Measuring faith development.

Parker, Stephen

James Fowler's theory of faith development has had a significant influence on religious education, pastoral care, and developmental psychology. Since he introduced the notion of faith "stages," there have been several attempts to measure these, beginning with Fowler's own work. This article reviews and evaluates the adequacy of the various instruments used to measure Fowler's theory of faith development.

**********

James Fowler's theory of faith development is one of the most significant models of religious/spiritual development to emerge in the last thirty years; it has influenced religious education, pastoral care, and developmental psychology (Astley & Francis, 1992; Santrock, 1999). His model of "faith stages" has inspired over 200 research projects (Fowler, 1990) and continues to generate debate and appeal (Streib, 2001, 2005). Given the wide interest in Fowler's model, a review and evaluation of the various instruments designed to measure Fowler's faith development stages will be beneficial to those desiring to contribute to and evaluate the research on faith development.

FOWLER'S THEORY OF FAITH DEVELOPMENT

Those unfamiliar with Fowler's model will find a helpful summary in Fowler (1992), or a fuller introduction in Stages of Faith (Fowler, 1981). Only a brief sketch is offered here. It also is not the purpose of this article to address the myriad of questions concerning the adequacy of Fowler's theory (e.g., Broughton, 1986; Day, 2001; Jardin & Viljoen, 1992; McBride, 1976; Nelson & Aleshire, 1986; Parks 1991; Streib, 2001, 2005). Those interested in the ongoing debate about Fowler's way of conceptualizing faith should consult the collections in Dykstra and Parks (1986), Astley and Francis (1992), and a recent issue of the International Journal for the Psychology of Religion (Streib, 2001).

Modeling his work after the cognitive-structural psychologies of Piaget (1970) and Kohlberg (1976), Fowler proposed seven "stages of faith," by which he meant the developmental path of a person's way of making sense of and relating to, what he termed, one's "ultimate environment" (Fowler, 1981). Faith, as this universal human activity of meaning-making, is rooted in certain "structures" (inherent in human interaction) that give shape to how humans construe and relate to self and world. These structures (e.g., cognitive development, level of moral reasoning, and locus of authority, among others) are distinguished from the content of faith. That is, faith is understood, not as a set of beliefs, but as a way of knowing, a way of constituting one's experience of the world. It is the structures (see Table 1), rather than the content of faith, that allows for determination of one's faith "stage."

Fowler (1981) further argues for an interactive complementarity among all of these structures, so that faith cannot be thought of as reducible to any single structure (such as moral development). The structural integrity or wholeness of these seven structures is Fowler's way of addressing the multifaceted nature of faith and is designed to acknowledge the cognitive, affective, and relational aspects of faith. Stage progression, when it occurs, involves movement toward greater complexity and comprehensiveness in each of these structural aspects. Table 2 summarizes Fowler's faith stages.

MEASURES OF FOWLER'S STAGES

There are several instruments that try to operationally measure Fowler's faith stages. This article describes and evaluates the instruments reported in the research literature, including two that have been previously reviewed (Hill & Hood, 1999).

The Faith Development Interview (FDI)

The chief instrument by which Fowler's faith stages have been determined has been the "faith development interview" (FDI). Developed for Fowler's own research, the FDI is a "semi-clinical" interview focused on significant life experiences and the meanings attributed to them (Fowler, 1981; cf. Burris, 1999). It remains the only authorized measure of faith development stages (Moseley, Jarvis, Fowler & DeNicola, 1993), and is one of two instruments for measuring Fowler's stages reviewed in the Measures of Religiosity compendium (Hill & Hood, 1999).

FDI Structure. The FDI covers four broad areas and is designed to elicit answers that will allow the researcher to assess the current level of complexity on the seven structural aspects that define faith (Moseley et al., 1993). The four broad areas include: (1) a general life review, which is inclusive of date and place of birth, siblings, parent's occupations, ethnic, racial, and religious identifiers, and social class indicators; this section also asks for major turning points or changes in one's life and closes by asking the meaning of life; (2) a review of life shaping experiences and relationships, which focuses on such things as childhood memories of parents, significant losses, "peak experiences" and taboos; (3) a description of present values and commitments, which focuses on such items as the purpose of human life, one's ethical sense of right, rituals, symbols and groups one finds important, as well as thoughts about death and suffering; (4) specific questions about religion, which includes questions on the nature of prayer, sin, and the relationship of religion to morality. Specifically religious questions are asked intentionally last in order to "overhear [a person's] ways of shaping and interpreting meanings from their lives" (Fowler, 1981, p. 308) without undue imposition of religious categories for this task. Since 1986 the information in section (1) has been covered by an optional autobiographical questionnaire (the "Life Tapestry Exercise") that interviewees can fill out prior to the interview. This is then reviewed during the interview (Moseley et al., 1993). The intent of the FDI is to elicit answers that provide insight into all of the structural elements Fowler proposes as well as inquire into the cognitive, affective, and relational dimensions of faith.

FDI Administration. In its current form, the FDI is a semi-structured interview that asks certain questions of all interviewees, though the order and form of the questions may vary somewhat, based on the person's previous comments. Subjects are usually taped as they respond to a series of questions on "attitudes and values in life" (Fowler, 1981, p. 308). Interviews generally last two to three hours although those with children aged four to eleven are typically shorter and are modified to include doll-play and story completion (Moseley & Brockenbrough, 1988). Administration of the FDI requires some sensitivity to interviewees and their reactions to the issues the FDI raises; it is not intended to be done in a mechanical way (Burris, 1999; Fowler, 1981; Snarey, 1991).

