Value of the Lofland and Stark Conditions
hi ^ 84
Group
No conversion
Conversion
Correct prediction for all: 89%
Canonical correlation: .76
Chi: 62.52 p: < .001
n
43
41
Prediction
no conversion
39 (91%)
5 (12%)
conversion
4 ( 9%)
36 (88%)
Eighty-four of the 92 cases score validly on all selected variables.
Four of the five conditions of Lofland and Stark contribute to prediction in the
discriminant analysis. In view of the discriminant analysis function coefficient figure,
the contribution by these four variables is a substantial one. Two points are worth
noting: (1) Condition 2 does not contribute to prediction. Evidently, a religious problemsolving
perspective is not an independent condition for conversion; in fact, as we saw
in Table 3, it is no predictor at all. (2) Tlie contribution of the turning point identified
as "job change" runs counter to expectation. Converts change jobs less frequently
than nonconverts. Evidently, to gain new employment does not constitute a turning
point that would destabilize the life situation, which Lofland and Stark assumed it
would; rather, it indicates ability to shift from one function to another. Converts seem
less predisposed to change jobs. Turning points that destabilize the life situation are
more frequent among converts; they seem less able to create new transitions themselves.
In contrast to Lofland and Stark's theory, the conditions are largely independent
of each other. No significant correlations were found among the five conditions. This
finding contradicts the supposed cumulativity of the Lofland and Stark model (the
"tunnel"). Their claim is that a convert must meet a series of conditions in cumulative
succession, which means that these conditions cannot be mutually independent. If
Lofland and Stark were right, the discriminant analysis would bring out correlating
conditions via the best predictor. The others would not essentially contribute to prediction
and would disappear from the analysis. But this is not the case. A cumulative
sequence cannot be constructed on this basis: only five out of 46 converts proved
to meet the first three Lofland and Stark conditions. Our findings, then, contradict
the sequence of conditions which these authors assumed for their model.
It may be concluded, therefore, that the Lofland and Stark conditions provide
a good picture of converts and conversions, but not of the conversion process.
SUMMARY AND DISCUSSION
At the beginning of this paper we demonstrated that inquiry into religious conversion
has progressed little over the past twenty years. Our data show that the main
reason for this is inadequate research design. The combination of Travisano's defini-
238 SOCIOLOGICAL ANALYSIS
tion of conversion and the model developed by Lofland and Stark permits us to predict
conversion to religious groups quite well. But the model should not be understood
as being cumulative. The conditions listed first in the model are no prerequisites for
subsequent ones, nor does the convert have to meet all of them in order to arrive
at conversion. This means that Lofland and Stark have not designed a step-by-step
model of conversion; rather, they have pointed up a set of conditions for conversion.^
Two of the dispositional conditions in the model proved outstandingly important
in our study: enduring tension and seekership. Lasting tensions experienced by converts
are due primarily to dissatisfaction with their own personality. This unhappiness seems
to go hand in hand with a relatively strong fixation on the self; involvement with
social problems is relatively slight among these young people. They are in search of
solving their personal problems more than others are; they define themselves as seekers
more frequently and undertake more activities to find a solution. Noteworthy is that
the condition "religious perspective" does not influence conversion. This suggests that
an experienced lack is more conducive to conversion than a positive world view. This
lack may arise in dissatisfaction with one's own person or, more generally, dissatisfaction
with one's situation — at least, we are inclined to understand seekership as indicative
of this.
As far as the situational conditions for conversion are concerned, the life situation
variables and the personal network seem important. Incisive situational changes
and sudden negative life events are far more frequent among converts. Their personal
networks seem to offer little solace for their problems. Neither parents nor peers are
expected to give them much social support.
These factors render charismatic groups attractive to them for two reasons. First,
such groups offer a wholly other perspective on life and so deliver them firom their
fixation on their own problems. To look for solutions by oneself is no longer necessary;
henceforth a higher, charismatic power will do this. The fact that a previous religious
perspective does not point the way to conversion indicates that it is not so much the
content of the new perspective that appeals to converts. At issue is a formal aspect
of world views: they direct life and offer certainty. Next to this solution on the level
of consciousness the group provides factual supix)rt.
Eighty percent of the converts establish affective bonds with other members of
the group. This is very meaningful to people who experience little support from parents
and peers. It seems justified to suppose that religious groups have a twofold appeal:
ideological, by offering a new perspective on life, and social, by providing a satisfactory
social network.
^Strictly speaking we can make no claims regarding generalization of our results. In view of the marked
differences found between the control group and converts who joined two charismatic Christian
movements (Unification Church and Pentecostal Church), we assume that these differences will apply
also when nonconverts are compared to converts to other charismatic Christian groups.
RELIGIOUS CONVERSION OF ADOLESCENTS 239
REFERENCES
Austin, R. L. 1977. "Empirical adequacy of Lofland's conversion model." Review of Religious Research
18:282-87.
Baer, H. A. 1978. "A field persp)ective of religious converstion: the Levites of Utah." Review of Religious
Research 19:279-94.
Balch, R. W. and D. Taylor. 1977. "Seekers and saucers: the role of the cultic milieu in joining a UFO
cult." American Behavioral Scientist 20:839-60.
Bromley, D. G. and A. D. Shupe. 1979. "Just a few years seem like a lifetime: a role theory approach
to participation in new religious movements." Research in Social Movements, Conflicts and Change
2:159-85.
