The multidimensional structure of the quest construct.
The empirical study of religion as an important social and personal variable has been a viable, if somewhat less than mainstream, pursuit in psychology for several decades. Psychology of religion research increased substantially when Allport and Ross (1967) developed their measure of intrinsic (I) and extrinsic (E) motivation (the Religious Orientation Scale, ROS). The measure was based upon Gordon Ailport's (Ailport, 1950) conceptualization of mature versus immature religious motivation. Allport (1950) proposed his model distinguishing between intrinsic and extrinsic motivation as a way to understand the observed association between prejudice and religious belief. While the intrinsic and extrinsic constructs and the ROS measure have been subjected to extensive review and critique (Donahue, 1985; Gorsuch & McPherson, 1989; Kirkpatrick, 1988), the I/E model of religious motivation has proved to be useful (Kirkpatrick & Hood, 1990) for psychology of religion research.
Batson and his colleagues (Batson, 1976; Batson & Ventis, 1982; Batson, Schoenrade, & Ventis, 1993) provided another advance in psychology of religion with their proposal of a third dimension of religious motivation, which they called religion as Quest. They argued that the Religious Orientation Scale (ROS) did not adequately measure all facets of Allport's original model of mature religion. Specifically, they suggested that the ROS excluded salient facets such as openness, flexibility, tentativeness, and doubt (Batson, Schoenrade, & Ventis, 1993). They argued that these facets were particularly important for understanding the relationship between religion and prejudice. The quest motivation describes individuals for whom religion is an on-going, open-minded exploration of existential questions, acceptance of the fact that many important religious questions do not have clear-cut answers, and a tentativeness of belief that remains open to change as one grows and develops. Batson & Ventis (1982) proposed a six-item measure of quest called the Interactional Scale (IS). The items were formulated to measure the three quest facets: openness, doubt, and questioning. The authors offer the following description of their original scale:
The six items in the scale were written to tap three different aspects of this dialogue. Three items tap a person's readiness to face existential questions without reducing their complexity: "I have been driven to ask religious questions out of a growing awareness of the tensions in my world and in my relation to my world"; "My religious development has emerged out of my growing sense of personal identity"; and "God wasn't very important to me until I began to ask questions about the meaning of my own life." Two items tap a person's self-criticism and perception of religious doubts as positive: "It might be said that I value my religious doubts and uncertainties"; and "Questions are far more central to my religious experience than are answers." One item taps a person's openness to change: "I do not expect my religious convictions to change in the next few years" (reverse scored, p. 432). The construct of quest was formulated to be independent of cither the extrinsic (means) or intrinsic (end) dimensions. Batson and Schoenrade (1991b) summarized several studies that support the independence of their quest measure from the ROS intrinsic and extrinsic dimension.
Since it was originally published, both the quest construct and the Interactional Scale (IS or the Quest Scale) have been subjected to criticism regarding both its reliability and validity (e.g., Donahue, 1985; Hood & Morris, 1985; Kirkpatrick & Hood, 1990), Batson has provided extensive responses to the critics of quest (Batson & Ventis, 1985; Batson & Schoenrade, 1991a, 1991b). Batson & Schoenrade (1991a) conducted a systematic review of research on quest that supports its construct validity. They reviewed studies showing that the Quest scale significantly differentiated among known groups in ways consistent with the theory. In addition, they identify several studies in which quest scores correlated significantly with important social variables such as principled moral reasoning, lower levels of racial and sexual prejudice, and higher levels of helping behavior (p. 426). They ended their review with the following conclusion:
After reviewing the major validity concerns raised about the Quest scale, and considering that, in light of the available empirical evidence, we find the results quite reassuring. The Quest scale does seem to be measuring a dimension of religious orientation distinct from that measured by the Extrinsic or Intrinsic scale, and moreover, this dimension seems to be very much like the one the scale was designed to measure, (p. 427)
Another persistent criticism of the quest measure has been the scales reliability. Almost all of the estimates of the quest scale's reliability have been based on Coefficient Alpha. A number of studies have found that the Interactional scale had very low internal consistency with coefficients between.45 and.50. Gorsuch (1988) suggested that "until studies accrue that use more reliable measures of Quest ... few conclusions can be theoretically meaningful" (p.213). A problem with the reliability criticisms of the Quest scale is that Coefficient Alpha is a statistical measure of internal consistency. It provides an estimate of reliability, but it is not a direct measure of reliability. The gold standard for estimating reliability is the test-retest paradigm. However, test-retest research takes time and resources. The model also assumes the stability of the construct over the test-retest period and the independence of error between the two assessments. Alternatives to the test-retest model have been proposed. These include alternate forms and internal consistency estimates of reliability. Cronbach's Coefficient Alpha is the most widely used internal consistency statistic. It can be readily calculated using most standard statistical packages for any sample with a single administration of a questionnaire using scores on individual items. The value of Alpha is a function of the average of the correlations among the items in a measure. It is also a function of the number of items in a scale. Alpha is an indication of how homogeneous, or internally consistent, a set of items are. It can be used as an "estimate" of test "reliability" under certain assumptions (see Cortina, J.M., 1993; Clark, L.A. & Watson, D., 1995 for a more technical discussion of Coefficient Alpha).
A test can be reliable in the test-retest paradigm and have a low coefficient alpha. Under such circumstances, one would conclude that the test is reliable but consists of a heterogeneous pool of items. The total score derived from a set of heterogeneous items can be shown to be statistically valid in either criterion-related or construct validity paradigms. Such a test would be said to have low item homogeneity (low internal consistency) but high validity. It would not make sense, in light of positive validity evidence, to say a measure with low Alpha is "unreliable."
The challenge in construct validity studies with heterogeneous tests (with low coefficient alpha) is primarily one of conceptual interpretation rather than low reliability. The task in construct validity research is to identify the underlying dimensions (latent variables) that account for the observed relationship between the total score and the other constructs serving as criteria. When heterogeneous sets of items representing related but distinct latent variables are combined into a single score and used to predict other phenomenon, you cannot tell which of the constructs accounts for the observe relationship.
We maintain that this is the potential state of affairs with regard to the quest construct as defined and measured by Batson's original six-item measure, the reformulated 12-item version, as well as two other measures that are examined in the present study. After reviewing numerous studies on the validity of quest, Batson & Schoenrade (1991a) concluded that "in spite of its low internal consistency, the six-item Quest scale appears to be a valid measure of the construct it was designed to measure" (p. 434). They also note that the few studies that report test-retest reliabilities "give reassurance that it (the Quest scale) is reliable" (p. 434). It is a fundamental truth in psychometrics that the reliability of a measure is a necessary but not sufficient condition for it to be valid. If a test is unreliable, it cannot be shown to be valid. On the other hand, if a test is shown to be valid you can assume it is reliable. It is our contention that the 12-item Quest scale has low coefficient alphas because the items are heterogeneous, not because it is unreliable.
It's clear from their earliest descriptions of quest, that Batson and his colleagues envisioned quest as consisting of three related clusters of items or facets: readiness to face existential questions, self-criticism and doubt, and openness to change. They have referred to these as the "three subdimensions" of quest (p. 434). Batson and Schoenrade (1991b) proposed a revision of IS with their primary goal being to improve its reliability as measured by Coefficient Alpha. Their revised scale had 12 items. They reported a principal components analysis of the 12 items that clearly identified three and four factors, respectively, in two different samples. The factors they identified corresponded to the three facets of quest identified in their original formulation: openness to change, asking existential questions, and religious doubts as positive. Three similar components were previously identified for the six-item measure by Watson, Morris, and Hood (1989). They summarized their view on the dimensionality of quest as follows: First, the new 12-item Quest scale is certainly not perfect. To mention just one problem, the split- and cross-loadings of many items when the 12 items are factor analyzed indicates that we do not yet have clear measures of the three hypothesized subdimensions of the quest orientation. This limitation is not especially troubling if one wishes to examine the relationship between the quest dimension and some personal or social variable, but it is troubling if one wishes to consider whether one or another of the subdimensions is primarily responsible for the relationship. At present, factor scores derived from factor analyses like those presented in Table 4 could be used to measure the subdimensions in large samples. More items tapping each subdimension would be needed to provide clearly distinct scales measuring the three subdimensions. (p. 445)
Table 4 Principle Component Loadings for Dudley & Cruise (RMS) Items Dudley & Cruise RMS Items RMSF1 Beliefs Satisfying Open to Change Beliefs satisfying/ open to new RMS1 0.718 info Happy w beliefs/ open to new RMS2 0.716 insight Beliefs true, but could be RMS3 0.682 mistaken Questions/ hesitant to be RMS9 0.552 dogmatic No easy answers, faith RMS4 0.509 developmental process struggled to understand evil & RMS6 0.417 suffering beliefs not certain, act on RMS8 0.409 probably true beliefs same as 5 years ago RMS11 0.37 [R] people relationships as RMS10 0.303 important as God Need certainty religion RMS5 completely true [R] preach, NO involvement in RMS7 politics [R] [R] Item is reversed scored.