FDI Scoring. Stage assignments are made by comparing a respondent's answers from their FDI to formal stage descriptions and specific stage-level guidelines for each of the seven structures of faith development that are outlined in the Manual for Faith Development Research (Manual, hereafter; Moseley et al., 1993). Using a scoring guide listing the seven structural aspects, and formal stage descriptions provided in the Manual, statements by an interviewee can be compared with certain benchmarks that signify whether a given answer is inclusive of a specific stage or clearly signals that this way of thinking, perspective taking, using symbols, etc. is to be excluded from that stage. In his Stages of Faith, Fowler (1981) provides enough background information, transcript excerpts, and instruction with an illustrative case ("Mary") to see how a faith stage level is assigned from information gathered by the FDI. The Manual (Moseley et al., 1993) recommends having some practice scorings reviewed by an experienced FDI user to increase the reliability of stage assignments.

Final determination of a person's faith stage involves some simple arithmetic. Each interviewee response is assigned a stage level and a location among the aspects. By adding the stage scores for each response under each location (e.g., perspective taking), then dividing by number of responses, scorers arrive at an arithmetic average (with a range of 1-6) for each structural aspect. The averages from the structural aspects are then combined and averaged again to yield an overall score for each interviewee (again, with a range of 1-6). Those whose average comes to .39 or less are assigned to the stage lower while those with .70 and higher are assigned to the next higher stage. Person's whose arithmetic average falls between .40 and .69 are said to be in stage transition.

FDI Validity. Two studies on the construct validity of faith development theory offer the most direct support of the validity of the FDI. Snarey's (1991) research with a sample of non-theistic kibbutz leaders analyzed the construct validity of faith development theory using the FDI as his measure of faith development stages. His study provided confirmatory results for four proposed hypotheses: 1) stages of faith are structural wholes; 2) variations in level of faith development predict relevant outcomes in terms of religious, sociological, and psychological characteristics, 3) stages of faith are not reducible to or solely determined by stages of moral development, and 4) stages of faith are cross-culturally universal.

The structural wholeness of Fowler's stages was supported by several statistical analyses that, taken as a whole, point to faith development as a unified dimension. As one would anticipate if the seven structures are all aspects of faith, a correlation matrix of scores on the seven aspects with scores on each of the others were "all positive, moderately strong, and highly significant" (Snarey, 1991, p. 290). Furthermore, if the seven aspects are all necessary, then one would anticipate factor analysis to reveal only one factor. This also is what Snarey found: a factor analysis of scores on the 7 aspects showed one factor accounting for 73.7% of the variance with an eigenvalue of 5.16. A second factor had an eigenvalue of less than 1 confirming that only one factor is being measured. Snarey also found that when aspect scores did not cluster to a single stage, these scores occurred only in adjacent stages (as one would anticipate for structural stages).

Snarey (1991) also investigated criterion validity by looking at the predictive quality of faith stages. He found that faith stages were strongly predictive of education level (r = .49), occupational level (r = .45), social class (r = .43), and work complexity (r = .49). Multiple regression analysis provided the level of variance accounted for by each criterion. The correlation with educational level supports the notion that faith stages involve movement toward complexity in thought processing (e.g., Piaget's cognitive stages). The multiple regression analyses also confirmed a contribution for each variable; that is, the correlations do not simply reflect a common sociological artifact connected to class or education. Snarey also found no relationship between faith development and gender (exactly what one would expect from structural stages); his finding of no relationship between faith development and age is not surprising in this study since he only interviewed a rather narrow age range of adults.

Snarey's (1991) finding of moderately positive correlations between faith stages and moral development stages (.597) is more what one would anticipate if these two phenomena are related but not identical. This is clearly more supportive of Fowler's argument about the relationship of moral development and faith development than the higher correlations that others have reported (e.g., Mischey, 1981).

The universality of Fowler's stages finds support in a chart showing comparable mean FDI scores from Snarey's (1991) sample of non-theistic kibbutz leaders with FDI means from other studies of faith stages in Protestant, Catholic, and Jewish adults. Snarey's study confirms that non-Christian beliefs are not undervalued by Fowler's model. Snarey also interpreted his finding that city dwellers had higher average faith stages than kibbutz dwellers as supportive of the idea that diversity of experiences aid faith development, though it also may raise questions as to class bias in FDT (Timpe, 1999).

Driedger's (1998) study using the FDI replicated aspects of Snarey's (1991) work, confirming several of his findings, but challenging others. Factorial analysis of her study of 65 HIV-positive adults confirmed Snarey's finding that the FDI measured a single underlying construct that accounted for 84% of the variance in her sample. She also found significant positive correlations between faith development and educational level (r = .48), ego development level (r = .31), and Intrinsic religiousness (r = .38; Allport & Ross, 1967), but not with Extrinsic religiousness. Her finding of a significant correlation between faith stage and ego development (r = .31) points to the relationship between what Fowler is charting and the evolving sense of self that Loevinger (1976, 1979) describes, but also shows that the one is not identical to the other. The correlation with Intrinsic/Extrinsic (I/E) religiosity probably reflects the orientation toward meaning making characteristic of both Fowler's work and the Intrinsic component of the Religious Orientation instrument. Driedger's findings are similar to Chirban's (1981) finding that higher faith stages correlated with Intrinsic religiousness while lower faith stages correlated with Extrinsic religiousness. While Dreidger's finding of a higher correlation between moral development and faith development (r = .80) is closer to those found by others (Mishey, 1981; Shulik, 1979), and thus, raises questions about the amount of overlap between these two constructs, the overall support for the theory of faith development suggests that the instrument used to measure faith development is a valid measure of Fowler's stages.

Other research on faith development also offers some tentative support of the FDI's validity. Fowler's (1981) own research observed a general age/stage relationship, especially noticeable in the lower stages. He reports a slightly skewed bell distribution of the various stages: 6.4% at stage 1, 7% at stage 2, 24% at stage 3, 24.8% at stage 4, 7 % at stage 5 and .3% (1 person) at stage 6. The remaining percentages were people in stage transitions. Age ranged from 3 1/2 to 84 years. Progress through the stages was not inevitable and adults are reported as stabilized at all stages from 2 to 6. These relationships seem to hold in other studies as well (Moseley & Broken-brough, 1988; Tamminen, 1994).