Christensen, C. S. 1965. "Religious conversion in adolescence." Pastoral Psychology 9:17-28.
Downton, J. V. 1979. Sacred Journeys: The Conversion of Young Americans to Divine Light Mission. New
York: Columbia University Press.
Ebaugh, H. R. F. and S. L. Vaughn. 1984. "Ideology and recruitment in religious groups." Review of
Religious Research 26:148-57.
Felling, A. et al. 1981. "Gebroken identiteit [Shattered identity)." Jaarboek Katholiek Documentatiecentrum
11:25-81.
Gerlach, L. P. and V. H. Hine. 1970. People, Power, Change: Movements of Social Transformation.
Indianapolis: Bobbs-Merrill.
Gordon, D. F. 1974. "The Jesus people: an identity synthesis." Urban Life and Culture 3:159-78.
Harrison, M. 1. 1974a. "Preparation for life in the spirit: the process of initial commitment to a modern
religious group." Urban Life and Culture 2:387-413.
. 1974b. "Sources of recruitment to Catholic pentecostalism." Jourrial for the Scientific Study of
Religion 13:49-64.
Heirich, M. 1977. "Change of heart: a test of some widely held theories about religious conversion."
American Journal of Sociology 83:653-80.
Hoenkamp-Bisschops, A. M. 1979. Her verschijnsel bekering bij nieuwe religieuze bewegingen [The Phenomenon
of Conversion in "New Religious Movementjl. Nijmegen: Report Psychological Laboratory.
. 1982. "Bekeringsonderzoek: geschiedenis en methodologische problemen [Conversion research:
history and methodological problemsl." Gedrag 10:57-74.
Kilboume, B. and J. T. Richardson. 1988. "Paradigm conflict, types of conversion and conversion theories."
Sociological Analysis 50:1-21.
Kox, W. 1989. Van crisis naar Christus: Een modelmatig onderzoek naar de bekering van niet-religieuze jongeren
tot christeliyjke charismatische kerken [From Crisis to Christ: Model Research Concerning the Conversion
of"NonreligiousYoungsters to Charismatic Christian Churches]. Utrecht: Unpublished thesis.
Lans, J. M. van der and R. M. M. Dahlmans. 1982. "Toetreding tot religieuze bewegingen (Verenigingskerk
en Bhagwan-beweging): een bijdrage tot toetsing van het bekeringsmodel van Lofland [Joining
religious movements (Unified Family and Bhagwan Movement): a contribution to testing Lofland's
conversion model)." Gedrag 10:39-57.
Lofland, J. 1966. Doomsday Cult: A Study of Conversion, Proselytization and Mainterumce of Faith. Englewood
Cliffs, NJ: Prentice-Hall.
1977. Doomsday Cult: A Study of Conversion, Proselytization and Maintenance of Faith, enlarged
ed. New York: Irvington.
and R. Stark. 1965. "Becoming a world saver: a theory of conversion to a deviant perspective."
American Sociological Review 30:862-75.
. 1978. " 'Becoming a World Saver' revisited," in J. T. Richardson (ed.). Conversion Careers: In
and Out of the New Religions. Beverly Hills, CA: Sage.
Lynch, F. R. 1978. "Toward a theory of conversion and commitment to the occult," in J. T. Richardson
(ed.). Conversion Careers: In and Out of the hJew Religions. Beverly Hills, CA: Sage.
Meeus, W. 1989. "Parental and peer support in adolescence," in K. Hurrelmann and U. Engel (eds.),
The Social World of Adolescents. New York: De Gruyter.
Nordquist, T. A. 1978. Ananda Cooperative Village: A Study in the Beliefs, Values and Attitudes of a New
Age Religious Community. Uppsala: Religionhistoriska Institutionen.
Richardson, J. T., M. Harder, and R. B. Simmonds. 1972. "Thought reform and the Jesus movement."
Youth and Society 12:185-202.
240 SOCIOLOGICAL ANALYSIS
Richardson, J. T., M. W. Stewart, and R. B. Simmonds. 1979. Organized Miracles: A Study of a Contemporary,
Youth, Communal, Fundamentalist Organization. New Brunswick, NJ: Transaction Books.
Roberts, F. J. 1965. "Some psychological factors in religious conversion." British Journal of Social and
Clinical Psychology 4:185-87.
Schnabel, P. 1982. Tussen stigma en charisma [Between Stigma and Charisma]. Deventer: Van Loghum
Slaterus.
Seggar, J. F. and P. Kunz. 1972. "Conversion: evaluation of a step-like process for problem-solving."
Revieu; of Religious Researck 13:178-84.
Snow, D. A. and C. L. Phillips. 1980. "The Lofland-Stark conversion model: a critical reassessment."
Social Problems 27:430-47.
Straus, R. A. 1976. "Changing oneself: seekers and the creative transformation of life experience," in
J. Lofland (ed.). Doing Social Life: The Qualitative Study of Human Interaction. New York: Wiley.
Travisano, R. V. 1970. "Alternation and conversion as qualitatively different transformations," in G.
P. Stone and H. A. Farberman (eds.). Social Psychology Through Symbolic Interaction. Waltham, MA:
Ginn-Blaisdell.
Wimberley, R. C. et al. 1975. "Conversion in a Billy Graham crusade: spontaneous event or ritual performance?"
Sociological Quarterly 16:162-70.