Batson has consistently argued for measuring the quest construct as a single score. While there is substantial empirical support for the validity of the quest as a single score, the dimensionality of that construct remains in question. That the total Quest scale scores correlated with other variables in the predicted direction does not mean the Quest construct is a single dimension. We contend that the investigation of sub-dimensions is an important direction for future research on quest motivation. Identifying the items that measure these sub-dimensions is an important first step. That is the purpose of the present study.
Around the same time Batson & Schoenrade (1991b) developed the 12-item Quest scale, two other groups of researchers offered scales purporting to be improvements on the original 6-item IS scale. We also examined the convergence and divergence of items across three different measures of quest.
Altemeyer and Hunsberger (1992) proposed a measure of quest out of their work on authoritarianism, religion, and prejudice. Building on a modification of the IS scale by McFarland (1989), they developed a 16 item measure that they called "balanced quest." It was called balanced because they sought to include negatively keyed "antiquest" items to control for response sets. The negative items were worded to reflect closed-minded faith and beliefs that were not part of Batson's original model. The balanced quest scale has only one item in common with Batson's measure. The authors conducted a factor analysis of the balanced quest items in an item pool that also included 20 religious fundamentalism items. The analysis demonstrated that the quest and RF items were separate dimensions, but it did not shed any light on the possible dimensionality of the quest construct. Our preliminary analysis of their 16-item measure suggests that it is multidimensional (Edwards at al., 2002). We are not aware of any systematic investigation of the dimensions of quest that underlie the revised Batson measure and the balanced quest scale. Such an analysis will shed light on the convergence and divergence of these two measures.
Dudley & Cruise (1990) formulated a third approach to measuring quest that is included in our dimensional analysis. They objected to Batson's theoretical model of quest as independent of intrinsic religion. They sought to create a measure that combined the best of both the intrinsic and quest concepts to assess mature religion. They created a pool of 58 items, including intrinsic, extrinsic and quest related items which were subjected to factor analysis. Separate factors were identified for intrinsic, extrinsic, and quest (mature) religion. They identified 11 items as quest-related in these analyses. They call their measure the Religious Maturity Scale (RMS). It correlated 0.37 with the original IS scale. Since the RMS items were analyzed in combination with other non-quest items, their factor analysis does not provide any definitive information on the dimensionality of the RMS items.
Another group studying the dimensionality of quest reported their findings after our data gathering project was finished. Beck and his colleagues have pursued their study of the dimensions of quest by developing their own measures (Beck, Baker, Robbins, & Dow, 2001; Beck & Jessup, 2004). In their first study, Beck's group developed 10 items designed to measure the sub-dimensions of quest labeled Tentativeness and Change. A principal components analysis of the tentativeness and change items revealed two distinct dimensions. Subscales scores for the tentativeness and change dimensions correlated 0.35 and 0.31 with the IS scale, respectively. A version of Batson's IS scale did not correlate with any of the religious variables in their study. However, both of their quest subscales produced significant correlations with other religious variables. Specifically, tentativeness was positively correlated with intrinsic religiosity and existential well-being. Change was negatively correlated with intrinsic religiousness, existential and spiritual well-being and positively correlated with extrinsic religiousness (Beck, et al, 2001, p.289). These results indicate that Batson's suggestion of combining the separate clusters of quest items into a total score may obscure important relationships with other variables.
Beck & Jessup (2004) proposed a nine dimensional measure of quest, which included both tentativeness and change as sub-dimensions. In this study, tentativeness was not significantly correlated with other religiosity measures. Change was positively correlated with extrinsic religiosity and negatively correlated with intrinsic religiosity, existential well-being, religious well-being, and orthodoxy. The subscale scores of tentativeness and change correlated significantly with the 12-item quest scale at the 0.36 and 0.51, respectively. Results from the Beck group research supports the utility of measuring the sub-dimensions of quest separately. However, the question of how many dimensions are needed remains to be answered.
The present study reports an analysis of the dimensionality of the Quest construct at the item level using items from three quest-related measures: the 12 quest items (B Quest) developed by Batson & Shoenrade (1991b), the 16 balanced quest items (AQuest) developed by Altemeyer & Hunsberger (1992), and the 11 religious maturity items (RMS) Dudley & Cruise (1990). No study has been conducted which examines the dimensionality of quest by including items from three existing Quest-related measures. Prior analyses of quest dimensions have relied on exploratory procedures and examination of the Scree Plot to estimate the number of dimensions. In the present study we employ both exploratory and confirmatory factor analyses. The latter provides a statistical test of the goodness of fit for a specific dimensional solution. Altemeyer and Hunsberger (1992), after reviewing the controversy in the literature regarding quest stated that "it is possible that matters of controversy would be clearer 25 years after Allport and Ross (1967) if constructs had been defined and measured with greater precision" (p. 117). We intend for our analyses of different quest measures to move us closer to a clearer definition and more precise measurement of the quest construct.
Participants and Procedures
The sample consisted of 1024 undergraduate students from 8 Christian liberal arts colleges from across the United States. Subjects were recruited from a variety of courses, mostly in the social sciences, and participation was voluntary. The sample consisted of students from all four years in school, with the largest groups coming from the freshman and senior classes. There were 346 Freshman, 179 Sophomores, 189 Juniors, and 299 Seniors. There were 540 who identified themselves as females, 264 as males with the remaining subjects giving no indication of gender. Each subject completed the booklet containing the various measures in one sitting. We called the booklet the Faithful Change Questionnaire.
Items from three different Quest measures were included in the factor analysis. All of the questionnaire items were rated on a scale from 1 to 6, Strongly Disagree to Strongly Agree. The measures included in this study were as follows:
B Quest. The Batson & Shoenrade Quest Scale consists of 12 items developed to tap three different facets of quest religious motivation: L Readiness to face existential questions without reducing their complexity; 2. Self-criticism and perception of religious doubt as positive; 3. Openness to change. Batson and Schoenrade (1991b) developed the B Quest scale to improve the reliability of the original six item Interactional scale. They added 7 items and dropped one original item resulting in the 12 item revised scale (see Table 1), which had an internal consistency reliability of between.75 and.8L The new scale correlated with the old one.87.
A Quest. Altemeyer and Hunsberger (1992), building on the work of McFarland (1989), developed a 16 item self-report scale of Quest. This scale (see Table 1) uses only one of the Quest items (1) from Batson & Schoenrade (1991b) and three of the four McFarland items (2, 4 & 6). It contains several con-trait-items (2, 3, 5, 8, 10, 11, 13 & 15), which they formulated to create a scale balanced for positively and negatively stated attitudes. Their new scale had an internal consistency reliability of.88. The new scale proved to be a robust correlate with measures of fundamentalism, authoritarianism, and prejudice. We call this scale A Quest in the present study. Since this scale has one item in common with Quest B, there are 15 unique A Quest items.