The FDI is used with a wide variety of populations. There are published studies using the FDI with populations in the U.S. (Fowler, 1981), Israel (Snarey, 1991), and Finland (Tamminen, 1994); these studies included Christians (Fowler, 1981), Jews (Fowler, 1981), Buddhists (Furushima, 1985), and non-theistic Israelis (Snarey, 1991). The Manual reports that Stages of Faith (which includes the FDI as an appendix) has been translated into German, Korean, and Portuguese, and that reports of faith development research have appeared in Danish, Swedish, Finnish and Indonesian (Moseley et al., 1993).

FDI Reliability. The Manual reports results from efforts to standardize the administration and scoring of the interviews (Moseley et al., 1993). When the current scoring protocol is used, Fowler and his colleagues report an inter-rater reliability for trained scorers as being "in the range of 85 to 90 % agreement" (Fowler, 1981, p. 314; cf. Moseley et al., 1993). Snarey (1991) reported an inter-rater reliability of .88 for his study.

Conclusions. Overall, the FDI is the best validated of the instruments designed to measure Fowler's faith stages. The questions in the FDI elicit answers that encompass all of the structural elements Fowler proposes, as well as the cognitive, relational, and affective dimensions of faith. Its validity is clearly adequate for research purposes (cf. Snarey, 1991). Copies of the FDI can be found in several places: Burris (1999), Fowler (1981), and Moseley et al. (1993).

However, use of the FDI is not without problems. The enormous amount of time needed to administer (and then transcribe) an FDI tends to prohibit its use in research (Moseley et al., 1993). The need for some level of clinical sensitivity and training in scoring are other practical deterrents to its use. In addition, those using it should consider its operational limitations. Despite its attempts to attend to affective and relational dimensions of faith, the nature of the FDI is better attuned to uncovering the cognitive dimensions of faith, perhaps because the FDI attempts to cover these latter structures via reflections (cf. Broughton, 1986: Fernhout, 1986; Jardin & Viljoen, 1992). The heavy focus of the current scoring guide on the structural elements of faith also tends to favor the cognitive elements of faith development for stage determination. While the relational aspects of faith development receive some attention in the structures labeled "bounds of social awareness" and "locus of authority," Fowler's theorizing about the affective elements does not translate well in the FDI scoring of the various structures (Streib, 2005). How one might overcome the problem of cognitive inquiry about affect is not an easy issue to solve. Perhaps the addition of some sort of "relationship" structure--similar to the hierarchical arrangement of stages in object relatedness that Kernberg (1976) outlines--would improve the FDI's ability to measure faith stages. Or perhaps studies comparing different scoring procedures would better account for the relational and affective dimensions of faith (cf. Moseley et al., 1993).

In addition, the current method of computing an arithmetic average tends to flatten out scores, thus covering over any significant fluctuations that might be present across structures. For instance, there is the potential that someone might have a score of "2" on one structure but average a "4" on another. These in turn indicate an average of "3" which is misleading in terms of the actual level of the person's answer. Current scoring yields a continuum of scores, rather than scores that cluster to specific stage levels (Fowler, 1981; cf. Batson, Schoenrade & Ventis, 1993). While it is unlikely that one would have huge disparity when all seven structures are assessed, the spread that can sometimes be present is lost (Dreidger, 1998). Similarly, Moseley et al.'s (1993) observation that there are differences in the way people strongly disposed to one structural aspect (e.g., someone predominantly interpersonal vs. someone predominantly cognitive) engage the other structural aspects raises a question as to whether some sort of weighting of the structures ought to be part of the scoring. While the current procedure is adequate for research (cf. Snarey, 1991), the current scoring protocol also obscures important information (Streib, 2005). A recently revised Manual for Faith Development Research (Fowler, Strieb, & Keller, 2004) suggests reporting "profiles" in addition to averages as a way to show discrepancies across stages that get obscured when reporting an averaged faith stage score. This option in scoring also is designed to ameliorate the preferential treatment of cognitive material by making differences between cognitive and affective material more visible to researchers comparing FDI scores. It is too early to know how this option will impact research using the FDI, but is an attempt to address some of the criticisms of the FDI as a research instrument.

The Faith Styles Scale (FSS)

FSS Structure. Barnes, Doyle and Johnson (1989) have produced the most widely used questionnaire for assessing faith stages (Timpe, 1999). Assuming "face plausibility" to Fowler's theory, Barnes et al. constructed a forced option instrument for use with adults; it consists of nine paired statements designed to assess the kinds of thinking characteristic of stages two through five. (Stages 1 and 6 were omitted because they do not appear with frequency in adults). The statements were written to reflect specific qualities that Fowler attributed to each of the chosen stages. While these characteristics are not consistently identifiable in terms of Fowler's seven structures, they are reflective of broad stage characteristics, such as concern for group standards for stage three or a concern for connecting all aspects of reality for stage four. A sample statement indicative of Fowler's stage two is "those who do what God wants are given special rewards," while "love of neighbor requires being open to new ideas and values" is a statement considered indicative of stage five. Those taking the survey respond to Likert-like options (A a ? b B) indicating agreement with one of the two paired statements. The pairings require choices between statements reflecting different stages. There are five statements each reflecting stages 3 through 5 and three statements that reflect a preference for stage two. Because the lack of longitudinal data and a sufficiently diverse age group prevent any speculation on whether their respondents experienced stages sequentially, Barnes, et al., identify the differences they found as faith "styles" rather than "stages." However, their statements consciously attempted to reflect Fowler's stage differences.

FSS Administration. The FDS is generally given as a paper-pencil instrument, though it can be administered orally (Timpe, 1999). The nature of the statements requires some reflective ability on the part of the responder. Thus, its use with young children or those not given to serious reflection about religion will be limited. No special training is needed to administer the FDS; scoring also is an easily learned procedure.

FSS Scoring. Respondents who choose statements reflective of a given stage four out of the five possible times (two out of three for stage two) are identified as being at that faith style (stage). Either choice for a given statement (e.g. A or a) is scored as agreement.

FSS Validity. Although the statements in the Barnes et al. (1989) questionnaire are a conscious attempt to reflect stage characteristics, one cannot be sure that it actually measures faith "stages." Two findings support the possibility that their instrument might be distinguishing stages. Barnes et al. report that 65% of their respondents consistently chose (four out of five options for most styles) the same stage identifying answer, while another 29% spread their answers only to the two adjacent styles. These are results one would anticipate from structural stages. In addition, Barnes et al. found a positive correlation between different faith styles and how literally vs. symbolically people interpret their religious beliefs; people with advanced faith styles preferred symbolic interpretations. However, contrary to Fowler's data with the FDI, Barnes et al. also found much higher levels of stage 5 and stage 4 faith in their sample, with more stage 5's than any other. This may be due to sampling selection-a higher degree of their sample were people with master's and doctorate degrees--or it might reflect a problem with the instrument. Barnes et al. found that education level was positively related to the level of faith style.

The Barnes et al. (1989) scale has seen moderate use in other studies (Timpe 1999). Sometimes the results with the Barnes et al. questionnaire confirm findings using the FDI; sometimes they do not. Timpe reports a study that found a moderate correlation between the FSS and the Defining Issues Test (Rest, 1979); a finding one would anticipate given the relationship between Fowler's (1981) and Kohlberg's (1976) theories. Compliment (1998) found that people with different faith styles use different coping strategies (also a finding one would anticipate from Fowler given that faith stages involve variant ways of making sense of the world), but he also found differences in personality characteristics and faith style (a more complex finding that requires further investigation of whether faith stage may involve personality variables in addition to the seven structures. Cf. Moseley et al. 1993). Using the Barnes et al. scale, James and Samuels (1999) found no significant relationship between faith style and I/E religious orientation (Gorsuch & McPherson, 1989), a finding at some variance with studies by Chirban (1981) and Driedger (1998) using the FDI. However, James and Samuels use a slightly different measure of Intrinsic religiosity than these latter researchers who found a positive relationship between faith stage and Intrinsic religiosity. Wroblewski's (1996) use of the Barnes et al. scale found no relationship between faith style and a measure of psychosocial development, a finding that contradicts similar investigations using the FDI (Driedger, 1997; Snarey, 1991). Wroblewski also reports additional factor analysis with the Barnes et al. scale that indicates it is measuring several factors. While this finding does not settle the question of whether this scale is measuring Fowler's constructs, the presence of multiple factors strongly suggests that this scale measures something different than the FDI measures.

FSS Reliability. Barnes and his colleagues (1989) report no reliability studies. Leak, Loucks, & Bowlin (1999) reported a coefficient alpha of .53 when they assessed internal consistency of the Barnes et al. scale items, and an average inter-item correlation of .11. James and Samuels (1999) report a three month test-retest reliability of .62 (Spearman Rho) on the FSS.

Conclusions. The brevity and ease of use of the FSS makes it attractive for research. Copies can be found in Timpe (1999) and Barnes et al. (1989). However, there is a serious drawback to its use: one cannot be sure it actually measures Fowler's stages. In fact, the discrepancies between findings using the FDI and the FSS strongly suggest it is not. The failure of any of those using the FSS in research to compare FSS scores with FDI scores is a serious deficiency in the research on this instrument.

The Faith Development Scale (FDS)

FDS Structure. A second questionnaire designed to measure Fowler's way of conceptualizing faith development was produced by Leak et al. (1999). It also is a forced-choice, paired item scale, consisting of eight pairs of statements, with one statement representing a "more mature" statement of faith development. The statements are designed to explicitly reflect Fowler's characterization of the various faith stages, but are not designed to reflect stage levels; instead, the FDS yields an overall "index" of current faith development. A sample pair of items is: (a) "it does not bother me to become exposed to other religions," (b) "I don't find value in becoming exposed to other religions." In this pair, statement (a) reflects greater faith development. The FDS was developed for use with adults and like the Barnes et al. (1989) scale omits statements reflective of stage 1 and 6 given their infrequent occurrence in adults.

FDS Administration. The FDS is generally administered in paper-pencil form, but there is nothing that would prohibit its oral administration. Like the Barnes et al.(1989) scale, it is designed for use with adults, and thus shares the same limits in use with children and those given to little reflection on religion. No special training is needed for administration or scoring.

FDS Scoring. The FDS yields an overall index of faith development rather than an indication of "stage" (or even style). Scoring is done by counting the number of times a respondent chooses the more "mature" statement over the less mature. The potential range is 0-8. Thus, it is important to remember that a score of 4 does not indicate faith development characteristic of Fowler's stage 4, but simply that a person chose the more mature statement four out of eight possible times.