MLIIDIMENSIONAL STRUCTURE OF QUEST CONSTRUCT TABLE 1 Quest. Iterns from Three Quest Measures QB1 I was not very interested in religion until I began to ask questions about the meaning and purpose of life QB2 I have been driven to ask religious questions out of a growing awareness of the tensions in my world and in my relation to my world QB3 My life experiences have led me to rethink my religious convictions QB4 God wasn't very important for me until I began to ask questions about the meaning of my own life QB5 It might be said that I value my religious doubts and uncertainties QB6 For me, doubting is an important part of what it means to be religious QB7 I find religious doubts upsetting QB8 Questions are far more central to my religious experience than are answers. QB9 As I grow and change, I expect my religion also to grow and change. QB10 I am constantly questioning my religious beliefs QB11 I do not expect my religious convictions to change in the next few years QB12 There are many religious issues on which my views are still changing QA1 It is better for a person's religious beliefs to be firm and free of doubt QA2 You don't find the true religion by studying all the facts in the universe; you find it by praying to God for grace, humility, and enlightenment QA3 Religious doubts allow us to learn QA4 When my religious beliefs are challenged by famine, disease, and other evils in the world, it only makes me believe in God's goodness more fervently than ever. QA5 My religious beliefs may change in the future as I mature and learn QA6 Religion should just be an aspect of a more basic quest to discover the truth about everything, without prejudice and taking nothing on faith QA7 My religious beliefs are far too important to me to be jeopardized by a lot of skepticism and critical examination QA8 The point of life is to search for the truth, with as open a mind as you can, NOT to memorize the "eternal truths" that he been handed down from generation to generation as matters of faith QA9 I am glad my religious beliefs are based upon faith; it would not mean as much to God, and to me, if these beliefs could be "scientifically proven" beyond a doubt. QA10 We were NOT put on this earth to go "searching for the truth, whatever it is", but instead to live our lives according to the revealed word of God QA11 If an honest quest for the truth leads to the conclusion that there is no God, then that is what one must conclude QA12 The human mind is too limited to discover God and the Truth by itself; we simply have to accept the truths that have been revealed QA13 The real goal of religion ought to be to make us wonder, think, and search, NOT take the word of some earlier teachings QA14 When my religious beliefs are challenged by personal unhappiness, or by some clever argument, it just makes me believe stronger than ever QA15 My goal is to discover the truth, even if that means changing my religious beliefs RMS1 My religious beliefs provide me with satisfying answers at this stage of my development, but I am prepared to alter them as new information becomes available RMS2 I am happy with my present religion but wish to be open to new insights and ways of understanding the meaning of life RMS3 As best I can determine, my religion is true, but I recognize that I could be mistaken on some points RMS4 Important questions about the meaning of life do not have simple or easy answers; therefore faith is a developmental process RMS5 I could not commit myself to a religion unless 1 was certain that it is completely true RMS6 I have struggled in trying to understand the problems of evil, suffering, and death that mark this world RMS7 Churches should concentrate on proclaiming the gospel and not become involved in trying to change society or political action RMS8 While we can never be quite sure that what we believe is absolutely true, it is worth acting on the probability that it may be RMS9 I have found many religious questions to be difficult and complex so I am hesitant to be dogmatic or final in my assertions RMS10 In my religion my relationships with other people are as fundamental as my relationship with God RMS11 My religious beliefs are pretty much the same today as they were five years ago QB Batso& Schoenrade (1991b). QA Altemeyer and Hunsberger (1992). RMS Religious Maturity Scale Dudley & Cruse(1990). 38 item is scored on two subscales: FMS-Vertical, total scale. Scale development was informed by
Religious Maturity Scale. Dudley and Cruise (1990) developed the Religious Maturity Scale (RMS) in an attempt to improve on the ROS and Quest as measures of Allport's conception of religious maturity (Allport, 1950). Dudley and Cruise believe that the following statement, of what a mature religionist might say, reflects a better conception of what Allport actually meant by intrinsic religion:
I have studied the evidence available to me and have developed some satisfying answers to existential questions. My religion makes sense to me; more sense than anything else I have considered. Therefore I am committed to believing it and advancing it. However, I realize that it is incomplete. I will continue to study, and as new information becomes available to me as I grow in understanding, it is very possible that I will see things in a somewhat different light. I want to be ready to progress in my understanding when a new piece of "truth' becomes clear to me. In the meantime I will live by the light I have. And while on the basis of my present experience I prefer my religious beliefs over any others, I fully recognize that other people have a right to their beliefs and, given their perspective on truth may be closer to reality than I am. This recognition however, does not make me any less committed to my faith as I now understand it. (Dudley and Cruise, 1990, p. 101)
They defined mature religion as holding contrasting ideas in tension. The contrast is not as Batson suggests one of faith versus doubt but one of commitment versus tentativeness and open-mindedness. The RMS scale attempts to measure a faith that holds in tension both commitment and openness to further growth in religious understanding. The RMS is an 11 item self-report instrument. It has rather low internal consistency estimates of reliability ranging from.58 to.70 (Dudley and Cruise, 1990). The 11 RMS items are shown in Table 1
Religious Fundamentalism. Paloutzian (1996) has noted that there has been an advance in the measurement of fundamentalism. In the past, fundamentalism has been measured by the content of one's beliefs. However, this makes it difficult to differentiate between fundamentalism and orthodoxy. Altemeyer and Hunsberger (1992) define fundamentalism as, The belief that there is one set of religious teachings that clearly contains the fundamental, basic, intrinsic, essential, inerrant truth about humanity and deity; that this essential truth is fundamentally opposed by forces of evil which must be vigorously fought; that this truth must be followed today according to the fundamental, unchangeable practices of the past; and those who believe and follow these fundamental teachings have a special relationship with the deity, (p. 118)
This method of defining fundamentalism is hypothetical neutral with respect to belief content and applies equally well to fundamentalism in differentreligions. Altemeyer and Hunsberger (1992) have developed a 20 item self-report measure of fundamentalism (RFS). The scale was found to be uncorrelated with a measure of Christian Orthodoxy. The RFS was included in the present study.
Several other measures of religious functioning were included in the Faithful Change Questionnaire. This provided an opportunity to determine how the various facets of quest correlated with these measures. The measures used in the present analyses are described below.
Faith Maturity Scale. The FMS was developed by Benson, Donahue, and Erickson (1993). It consists of 38 items scored on two subscales: FMS-Vertical, which measures experience of God and FMS- Horizontal which measures attitudes about personal, social, and relationship issues. Rather than measuring faith itself, the FMS focuses on what Benson, Donahue, and Erickson (1993) describe as "the degree to which a person embodies the priorities, commitments, and perspectives characteristic of vibrant and life-transforming faith, as these have been understood in mainline Protestant traditions (p. 14)." Thus, in this model, faith is a way of living, not just knowledge of or adherence to doctrine, dogma, or tradition.
At the core of the FMS is an understanding of faith as having "vertical" and "horizontal" dimensions, with faith maturity being the integration of the two (integrated faith). The vertical dimension emphasizes the self and its relationship to God or the divine, or the inward journey. The horizontal dimension emphasizes obligation and action on the human plane through acts of service and justice, or the outward journey.
Spiritual Experience Index. Genia (1991) developed the SEI to measure Western theistic spiritual maturity. The original version consisted of 38 items with coefficient alphas from 0.87 to 0.92 for the total scale. Scale development was informed by faith development theory and definitions of spiritual maturity by a number of researchers and theorists. Genia (1997) revised the SEI and reduced it to 23 items. She defined spiritual maturity as a balance of Spiritual Support (SSS) from the divine and Spiritual Openness (SO) defined as an open and inclusive approach to faith. Genia reported that regression analyses suggested that SO is distinct from quest. Edwards et al. (2001) had identified a subset of the SOS items that appeared to be similar in content to those in the quest construct. In the present study, the SSS and SOS scales were correlated to the quest measures.
Spiritual Assessment Inventory. The SAI was developed by Hall & Edwards (1996; 2001) as a multidimensional, theistic measure of spiritual maturity based on object relations and attachment theories. It consists of 51 items scored on 5 subscales: Awareness (A), which assesses one's awareness of God's presence and communication; Realistic Acceptance (RA), which assesses one's capacity to work through difficult experiences with God; Disappointment (D), which assesses disappointment with God; Grandiosity (G), which assesses an inflated view of oneself presumed to be a defense against feelings of inadequacy; and Instability (I), which assesses instability in relationship with God. The SAI has exhibited its theoretically derived factor structure across numerous studies (Hall &: Edwards, 1996, 2002). In addition, construct and incremental validity have been demonstrated in numerous studies (Hall & Edwards, 1996, 2002).
Spiritual Maturity Index. The Spiritual Maturity Index (SMI) is a 30-item instrument developed by Ellison (1983) to assess spiritual maturity within a theistic, and specifically Christian, religious framework. I he items generally assess Christian faith being central to one's life (e.g., "My relationships with others are guided by my desire to express the love of Christ."). Several items are reverse scored (e.g., "I don't regularly have times of deep communion with God in personal (private) prayer."). Although there is some evidence that the SMI may be factorially complex (Edwards, Slater, Hall, Oda, & Eck, 2001), we used a total score for all 30 items as intended by the original development of the SMI.