FDS Validity. Leak and his colleagues provide information from several validation studies on the FDS. Items in the instrument were evaluated for content validity by expert judges who teach Fowler's theory. Leak et al. (1999) correlated scores on their faith development scale (FDS) with various measures of religious maturity, belief, and experience. As anticipated, they found positive correlations between high scores on the FDS and the Quest scale (Batson et al., 1993), people's openness to change in the church, a measure of psychological openness (Costa & McCrae, 1985), a measure of Fowler's higher stages, peer ratings of more mature faith, and self-report that one's beliefs were formed from personal experience rather than the influence of others. In addition, there were negative correlations between FDS scores and Extrinsic religiosity (Allport & Ross, 1967), having one's beliefs influenced more by others, a measure of Fowler's lower stages, and peer ratings of lower faith development. Scores on the FDS were not related to participation in overt religious acts or to a measure of social desirability. These correlations reflect the directions one would anticipate from Fowler's theory. Contrary to anticipated directions, Leak, et al. found that scores on the FDS were not correlated with Intrinsic religiosity, nor with a psychological measure of agreeableness (Costa & McCrae, 1985). Longitudinal studies on the FDS are underway (Leak et al., 1999). Overall, the correlations of the FDS with other measures of religiosity are clearly in the anticipated directions. Though mostly in the moderate range, the correlations are strong enough to merit use of the FDS for research.

FDS Reliability. Leak, et al. (1999) report several items on the reliability of the FDS. They report a small sample (n = 11) test-retest reliability of .96 after five weeks. After deleting items using iterative item analysis, the internal consistency for the remaining eight paired items had a coefficient alpha of .71. A factor analysis of the FDS indicated only one dominant factor.

Conclusions. While the FDS is an acceptable, short global measure of faith development, it is not a measure of Fowler's faith stages. A major deficit in the validity studies with the FDS is the absence of any correlation of its scores with scores from an FD1. As a global measure, it might serve as a supplement or alternative to the other widely used measure of religious maturity, the I/E scales (Allport & Ross, 1967), but will not serve those desiring to measure Fowler's faith stages as a variable. The FDS can be found in Leak et al. (1999).

Other Measures of Faith Development

There are several other attempts to design briefer instruments to measure Fowler's faith stages noted in the literature. Most of these have not been used in research beyond the initial reports. They can be found in the original citation source except where noted.

Green and Hoffman (1989). These researchers devised a questionnaire in which they asked respondents to select one of four statements as best describing their views toward religious faith. Each of the four statements was crafted to reflect elements of stages 2 to 5 as described by Fowler. Green and Hoffman report no validity or reliability information about their instrument; also absent is any correlation of their measure with scores using the FDI. Using this instrument, Green and Hoffman found support for Fowler's notion that faith stage is correlated with how one draws social boundaries for in-groups/out-groups. However, they also reported high levels of stage 5 respondents in an undergraduate population. This finding is problematical given Fowler's findings regarding the rarity of the upper stages and his argument about the hierarchical nature of the stages. These findings, along with the absence of any further studies with Green and Hoffman's measure, suggest that it is not an adequate measure of faith stages.

Rose (1991). Rose designed a forced choice questionnaire to which respondents indicated their level of agreement with belief statements about the faith issues raised by a series of four brief vignettes. Using Rest's (1979) model for the Defining Issues Test (DIT, a shorter measure for Kohlberg's stages), Rose had subjects responded on a nine point Likert scale to questions about four scenarios. Scores on this instrument correlated moderately (r = .32) with estimates of faith stages by trained pastors. The reliability estimates (coefficient alphas) were high for items discriminating stages 2 (.75) and 4 (.78) but low for stages 3 (.16) and 5 (.19), indicating some problems with the internal consistency of the measure on these two stages. Some of the criterion measures Rose used to assess validity also ran counter to his hypotheses and Fowler's theory (e.g., faith stage, rather than being generic, was connected to traditional religious acts; fundamentalist protestants had higher faith stages than mainline protestants). Factor analysis indicated multiple factors being measured, which would be inconsistent with the notion that faith is a coherent or unified structure.

Clearly, in its current form there are serious flaws in Rose's instrument. Some findings may be the result of inadequate criterion measures (e.g., church attendance as the measure of traditional religious behavior), but some are the result of the low internal consistency for two of the stages. In devising his instrument Rose noted that he used Fowler's description of three broad dimensions of faith (relational, cognitive, affective), rather than the seven structures around which the FDI is constructed. This choice may account for some of the problems with his instrument. Because of its similarity to the DIT, an instrument with a moderate level of validity as a measure of Kohlberg's stages, Rose's instrument, with revisions to certain items that would increase the internal consistency for stages three and five, could become a useful shorter measure of Fowler's stages.

Hoffman (1994). Hoffman has developed the only alternative measure of Fowler's faith stages that reports correlations with scores from the FDI. Allowing a half-stage variation, he found an 83.8% agreement rate between faith stage level as determined by FDI and his "Faith Development Essay" in a sample of twelve undergraduate students at a Catholic college. Hoffman's instrument requires subjects to write answers around a cluster of three questions (e.g., how do you choose friends, how do you know something is right, how do you account for suffering). Answers are scored according to a guide similar to the one Fowler provides in the Manual (Moseley et al. 1993). Hoffman reports that it takes 30-45 minutes to complete the essays. Hoffman found that his essay instrument correlated with the FDI (r = .59) and also with the DIT (r = .41), confirming a relationship but not identity between faith development and moral judgment. He found no differences between gender and faith development although the small number in his sample makes all his findings tentative. Although there has been no further reported use of this measure, with further studies on its validity and reliability, it holds potential as a valid shorter measure of Fowler's faith stages.

Clore (1997). Clore developed a 30 item questionnaire designed to account for the complexity of Fowler's stages (five "stages" across six aspects or structures) and has used this in one subsequent study (Clore and Fitzgerald, 2002). In both samples, factor analysis yielded four factors which he argued corresponded to (though in combined form) and supported Fowler's argument regarding different aspects or structures to faith. However, his instrument did not distinguish different "stages" of faith, but rather four "ways" of faith that operate in a more interactive fashion than Fowler's "stages" would allow. While these findings indicate that Clore is measuring some aspects of faith development, they also indicate that his instrument is not measuring Fowler's faith stages.

Leak et al. (1999). Leak and colleagues report on the development and preliminary validation of an instrument to measure Fowler's faith stages used in their validation studies of the FDS. They constructed four "scales" that provide a measure of Fowler's stages 2-5. The scales consists of 30 items, written to explicitly reflect Fowler's descriptions of the various stages. The items were validated for content by judges, and reduced from a larger pool through item analysis. The alpha reliability estimates for the final scales were .55, .80, .72, and .55 for stages 2-5, respectively. Leak, et al. report a preliminary study (Jugel, 1992) on content and criterion validity of the scales. The results using this measure showed that the four faith stages related as expected to measures of conventional religiousness and openness to religious issues. No further research with this instrument has appeared in the literature; it is unpublished.

Abbreviated Interviews. Several researchers have tried to assess Fowler's faith stages by asking fewer questions of the type used in the FDI (Gorman, 1977; Swenson, Fuller & Clements, 1993; Watt, 1997). Only Swenson et al. (1993) report any reliability or validity data. They report that using a prepared answer guide for assigning various stages, and a scale for scoring answers to their five questions, independent judges were able to assess faith level with a fair degree of accuracy. However, in a subsequent study, Clements (1998) questioned the reliability of this shorter measure. Das and Harries (1996) report no validity or reliability data on their use of written answers to only two questions (are you religious; what is the meaning of life) to assign a faith stage. However, the utter brevity of their measure makes validation as a measure of faith stages unlikely. None of these shorter interviews have been correlated with stage determinations using the longer FDI.

CONCLUSIONS

Those wishing to do research in faith development theory have a variety of instruments to sort through, each with its limitations and attractions. It is clear that they are not all measuring the same thing. All of these instruments, including Fowler's FDI, neglect some aspects of faith development theory. On the other hand, each instrument captures aspects of Fowler's conceptualization of faith development. The FDI remains the best validated, and most reliable of the measures. Its limitations include the enormous amount of time needed to administer the FDI and a scoring protocol that neglects the relational and affective dimensions in FDI answers.

Their brevity and ease of use in research make the shorter measures attractive. However, the central question regarding the shorter measures is the extent to which they are valid measures of faith "stages" (some, such as Leak et al. [1999], have foregone any claim to measure stages). Those who use these instruments must contend with McDargh's (2001) critique that any shorter measure will likely omit the very aspects of faith that are unique to Fowler's theory. Given Fowler's argument regarding the complexity of faith (and the concomitant complexity of the FDI to try to capture this), the general failure to compare the findings obtained with these measures against stage determinations using the FDI is a limitation of these instruments that needs addressed in future research.

Finally, one should note that it is not always easy to distinguish between an instrument problem and a theoretical one. For instance, Leak et al. (1999) found no significant correlations between faith stages and the other factors of the NEO personality inventory. This is surprising given that Leak and his colleagues argue that the FDS measures religious maturity. As they note, one would anticipate a positive correlation of FDS scores with high scores on the agreeableness factor of the NEO since this factor measures such things as helpfulness and forgiveness (Costa and McCrae, 1985)--qualities usually associated with religious maturity. Is the lack of correlation between agreeableness and FDS score to be interpreted as a problem in Leak et al's instrument or with faith development theory? That is, is faith development theory too cognitive, omitting aspects like forgiveness that generally are associated with deep religious commitment (Fernhout, 1986)? This is an important question and one that points to the connection between evaluating faith development measures and evaluating faith development theory. While an adequate answer to this question lies beyond the scope of this article, those who wish to evaluate the adequacy of the faith development measures reviewed here must remember that answers regarding the reliability and validity of the instruments do not necessarily answer these larger questions regarding theoretical adequacy (Clore & Fitzgerald, 2002; Streib, 2005).

REFERENCES

Allport, G. & Ross, J. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5, 432-443.

Astley, J. & Francis, L. (Eds.). (1992). Christian perspectives on faith development. Leominster, England: Gracewing.

Barnes, M., Doyle, D., & Johnson, B. (1989). The formulation of a Fowler scale: an empirical assessment among Catholics. Review of Religious Research, 30(4), 412-420.

Batson, C., Schoenrade, P. & Ventis, L. (1993). Religion and the individual. New York: Oxford University Press.

Broughton, J. (1986). The political psychology of faith development theory. In C. Dykstra & S. Parks (Eds.), Faith development and Fowler. Birmingham, AL: Religious Education Press.

Burris, C. (1999). Faith development interview guide (Fowler, 1981). In P. Hill & R. Hood (Eds.). Measures of Religiosity. Birmingham, AL: Religious Education Press.

Chirban, J. (1981). Intrinsic and extrinsic religious motivation and stages of faith (Doctoral dissertation, Harvard Divinity School, 1981).

Clements, R. (1998). Intrinsic religious motivation and attitudes toward death among the elderly. Current Psychology: Developmental, Learning, Personality, Social, 17(2-3), 237-248.

Clore, V. (1997). Faith development in adults: Scale of measurement and relation to attachment. (Doctoral dissertation, Wayne State University, 1997). Dissertation Abstracts International, 58(11), 6256B.

Clore, V. & Fitzgerald, J. (2002). Intentional faith: An alternative view of faith development. Journal of Adult Development, 9, 97-107.

Compliment, B. (1998). The other victims of scandal in the church: A study of the effects of clergy sexual misconduct on ministerial identity, coping responses and style of faith development on non-offending priests and laity in the Roman Catholic church. Dissertation Abstracts International, 58 (08), 4440B.

Costa, P. & McCrae, R. (1985). The NEO personality inventory manual. Odessa, FL: Psychological Assessment Resources.

Das, A. & Harries, B. (1996). Validating Fowler's theory of faith development with college students. Psychological Reports, 78(2), 675-679.