Spiritual Well-Being Scale. The Spiritual Weil-Being Scale (SWBS; Ellison, 1983; Ellison & Smith, 1991) is a 20-item self-report scale in which participants rate item endorsement on a 1 to 6 Likert scale. The SWBS has two 10-item subscales assessing Religious Well-Being and Existential Well-Being. The Religious Well-Being subscale is intended to assess one's perception of the quality of their relationship with God. Alternatively, the Existential Well-Being subscale assesses one's perception of having a sense of life meaning and satisfaction with life. Although some questions have been raised concerning the factor structure of the SWBS (Scott, Agresti, & Fitchett, 1998), it has generally demonstrated good reliability and validity coefficients (Bufford, Paloutzian, & Ellison, 1991). Specifically, the SWBS has demonstrated good convergent validity with other measures of general well-being (Bufford et al., 1991).
Each of the three quest measures were subject to a principal components analysis (PCA) of their separate item pools. Each analysis followed the same procedure. The plot of the eigenvalues for each set of items was examined using Catell's Scree Test for determining the number of factors to extract. Determining the number of factors to extract from a given set of data in PCA is not an exact science. The Scree Test provides only an approximation based on the relative change in eigenvalues, which must be determined from a plot. An estimate of the number of components is provided by locating a clear change in the plot from a rapidly descending line to a gradually sloping horizontal line. Once we determined the number of components to extract, each solution with more than one factor was subject to oblique rotation. The goodness of fit of each rotated solution was tested with the Confirmatory Factor Analysis described by Asparouhov & Muthen (2009). The goal of the within-scale analyses was to estimate the dimensionality of each measure.
We then conducted an exploratory PCA on a pool of selected quest items from all three scales. Items that had loadings less than 0.4 or loaded on more than one within-scale factor were omitted from the component analysis of the combined items. Deleting items in this manner does not mean they are not useful for measuring quest. It just means they may not be pure measures of a sub-dimension of quest in their current form. In fact, Batson & Schoenrade (1991) acknowledged that they wrote some of the items for their revised quest scale trying to capture the complexity of the quest construct.
The Scree test was again used to estimate the number of components to extract. The components were subject to oblique rotation. CFA goodness of fit statistics were calculated for the number of components extracted. Quest sub-dimension scale scores were computed for each of the components using the average of student responses to the items with pattern coefficients above 0.40.
The correlation among these component scales and their correlations with the original quest scale scores were calculated. The original and quest component scales were correlated with several other religious variables including Religious Fundamentalism (RF). Finally, differences among students classified by year in school on the derived quest sub-dimension scales were examined using a mixed-model repeated-measures ANOVA followed by a series of within-measures, pair-wise comparisons between means by year in school on each scale when the main effects were significant.
Table 2 Principle Component Loadings for the Batson & Schoenrade (BQuest) Items Batson & Schoenrade Quest Items BF1 BF2 BF3 Change Questions Doubts Religious views still QB12 0.72 changing NO belief change expected QB11 0.72 Expect religious growth QB9 0.70 and change Constantly question my QB10 0.60 beliefs Questions more central QB8 0.36 0.34 than answers Questions increased QB1 0.84 religious interest Questions increased QB4 0.83 importance of God Awareness of tension QB2 0.58 >asked questions Life experiences > QB3 0.34 0.46 rethink belief Doubting important in QB6 0.81 religion I value religious doubts QB5 0.72 Religious doubts upsetting QB7 0.67 [R] [R] Item is reversed scored.
Results and Discussion
In the present study, we wanted to identify the sub-dimensions of quest using three different scales claiming to measure the quest construct. We used a sequence of Principal Components Analyses (PCA) to identify homogeneous clusters of quest items within each measure. We then examined the convergence of these items across the three measures using PCA on a combined pool of items. All of the items for each of the three quest scales included in the analyses are shown in Table 1. Separate Principal Components Analyses (PCA) were conducted for each of the three quest scales. The Scree plots of the eigenvalues are shown in Figure 1 For the 12-item BQuest scale and the 16-item for AQuest, three components were extracted and subject to oblique rotation. The pattern loadings for the obliquely rotated components are shown in Tables 2 & 3. The plot for the RMS items indicated only one component was warranted. The related component loadings are shown in Table 4. A brief paraphrase of each item is given in the tables to identify the content of the items.
The three BQuest components correspond to some degree to the three facets Batson & Schoenrade (1991b) designed into the scale. We labeled the first component, "Change expected" (BF1) since the items with the three highest loadings explicitly mention growth and change. The item, "I am constantly questioning my religious beliefs" also loaded on this factor. Even though the word "questioning" is in this item, the instability of one's beliefs implied by this question may explain why it loads on the change component. The second component consisted of two items that emphasized the role of questions in increasing religious interest and one on how life experience motivated rethinking (questioning) beliefs. We labeled this component, "Questions good." (BF2). The third component consisted of three items that indicate the importance of doubt. We labeled this component, "Doubt valued" (BF3). The correlation between pairs of the three components were modest: BF1-BF2 = 0.24; BF1-BF3 = 0.26; BF2-BF3 = 0.13, Items 3 and 8, with their low, cross-loadings were not used in the dimensional analysis of the combined item.
Next, we conducted a Confirmatory Factor Analysis (CFA) for the 3-component model of BQuest items. Four indices that are used to evaluate goodness of fit for CFA models are shown in Table 6. While the Chi-Square statistic is reported, methodologists acknowledge that it is almost always significant with large sample sizes. (Camp-bell-Stiles & Brown, 2005; Methuen &C Methuen, 2004; Steiger, 1990). As expected, all the chi-square statistics reported in Table 6 are significant at p ----- .0001. All three of the goodness of fit statistics in Table 6 for the BQuest items indicate the three factor model is a good fit. Since the BQuest items have been used to obtain a single score for the quest concept, we calculated the goodness of fit statistics for the 1-factor model. All the indexes in Table 6 indicate the one factor model is a poor fit. The chi-square difference statistic is useful to compare two CFA models on the same set of items. The difference in chi-squares between the 1-factor and 3-factor models is 929 with 21 degrees of freedom, p ----- .0001, indicating that the 3-factor model is a significantly better fit. The 1-factor model is not a good fit for the BQuest items.
Table 6. Goodness of Fit Statistics for Confirmatory Factor Analyses Measure Model Chi-Square DF CFI RMSEA SRMR Goodness > < < of Fit .90 .08 .05 Criteria Bquest 1 1092 54 0.62 0.14 0.09 Items Factor 3 163 33 0.95 0.06 0.03 Factor Aquest 3 307 63 0.89 0.06 0.03 Items Factor RMS 1 214 44 0.86 0.06 0.05 Items Factor Combined 4 1141 227 0.85 0.06 0.04 Items Factor 5 813 205 0.90 0.05 0.03 Factor Note. All Chi-Square values are significant at p ----- .0001
The three components of the AQuest items are shown in Table 3. Component B includes items that we summarize with the label, "Reason vs faith." (AF1) All the items loading on this component reflect the triumph of faith over challenges or doubts. The items are reversed scored to measure the Quest construct, so a high score on this component would reflect the importance of reason over faith. The items on AF1 were uniquely developed by Altemeyer & Hunsberger to measure antiquesting, reflecting authoritarian beliefs about faith. The second component consists of four items valuing doubt and two expecting change. Note that the item labeled QAO is the same as item QB5 from the Batson scale. We labeled this component, "Doubts valued." (AF2) We labeled the third component for AQuest "Religion is a Quest." (AF3) The items all emphasize religion as a quest with a skeptical stance toward truth. Items 7, 11, 12, and 15 with their low, cross-loadings were not used in the combined dimensional analyses. The statistics in Table 6 indicate that the 3-factor model is a reasonably good fit for the AQuest items with the CFI on the borderline (CFI=0.89).
Table 3 Principle Component Loadings for Altmeyer [??] Hunsberger (AQuest) Items Altmeyer & Hunsberger AF1 AF2 AF3 Quest Items Reason Doubts Religion v Valued a Quest Faith Beliefs challenged, believe more QA4 0.67 fervently [R] Beliefs challenged, believe stronger QAM 0.66 [R] Religion not facts, pray for QA2 0.65 enlightenment [R] Beliefs based on faith, not proof, a QA9 0.62 virtue [R] No search for truth, believe QA10 0.61 revelation [R] Don't jeopardize beliefs with QA7 0.47 0.39 skepticism [R] Mind limited, accept revealed truth QA12 0.39 [R] Doubts allow learning QA3 0.73 Better/ beliefs firm, NO doubt [R] QA1 0.67 I value religious doubts QAO 0.62 Beliefs may change as mature, learn QA5 0.56 Goal is truth/ would change beliefs QA15 0.33 0.35 Religion a quest, take nothing on QA6 0.70 faith Open mind, no eternal truth QA8 0.70 Religion a search, NOT earlier QA13 0.65 teachings Honest quest -- > No God, QA11 0.32 0.47 believe it [R] Item is reversed scored.
The loadings for the single component derived from the RMS items are shown in Table 4. These items all express the idea of satisfaction with beliefs along with openness, tolerance of uncertainty, and willingness to change. Seven of the RMS items with loadings above 0.40 were retained for the combined analyses. The statistics in Table 6 indicate that the 1-factor model is a less-than-optimum fit for the 11 items with the CFI at 0.86. Dropping the 4 items that load below 0.40 would improve the goodness of fit.
The 32 quest items that were retained for the combined analysis were subject to exploratory PCA with oblique rotation. The Scree test was again used to estimate the number of components to extract and rotate. The Scree test for the combined group of selected items is shown in Figure 2. The plot appears to support the extraction of 4-components. The four components were extracted and rotated. The 3 items that loaded on the AQuest component Religion a Quest (AF3) had low loadings in the 4-component solution, so we also examined a 5-component solution. The 5-component solution had a component containing these 3 items not accounted for in the 4-component solution, so we retained the 5-components. The goodness of fit statistics for both the 4-component and 5-component solutions are shown in table 6. The chi-square difference between the two models is 227 which is significant at the p ----- .0001 and the CFI for the 5-component model is above 0.90. The 5-component model is a better fit for the data than the 4-component model.
The pattern loadings for the 5 oblique components are shown in Table 5. The first component (QF1) has six items from RMS, four from BQuest, and one from AQuest. The content conveys a dynamic, developmental view of faith as satisfying but changing over time as a result of growing/maturing. The second component (QF2) corresponds to the Reason versus Faith dimension (AF2) of the AQuest measure. The third component (QF3) has three items from the BQuest scale that have to do with the value of questions for faith motivation. Specifically, two of the items with the highest loadings express the value of questions for making one's religious quest more interesting and important. These are the same items that loaded on BF2. Component QF3 also contained item RMS6, "struggle to understand evil and suffering." It seems reasonable to assume that such a struggle would reflect motivation, tension, and questioning. Component 4 (QF4) is a convergence of items from the BQuest and AQuest measure that affirm the value of doubt (BF3 and AF2). The last component (QF5) has the three items from the AQuest component (AF3) that explicitly identifies religion as an open-minded quest with a skeptical attitude regarding eternal truths.
Table 5 Principle Component Analysis of Combined Quest Items COMBINED QUEST ITEMS Pattern Coefficients QF1 Beliefs Tentative Open to Change Beliefs may change as QA5 0.63 mature, learn Expect religious growth QB9 0.55 and change Constantly question my QB10 0.34 beliefs NO belief change expected QB11 0.42 [R] Religious views still QB12 0.47 changing Beliefs satisfying/open to RMS1 0.66 new information Happy with beliefs/ open RMS2 0.63 to new insight Beliefs true, but could be RMS3 0.63 mistaken No easy answers, faith RMS4 0.44 developmental process beliefs not certain, act RMS8 0.47 on probably true Questions/ hesitant to be RMS9 0.44 dogmatic QF2 Reason versus Faith Religion not facts, pray QA2 0.57 for enlightenment [R] Beliefs challenged, QA4 0.75 believe more fervently [R] Beliefs based on faith, QA9 0.60 not proof, a virtue [R] No search for truth, QA10 0.55 believe revelation [R] Beliefs challenged, QA14 0.74 believe stronger [R] QF3 Questions Good Questions increased QB1 0.76 religious interest Awareness of tension QB2 0.56 > asked questions Questions increased QB4 0.75 importance of God struggled to understand RMS6 0.50 evil & suffering QF4 Doubts Valued Better/ beliefs firm, NO QA1 0.67 doubt [R] Doubts allow learning QA3 0.53 Constantly question my QB10 0.35 beliefs I value religious doubts QB5 0.63 Doubting important in QB6 0.74 religion Religious doubts upsetting QB7 0.64 [R] QF5 Religion a Quest Religion a quest, take QA6 0.70 nothing on faith Open mind, no eternal QA8 0.64 truth Religion a search, NOT QA13 0.54 earlier teachings Note. [R] Items reversed scored. QB10 is the only item loading on 2 components > .30/
We calculated scale scores for the components within each quest measure and five components from the combined items. In each case, the component score (QF) was the average of the items with loadings above 0.40. Table 7 shows the correlations of the components derived from the combined set of items with the component scores from the separate quest scales. The QF1 component shows convergence on items from the BQuest, AQuest, and RMS items that measure Beliefs Tentative Open to Change. The QF4 component shows a convergence on items which come from the BQuest and AQuest measures that express the Value of Doubts. The other three QF components from the combined set of items (QF2, QF3, & QF5) correspond to components AF1, BF2, and AF3, respectively. Table 8 lists the names of the QF components and shows the correspondence with the components from the separate scales. The correlations among the individual components are shown in Table 9. The correlations among the components from the combined items are shown in Table 10. The coefficient alphas for each of the component scales are shown on the diagonals of the correlation matrices in the two tables. The alphas for the QF scales ranged from 0.57 for the 3-item Religion as a Quest (QF5) to 0.78 for the Beliefs Tentative Open to Change component (QF1). Table 11 shows the correlation of all the quest components with the original quest scale.
TABLE 7. Correlation of Separate Quest Scale Combined whth Combined Item Components Separate Quest Scale Components Combined BF1 RMSF1 AF2 AF1 BF2 BF3 AF3 Quest Items Components QF1 0.78 0.88 0.63 0.10 0.18 0.31 0.30 QF2 0.15 0.1 0.13 1 0.02 0.19 0.15 QF3 0.22 0.35 0.20 0.04 0.94 0.18 0.23 QF4 0.43 0.33 0.85 0.19 0.11 0.92 0.14 QF5 0.24 0.29 0.18 0.15 0.24 0.15 1 Table 8. Correspondence of Combined Item Components to Separate Quest Scale Components Sub-Dimension Combined Separate Items Scales Component Components Beliefs QF1 BF1 AF2 RMSF1 Tentative Open to Change Reason versus QF2 AF1 Faith Questions QF3 BF2 Good Doubts Valued QF4 BF3 AF2 Religion a QF5 AF3 Quest Table 9. Correlation of Separate Quest Components BFl RMSF1 AF2 AF1 BF2 BF3 AF3 BFl (.67) RMSF1 .48 (.69) AF2 .61 .48 (.60) AF1 .15 .10 .13 (.69) BF2 .18 .21 .13 .02 (.66) BF3 .34 .28 .69 .19 .14 (.65) AF3 .24 .29 .18 .15 .24 .15 (.57) Note: Coefficient Alpha for each component is on the diagonal Table 10. Correlation of Combined Item Components QF1 QF2 QF3 QF4 QF5 QF1 (.78) QF2 .10 (.69) QF3 .24 .04 (.63) QF4 .39 .19 .17 (.72) QF5 .30 .15 .23 .14 (.57) Note: Coefficient Alpha for each component is on the diagonal Table 11. Correlation of Quest Components with Original Quest Measures and Religious Maturity Scale QUEST BATSON ALTEMEYER RELIGIOUS MATURITY COMPONENTS QUEST QUEST SCALE QF1 0.66 0.48 0.85 BF1 0.78 0.46 0.52 RMSF1 0.50 0.42 0.90 AF2 0.69 0.61 0.54 QF2 = AF1 0.18 0.70 0.14 QF3 0.61 0.18 0.31 BF2 0.59 0.13 0.20 QF4 0.66 0.55 0.38 BF3 0.65 0.46 0.32 QF5 = AF3 0.33 0.40 0.27
Taken together, the data in Tables 7 through 11 shows the following patterns. The subcomponents of each quest scale are distinct enough to be represented as separate subscales. The only one of the three quest scales that can be considered a 1-component measure psychometrically is the RMS. The correlations among the QF2 through QF5 components is quite low with only one over 0.20 (0.23 between the Questions Good and Religion a Quest). The QF1 component, Beliefs Tentative and Open to Change, correlated 0.31 with Doubts Valued (QF4/BF3) and 0.30 with Religion a Quest (QF5/AF3).