Day, J. (2001). From structuralism to eternity? Re-imagining the psychology of religious development after the cognitive-developmental paradigm. The International Journal for the Psychology of Religion, 11(3), 173-183.

Driedger, S. (1998). Relationships among faith development, ego development, and religious orientation in HIV+ individuals. Dissertation Abstracts International, B 58 (12), p. 6842.

Dykstra, C. & Parks, S. (1986). Faith development and Fowler. Birmingham, AL: Religious Education Press.

Erikson, E. (1968). Identity, youth and crisis. New York: WW. Norton.

Fernhout, J. (1986). Where is faith? Searching for the core of the cube. In C. Dykstra & S. Parks (Eds.), Faith development and Fowler. Birmingham, AL: Religious Education Press.

Fowler, J. (1981). Stages of faith. San Francisco: Harper and Row.

Fowler, J. (1990). Faith development research. In R. Hunter (Ed.), Dictionary of Pastoral Care and Counseling. Nashville: Abingdon.

Fowler, J. (1992). Forward. In J. Astley & L. Francis (Eds.) Christian perspectives on faith development. Leominster, England: Gracewing.

Fowler, J., Streib, H., & Keller, B. (2004). Manual for faith development research. Atlanta: Center for Research in Faith and Moral Development.

Furushima, R. (1985). Faith development in a cross-cultural perspective. Religious Education 80(3), 414-420.

Gorman, M. (1977). Moral and faith development in seventeen-year-old students. Religious Education, 72 (5), 491-504.

Gorsuch, R. & McPherson, S. (1989). I/E measurement: I/E -revised and single item scales. Journal for the Scientific Study of Religion, 28, 348-354.

Green, C. & Hoffman, C. (1989). Stages of faith and perceptions of similar and dissimilar others. Review of Religious Research 30(3), 246-254.

Hill, P. & Hood, R. (1999). Measures of religiosity. Birmingham, AL: Religious Education Press.

Hoffman, D. (1994). Nurturing spirituality: An assessment of the cognitive, moral judgment, and faith development of first year college students participating in a developmental retreat program. (Doctoral dissertation, University of Maryland, 1994). Dissertation Abstracts International, 56, 108A.

James, B. & Samuels, C. (1999). High stress life events and spiritual development. Journal of Psychology and Theology 27(3), 250-260.

Jardin, M. & Viljoen, H. (1992). Fowler's theory of faith development: An evaluative discussion. Religious Education, 87(1), 74-85.

Jugel, L. (1992). Religious and personality variables: Assessment of religious maturity. Unpublished masters thesis, University of Nebraska-Lincoln.

Kernberg, O. (1976). Object relations theory and clinical psychoanalysis. New York: Jason Aronson.

Kohlberg, L. (1976). Moral stages and moralization. In T. Lickona (Ed.), Moral Development and Behavior, New York: Holt, Rinehart and Winston.

Leak, G., Loucks, A. & Bowlin, P. (1999). Development and initial validation of an objective measure of faith development. International Journal for the Psychology of Religion 9(2), 105-124.

Loevinger, J. (1976). Ego development. San Francisco: Jossey-Bass.

Loevinger, J. (1979). Construct validity of the sentence completion test of ego development. Applied Psychological Measurement 3(3), 281-311.

McBride, A. (1976). Reaction to Fowler: Doubts about procedure. In T. Hennessy (Ed.), Values and Moral Development. New York: Paulist.

McDargh, J. (2001). Faith development theory and the postmodern problem of foundations. The International Journal for the Psychology of Religion, 11(3), 185-199.

Mischey, E. (1981). Faith, identity, and morality in late adolescence. Character Potential: A Record of Research 9, 175-191.

Moseley, R. & Brockenbrough, K. (1988). Faith development in the pre-school years. In D. Ratcliff (Ed.), Handbook of pre-school religious education. Birmingham, AL: Religious Education Press.

Moseley, R., Jarvis, D., Fowler, J. & DeNicola, K. (1993). 1993 Manual for faith development research. Atlanta: Center for Research in Faith and Moral Development.

Nelson, C. & Aleshire, D. (1986). Research in faith development. In C. Dystra & S. Parks (Eds.), Faith development and Fowler. Birmingham, AL: Religious Education Press.

Parks, S. (1991). The North American critique of James Fowler's theory of faith development. In J. Fowler, K. Nipkow, & F. Schweitzer (Eds.), Stages of Faith and Religious Development. New York: Crossroad.

Piaget, J. (1970). Piaget's Theory. In P. Mussen (Ed.), Charmichel's Manual of Child Psychology, 3rd Ed., Vol. 1. New York: John Wiley and Sons.

Rest, J. (1979). Development in judging moral issues. Minneapolis: University of Minnesota Press.

Rose, D. (1991). An instrument to measure four of James Fowler's stages of faith Development. (Doctoral dissertation, California School of Professional Psychology, 1991). Dissertation Abstracts International, 52 (09), 4997B.

Santrock, J. (1999). Life-span development (Seventh edition). Boston: McGraw-Hill.

Shulik, R. (1979). Faith development, moral development and old age: An assessment of Fowler's faith development paradigm. (Doctoral dissertation, University of Chicago, 1979). Dissertation Abstracts International, 40, 2907B.

Snarey, J. (1991). Faith development, moral development, and non-theistic Judaism: A construct validity study. In W. Kurtines & J. Gerwitz (Eds.), Handbook of Moral Development and Behavior: Vol. 2, Research. Hillsdale, NJ: Lawrence Erlbaum.

Streib, H. (2001). Faith development theory revisited: The religious styles perspective. The International Journal for the Psychology of Religion, 11(3), 143-158.

Streib, H. (2005). Faith development research revisited: Accounting for diversity in structure, content, and narrativity of faith. International Journal for the Psychology of Religion, 15 (2), 99-121.