The quest items that were omitted from the dimensional analysis of the combined items were correlated with the QF component scores. The results are shown in Table 12. The omitted items tended to have one of two characteristics. Four of the items (RMS items 5, 7, 10 &c QA12) had very low correlations with all or most of the five QFs. These items do not appear to measure any of the quest sub-dimensions and could be omitted from future assessment of quest. The remaining six items in Table 9 have significant correlations with four or five of the QF component dimension. For example, item QB10, "I constantly question my beliefs," QA11 "If honest quest concludes 'no God/ believe it," and RMS 11 "Beliefs same as 5 years ago," have significant correlations with all 5 QFs. Items QB3, "Life experiences led to rethink belief," QB8, "Questions more central than answers," and QA15, "Goal is truth/ would change beliefs," correlated with 4 of the 5 QFs. These items do not show simple structure so they could be omitted from an investigation using sub-dimensions but could be useful in a global assessment of quest (single score).
Table 12. Correlation of Omitted Quest Items with Combined Item Components QUEST COMPONENTS Omitted Quest QF1 QF2 QF3 QF4 QF5 Items Constantly QB10 0.50 0.21 0.20 0.39 0.28 question my beliefs Honest quest QA11 0.19 0.32 0.17 0.13 0.26 > No God, believe it Beliefs same RMS 0.38 0.12 0.23 0.25 0.11 as 5 years ago 11 [R] Life QB3 0.42 0.38 0.28 0.23 experiences > rethink belief Questions more QB8 0.34 0.14 0.34 0.27 central than answers Goal is truth/ QA15 0.39 0.16 0.17 0.21 would change beliefs Need certainty RMS5 0.18 0.17 0.14 religion completely true [R] Preach, NO RMS7 0.14 -0.11 involvement in politics [R] People RMS 0.18 0.10 0.14 relationships 10 as import as God Mind limited, QAU 0.28 accept revealed truth [R] Note. [R] Items reversed scored. Correlation shown are significant at p----- .01, two-tailed test.
Complex items used to calculate total quest scores may be effective for prediction studies. Global quest scores may predict prejudice or other conceptually related behavior or attitudes. Factorially homogeneous subscales allow for investigating which sub-dimensions of quest account for the observed prediction. It is possible that two sub-dimensions could correlate in opposite directions with a dependent variable and the total quest score would correlate zero with the DV. This is what Beck et al. (2001) found when using their two subscales of quest (Tentativeness and Change) and Batson's scale total score to predict religious well-being. The correlations for Tentativeness were positive and for Change were negative with religious well-being but the total quest score had zero correlation with well-being.
We correlated the five Quest components and the three original quest scales with several measures of religious functioning. The results of these analyses are shown in Table 13. Three of the religious variables in Table 13 have two distinct subscales. In each case, one subscale has a vertical orientation emphasizing spiritual well-being and God-awareness (e.g., I have a personally meaningful relationship with God; FM_V, SW-B, & SEI-SS) and the other subscale has a horizontal orientation emphasizing social openness or personal well-being (e.g., I am active in efforts to promote world peace; FM_H, EW-B, 8c SEI-SO). The horizontal scales stress social responsibility, social inclusion, and personal well-being. The SEI-SO scale has several items that are very similar to the quest items of openness and tentativeness of personal beliefs. The table also includes correlations of both the quest variables and the religious functioning variables with Religious Fundamentalism (RF).
Table 13. Correlation of Religious Variables with Quest Components and Original Quest Measures Quest Components Religious Measures QF1 QF2 QF3 QF4 Religious Fundamentalism (RF) -0.34 -0.41 -0.13 -0.26 SEI-Spiritual Openness 0.50 0.21 0.18 0.31 Faith Maturity-Horizontal 0.15 -0.17 0.18 0.10 Existential Well-Being 0.18 -0.15 0.18 SAI-Disappointment 0.18 0.21 0.15 SAI-Grandiosity -0.10 0.14 -0.10 SAI-Instability 0.18 0.2 Faith Maturity-Vertical -0.43 Spiritual Well-Being -0.34 SEI-Spiritual Support -0.47 Spiritual Maturity Index -0.47 SAI-Awareness -0.46 SAI-Realistic Acceptance -0.37 Quest Scales Religious Measures QF5 BQUEST AQUEST RMS Religious Fundamentalism (RF) -0.38 -0.36 -0.46 -0.37 SEI-Spiritual Openness 0.41 0.46 0.39 0.52 Faith Maturity-Horizontal 0.14 0.21 Existential Well-Being 0.11 0.18 0.16 SAI-Disappointment 0.24 0.14 0.18 SAI-Grandiosity 0.18 SAI-Instability 0.13 0.20 0.15 Faith Maturity-Vertical -0.24 -0.11 -0.34 Spiritual Well-Being -0.21 SEI-Spiritual Support -0.24 -0.31 Spiritual Maturity Index -0.31 -0.16 -0.38 SAI-Awareness -0.16 -0.13 -0.36 SAI-Realistic Acceptance -0.17 -0.13 -0.29 Religious Measures RF Religious Fundamentalism (RF) SEI-Spiritual Openness -0.61 Faith Maturity-Horizontal Existential Well-Being SAI-Disappointment SAI-Grandiosity SAI-Instability Faith Maturity-Vertical 0.47 Spiritual Well-Being 0.29 SEI-Spiritual Support 0.47 Spiritual Maturity Index 0.50 SAI-Awareness 0.37 SAI-Realistic Acceptance 0.37 Note. Only correlations significant at p -----.01, two-tailed are displayed
The patterns of relationships show some differences in correlations of the quest components. Genia's SEI-Spiritual Openness had significant correlations with all of the quest components and the original quest scales. The highest correlation of SEI-SO is with QF1, the tentative/open component (0.50). These results support our observation that SEI-SO is the most quest-saturated of the religious functioning variables in Table 13. The other two horizontal religious variables had low positive or non-significant correlations with the quest variables, with the exception of QF2. Faith Maturity - Horizontal and Existential Well-Being had low significant negative correlations with QF2. All of the spiritual well-being and God-awareness measures had significant negative correlations with QF2 (Reason versus Faith) and QF5 (Religion as Quest) (-0.21 to -0.47). It appears that two of Altemeyer & Hunsberger's quest sub-dimensions (QF2 &C QF5) measure a questioning attitude toward religious authority that is at odds with personal spiritual well-being.
The spiritual well-being and God-awareness variables had no significant correlations with the other three quest dimensions (QF1, QF3, & QF4). The lack of correlation of vertical religious well-being with quest is consistent with Bateson's original claim that quest is an independent, third dimension of religious motivation. We did not include a specific measure of Intrinsic or Extrinsic motivation so our results offer tentative support for the Batson's theory. The spiritual well-being and God-awareness scales we did use contain a number of items expressing intrinsic religious motivation.
Five of the six correlations of the SAI Disappointment and Instability subscales with QF1, QF3, & QF4 were low and significantly positive. These two SAI scales measure negative aspects of vertical religious experience. These results tentatively suggest that the uncertainty of belief reflected in these quest dimensions has a modest negative impact on vertical religious experience. The SAI Grandiosity scale had low significant negative correlations with QF2 and QF4 and low significant positive correlations with QF3 and QF5. The Grandiosity sub-scale was designed to assess a narcissistic approach to personal religion (Hall & Edwards, 1996). These results suggest that a narcissistic tendency in personal religion corresponds to valuing faith over reason and devaluing doubt while endorsing one's questions as contributing to faith and endorsing faith as a quest. This is offered as a very tentative interpretation of the results since the correlations are very modest.