Swenson, C., Fuller, S. & Clements, R. (1993). Stage of religious faith and reactions to terminal cancer. Journal of Psychology and Theology 21(3), 238-245.

Tamminen, K. (1994). Comparing Oser's and Fowler's developmental stages. Journal of Empirical Theology, 7(2), 75-112.

Timpe, R. (1999). Faith development scale (Barnes, Doyle & Johnson, 1989). In P. Hill & R. Hood (Eds.), Measures of Religiosity. Birmingham, AL: Religious Education Press.

Watt, S. (1997). Identity and the making of meaning: Psychosocial identity, racial identity, womanist identity, self-esteem and the faith development of African-American college women. Dissertation Abstracts International, 58(06), 2091A.

Wroblewski, C. (1996). The relationship between religious faith maturity in Roman Catholic adults and their perceptions of eucharistic liturgy: An empirical study. (Doctoral dissertation, Loyola University, Baltimore, MD., 1996). Dissertation Abstracts International, 57,1473B.

AUTHOR

PARKER, STEVEN: Address: School of Psychology and Counseling, Regent University, 1000 Regent University Drive, Virginia Beach, VA 23464. Title: Associate Professor. Degree: Ph.D. Specializations: Religious and Spiritual Development, Integration of Theology and Personality Theory.

STEPHEN PARKER

Regent University

Correspondence regarding this article may be sent to Steven Parker, School of Psychology and Counseling, Regent University, 1000 Regent University Drive, Virginia Beach, VA, 23464.

TABLE 1 Structures of Faith

 The structures of faith include both adaptations and expansions of
concepts from other structural-developmental theorists as well as
concepts unique to Fowler:
 Form of logic: an adaptation of Piaget's (1970) variations in
 cognitive abilities. One develops from a "pre-logical" way of
 comprehending the world to a more concrete and eventually abstract
 logical comprehension.
 Perspective taking: an expansion of Selman's (1980) work on how one
 grows from a one dimensional (egocentric) perspective to the ability
 to see an event or experience from multiple perspectives.
 Form of moral judgment: an adaptation of Kohlberg's (1976) work on
 various stages of moral reasoning abilities.
 Bounds of social awareness: a structure that points to how inclusively
 or exclusively one draws the boundaries of one's social groups.
 Development of this structure is in the direction of increasing
 inclusiveness.
 Locus of authority: a structure that charts the movement from a more
 external to a more internal locus for decision-making and evaluation.
 Form of world coherence: a structure that has to do with both
 self-reflection and level of conscious awareness of one's cognitive
 and evaluative processes. The movement is from tacitly held ways of
 making sense of the world to more conscious, intentionally chosen ways
 of making sense.
 Symbolic function: this structure involves the way symbols are
 perceived and engaged. Movement is from a uni-dimensional engagement
 (symbols have one, often concrete meaning) toward multi-dimensional
 meanings and demythologizing of symbol content.

TABLE 2 Fowler's Faith Stages

 Beginning with childhood, Fowler (1981) charted a seven stage
progression of construing and relating to the "ultimate environment":
 Stage 0: Primal Faith. Actually a "pre-stage" since the various
structures of the subsequent stages are not available for the same
empirical verification as the remaining stages. Fowler sees the mutual
interaction between infant and primary caregiver during the first,
pre-verbal year of life as providing a foundation for faith in the
emergence of basic trust vs. mistrust (Erikson 1968).
 Stage 1: Intuitive-Projective Faith. The acquisition of language marks
the emergence of the first true stage which is characterized by the
abundant imagination of the pre-school child. This emotional and
idiosyncratic faith involves powerful images and a fluidity of thought
not bound by the logic of later cognitive structures (cf. Piaget 1970).
"God" is a powerful creature of the imagination, not unlike Superman or
Santa Claus.
 Stage 2: Mythic-Literal Faith. Maturation evokes a new way of knowing
and engaging the world. The child acquires what Piaget (1970) called
concrete operational thought. This new way of seeing and interacting
with the world and others allows the child to infer intentions and to
perceive continuity to actions; justice is concrete and reciprocal.
These abilities, held together by means of a narrative, give rise to a
faith in which the ultimate environment is inhabited by a cosmic judge
("God") who guarantees a kind of simple, reciprocal fairness.
 Stage 3: Synthetic-Conventional Faith. The emergence of formal
operational thought brings other possibilities for construing and
relating to the ultimate environment. In this stage, meaning-making and
committing to values takes on a more interpersonal dimension not
previously available. Self-identity and faith are closely tied to valued
others, and thought deeply felt, is unexamined. "God" takes on the
interpersonal qualities of a good friend.
 Stage 4: Individuative-Reflective Faith. This stage of faith is
characterized by intentional reflection on one's faith and its influence
on the self. This intense, critical reflection on one's faith (one's way
of making meaning) requires that inconsistencies and paradoxes are
vanquished, which may leave one estranged from previously valued faith
groups. "God" is the embodiment of principles of truth, justice, love,
etc.
 Stage 5: Conjuctive Faith. Mid-life sometimes brings a recognition
that the consistencies of one's more reflective faith have come with a
price; one may have dismissed other (unconscious) dimensions of knowing.
A yearning for a way to bring together the seeming paradoxes of faith
may emerge, along with a desire to enlarge the bounds of social
inclusiveness. Although one does not naively or uncritically accept
contradictions, "God" is seen to include mystery and paradox.
 Stage 6: Universalizing Faith. Finally, Fowler posits a movement
toward a style of "universalizing faith" that seeks inclusiveness while
still maintaining firm and clear commitments to values of universal
justice and love. Critics (Broughton 1986, Parks 1991) have pointed out
that this stage seems to be more content-oriented and less structural
than the previous ones.