The pattern of correlations of Religious Fundamentalism (RF) with quest, spiritual well-being, and God-awareness measures shown in Table 13 is interesting. RF has significant negative correlations with all of the quest measures and SEI-SO. This finding is consistent with the formulation of the quest concept as an open-minded approach to religion. On the other hand, RF has significant positive correlations with all of the spiritual well-being and God-awareness scale. The correlations of RF scale with the vertical religious scales are parallel to the correlations of the quest components QF2 and QF5. Altemeyer and Hunsberger (1992) explicitly formulated their measure of quest as an extension of their work on religious fundamentalism. They found that RF was not correlated with orthodox beliefs. Our results show that RF is positively correlated with vertical religious variables in a sample of Evangelical Christians. The questions that load on the QF2 were formulated as anti-quest items expressing belief over reason which also taps a fundamentalist mentality. It is reasonable to assume that individuals who hold orthodox beliefs, when faced with a conflict between their faith and seemingly contradictory information, would place their faith in religious authority. So it is also possible QF2 confounds orthodoxy and rigid fundamentalism. The pattern of correlations for the three other quest components (QF1, QF3, and QF4) suggests an alternative to direct measures of religious fundamentalism among Evangelical Christians. The three components correlated negatively with the RF scale validating the quest construct as a measure of openness versus rigidity in religious motivation. The fact that these same three components did not correlate with the spiritual well-being and God-awareness scales suggests that they are measuring openness, tentativeness, and doubt in the religious domain in a way that is not biased against vertical religious experience.
We had information on the student respondent's year in school. Since quest is considered to be a developmental process that is either enhanced or impeded by education, we expected differences in the level of quest across classes with quest scores being higher in the upper-class years. We compared the mean scores on the quest components as a function of year in school using a mixed-model Year-by-Quest Scale repeated measures ANOVA. We were able to treat the quest scales as a repeated measures factor because we were using the average score across items within each scale. So even though the subscale scores had different numbers of items, they were all measured on a scale of 1 to 5. There was a significant Year effect (F = 15.85; dfl= 3, dfl = 988; p ----- .001), Quest effect (F - 444; dfl = 4, dfl - 3952; p -----.001) and a Year by Quest interaction (F = 3.29; dfl = 12, dfl - 3952; p ----- .001). The interaction effect indicated that the pattern of means varied across year as a function of the quest sub-dimension. The means for each quest sub-dimension by year are shown in Table 14 and Figure 4. The differences among the quest components within each year are the largest effects. The mid-point of the scale is 3.5 with scores above this point indicating average agreement with the quest items. Two of the five components had averages about 3.5 across all four classes. Change Expected (QF1) means were around 4.0 with juniors significantly higher than the other three classes and seniors higher than freshman and sophomores. Doubt Valued (QF4) means were around 3.75 for freshman and sophomores, which were significantly less than juniors and seniors whose means were around 4.03. Means for Questions Good (QF3) were third highest with the means falling just below the mid-point of the rating scale (around 3.3). There were no differences between classes on QF3. Means for Religion a Quest (QF5) were next also averaging around 3.3. Juniors were significantly higher than freshman and sophomores. Seniors were not significantly different from the other three classes. The means for Reason over Faith (QF2) were the lowest across all classes with means averaging around 2.7. Seniors were significandy higher than the other three classes, which were not significantly different from each other. With the exception of QF2, Juniors had the highest scores within the other four scales and were significandy higher than FR or SO on 3 of the 5 scales. Seniors tended to be equal to or lower than Juniors. The general pattern was for the Quest scores to increase between the Sophomore and Junior year and then decrease from the Junior to Senior years. These are only cross-sectional data so the actual pattern of change across years needs to be studied in a longitudinal design. But these cross-sectional differences do support the concept of quest developing as a result of education. The fact that QF2, Reason versus Faith, is so low is disconcerting. Altemeyer & Hunsberger specifically formulated the balanced quest scale to measure an anti-questing attitude that would predict prejudice. As was noted above, in this sample of Evangelical Christians in higher education, a common working assumption in liberal arts education is that "all truth is God's truth." But when there is a conflict between secular authority and religious authority, such students are typically taught to give precedence to the latter.
Table 14. Means on Quest Components by Year in School Year Quest Component FR SO JR QF1 Open to Change 3.90 (a) 3.89.(a) 4.33 (b) QF2 Reason vs Faith 2.54 (a) 2.68 (a) 2.66 (a) QF3 Questions Good 3.26 (a) 3.26 (a) 3.41 (a) QF4 Doubts Valued 3.72 (a) 3.77 (a) 4.05 (b) QF5 Religion a Quest 2.97 (a) 2.91 (a) 3.23 (b) Year Quest Component SR F-Ratio QF1 Open to Change 4.12 (C) 14.89 * QF2 Reason vs Faith 2.92 (b) 2.45 (*) QF3 Questions Good 3.28 (a) 1.02 QF4 Doubts Valued 4.01.(b) 10.23 * QF5 Religion a Quest 3.11 (ab) 4.56+ Note. F-Ratio for Year Effect levels of significance are: * p ----- .001; + p ----- .01 Means with different superscripts are significantly different at p ----- .01
While sub-dimensions of the quest concept have been acknowledged from its inception, almost all of the research on quest to date has used a single total score. The results of our component analyses within each quest measure and across the measures indicate that these items are measuring distinct constructs. The components for the BQuest measure show a high degree of convergent validity with factors identified by Batson and Shoenrade (1991b). Dudley and Cruise (1992) formulated their Religious Maturity Scale as an alternative to Batson's original model. Our results show that the RMS converges with BQuest openness to change dimension. The BQuest and AQuest measures converge on the Doubts Valued dimension. Both of these components had Coefficient Alpha values above 0.70. They are the two components with the most number of items, which affects the value of alpha.
The third sub-dimension of quest originally identified by Batson emphasized the value of questions in promoting religious interest. This dimension is not as well measured by the current pool of quest items. The two items related to questions consistently loaded on the same distinct component in Batson and Schoenrade's (1991b) analysis and ours. The two items (QB1 & QB2) state that questions increased religious interest and increased importance of God, respectively. We would suggest that these questions do not differentiate levels of quest motivation. This suggestion is supported by the fact that QF3 was the only dimension with no significant difference among the four classes of students.
Two items related to questioning from BQuest, QB8 and QB10, did not consistently fit in the multidimensional model. In the dimensional analysis of the BQuest items, QB10 (I am constantly questioning my religious beliefs) loaded on the openness to change component. QB10 did not fit on this component in the combined item analyses, which included the more moderately worded RMS openness to change items. We would suggest that dropping the adjective "constantly" might improve the item's fit in the multidimensional model. Item QB8 "questions more central than answers" did not have simple structure. It has a more skeptical tone of valuing questions as an end rather than a means to faith.
In our preliminary study of religious measures (Edwards et al. 2001), we speculated that the dimensions of quest could be classified as assessing a "soft" and a "hard" version of the construct. The soft version corresponds to components QF1 and QF4, which allow for satisfying belief while remaining open to change and embracing doubts as an inevitable component of religious belief. The "faith with openness and doubt" perspective was the emphasis offered by Dudley and Cruise (1992) in their formulation of the RMS. We would suggest that components QF1 (Beliefs Tentative Open to Change) and QF4 (Doubts Valued) provide an excellent model for assessing the "soft" version of quest. The fact that these two scales correlate negatively with Religious Fundamentalism validates them as a measure of an open cognitive style specified in the quest construct. The fact that these two scales did not correlate with measures of vertical faith indicates they are not biased against individuals with satisfying faith commitments. The "hard" version of quest would correspond to components QF2 and QFS where faith and truth seeking are presented in a more polarized way. These two components had significant negative correlations with various measures of spiritual well-being and God-awareness.
One of the main purposes of the present study was to determine if the quest construct as measured by three existing instruments is multidimensional. The answer appears to be yes, both within the BQuest and AQuest measures as well as in the combined set of items. Identifying the specific sub-dimensions of quest remains a challenge. The only component showing convergence across all three measures was QF1, Beliefs Tentative Open to Change. The component QF4, Doubts Valued, showed convergence of items from BQuest and AQuest. We suavest that these two components may tentatively be viewed as core latent dimensions of the quest construct as formulated by Batson. It is noteworthy that these two dimensions represent what we have called soft quest. It appears that a conception of quest that is inversely related to a rigid (fundamentalist) cognitive style, and independent of--and thus not biased against--faith commitments and spiritual well-being, captures more of the essence of the original theory of quest as a dimension of mature religiousness. Our results suggest that hard quest is an indicator of immature religiousness.
Two of the approaches to measuring the quest construct investigated in this study are different in important ways. Some of the BQuest and AQuest items are taping different sub-dimensions and are not equivalent. The AQuest factor, QF2 Reason versus Faith, appears to primarily be a measure of Religious Fundamentalism that potentially confounds orthodoxy with rigid thinking. The AQuest factor, QF5 Religion as Quest, is a more direct measure of religion as a process of questing. The BQuest factor, QF3, on the importance of questions, needs further development with an expanded set of items. Beck and colleagues have taken yet another direction andhave developed their own multidimensional measure of quest (Beck et al.., 2OO1; Beck & Jessup, 2OO4). Fur ther research is needed on the convergence of these various measures of quest to identify a set of core latent constructs that can be used to investigate the validity and utility of each construct.
Is it important to measure the dimensions of quest separately? For the dimensions of change
(QF1) and doubt (QF4), we did not find differential correlations with other religious variables. Beck and his colleagues did find differential correlations (Beck et al., 2001; Beck & Jessup, Z004). We will need additional research on the convergence of all of these measures with other samples to provide a comprehensive conceptual map of the quest domain.
The primary focus of the present study has been on psychometric issues regarding the various measures of quest available today. The results are derived from a sample of college students at Evangelical Christian colleges. The fact that they represent a relatively homogeneous sample has advantages and disadvantages. The advantage is that the variation on the quest variables would be primarily a function of personal factors relevant to quest as a psychological construct. The disadvantage is that there may be some restriction in range of the variables measured which could attenuate the size of correlations observed. A major limitation is that the homogeneous sample limits the external validity of our results. Another limitation of the present study is that we did not include specific measures of Intrinsic and Extrinsic faith. Future research on the dimensions of quest should include these measures as well.
As noted earlier, the original measure of quest, the Interactional Scale, was subjected to a great deal of criticism regarding its validity and reliability in the 198O's. Batson and Shoenrade (1991a, 1991b) responded to their critics with an extensive defense of the quest construct. They also offered a revision of the IS scale to its current, more psychometrically sound 12-item version. They also made the appeal that researchers should move beyond psychometric preoccupation to doing substantive research on the functioning of quest in individual lives. We agree that psychometric research should be a means to achieving better measures of our constructs, which then facilitates better research on substantive questions. A multidimensional model of quest may provide a framework that integrates the quest literature and resolves some of the questions and doubts raised by quest critics. However, in the spirit of quest, such answers will need to be pursued with openness and tentativeness as we seek to map the domain of personal religious faith motivation.
Please address correspondence to Keith J. Edwards, Ph.D., Rose-mead School of Psychology, Biola University, 13800 Biola Ave., LaMirada, CA 90639. Email: email@example.com.
Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5, 432-443.
Altemeyer, B., & Hunsberger, B. (1992). Authoritarianism, religious fundamentalism, quest, and prejudice. The International Journal for the Psychology of Religion, 2, 113-133.
Asparouhov, T., & Muthen, B. (2009). Exploratory Structural Equation Modeling. Structural Equation Modeling, 16, 397-438,
Beck, R., Baker, L., Robbins, M, & Dow, S. (2001). A second look at quest motivation: Is Quest unidimensional or multidimensional? Journal of Psychology and Theology, 29, 148-157.
Beck, R., & Jessup, R. K. (2004). The Multidimensional Nature of Quest Motivation. Journal of Psychology and Theology, 32, 283-294.
Bassett, R. L., Camplin, W., Humphrey, D., Dorr, D., Biggs, S., Distaffen, R., Doxator, I., Flaherty, M., Hunsberger, P. J., Poage, R., & Thompson, H. (1991). Measuring Christian maturity: A comparison of several scales. Journal of Psychology and Tbeology, 19, 84-93.
Batson, C. D. (1976). Religion as prosocial: Agent or double agent? Journal for the scientific study of religion, 15, 29-45.
Batson, C. D.. & Schoenrade, P. A. (1991). Measuring religion as quest: Validity concerns. Journal for the Scientific Study of Religion, 30, 416-429.
Batson, C. D., & Schoenrade, P. A. (1991), Measuring religion as quest: Reliability concerns. Journal for the Scientific Study of Religion, 30, 430-447.
Batson, C. D., Schoenrade, P. A., & Ventis, W. L. (1993). Religion and the individual: A social-psychological perspective. New York: Oxford University Press.
Batson, C. D., & Ventis, W. L. (1982). The religious experience: A social-psychological perspective. New York: Oxford University Press.
Benner, D. G. (1988). Psychotherapy and the spiritual quest. Grand Rapids, MI: Baker Book House.
Benson, P. L., Donahue, M. J., & Erickson, J. A. (1993). The Faith
Maturity Scale: Conceptualization, measurement, and empirical validation. Research in the Social Scientific Study of Religion, 5, 1-26.
Buhrow, W., Calkins, P., Haws, J. & Rost, K. (1986). The Spiritual Maturity Index: A study of reliability and validity. Unpublished manuscript, Ailiance Theological Seminary.
Bufford, R. K., Paloutzian R. F., & Ellison C. W. (1991). Norms for the spiritual well-being scale. Journal of Psychology and Theology, 19,56-70.
Bufford, R. K. (1984, April). Empirical correlates of Spiritual Well-Being and Spiritual Maturity scales. Paper presented at the annual meeting of the Christian Association for Psychological Studies, Dallas TX.
Campbell-Sills, L., & Brown, T. A. (2005). Latent variable approaches to studying the psychopathology and classification of emotional disorders. In F. Andrasik (Ed.), Comprehensive handbook of personality and psychopathology (Vol.11: adult psychopathology, pp. 21-35). New York: John Wiley & and Sons.
Cannon, M. M. (1994). The psychological development of and spiritual development of women: An empirical study. Unpublished doctoral dissertation, Loyola College, Maryland.
Cooper, R. L. (1987). An empirical examination of the construct validity of the Spiritual Maturity Index. Dissertation Abstracts International, 47, 4645B. (University Microfilms International, 87-04712)
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology; 24, 349-354.
Davis, W. W., Longfellow, D. R., Moody, A. V., & Moynihan, W. S. (1987). Spiritual Maturity Index: Construct validation.
Unpublished manuscript, presented at Western Conservative Baptist Seminary, Portland, OR.
Dudley, R. L., & Cruise, R. J (1990). Measuring religious maturity: A proposed scale. Review of Religious Research, 32, 97-109.
Edwards, K. J., Slater, W., Hall, T. W., Oda, A., & Eck, B. (August, 2001). Spiritual change among college students: A National Study. Paper presented at the 109th Annual Convention of the American Psychological Association, San Francisco.
Ellison, C. W. (1983). Spiritual well-being: Conceptualization and measurement. Journal of Psychology and Theology, 11, 330-340.
Ellison, C. W.,& Smith, J. (1991). Toward an integrative measure of health and well-being. Journal of Psychology and Theology, 19,35-48.
Genia, V. (1991). The spiritual experience index: A measure of spiritual maturity. Journal of Religion and Health, 30, 337-347. Gorsuch, R. L. (1984). Measurement: The boon and bane of investigating religion. American Psychologist, 39, 228-236.
Gorsuch, R. L. (1988). Psychology of religion. Annual Review of Psychology, 39, 201-221.
Gorsuch, R. L., & McPherson, S. E. (1989). Intrinsic/extrinsic measurement: B/E revised and single-item scales. Journal for the Scientific Study of Religion, 28, 348-354.
Hall, T. W., & Edwards, K. J. (1996). The initial development and factor analysis of the Spiritual Assessment Inventory. Journal of Psychology and Theology, 24, 233-246.
Hood, R. W., Jr., & Morris, T. J. (1985). Conceptualization of Quest: A critical rejoinder Batson. Review of Religious Research, 26, 391-397.
Kauffman, J. H. (1979). "Social correlates of spiritual maturity among North American Mennonites." In David 0. Moberg (Ed.). Spiritual well-being: Sociological perspectives. Washington, DC: University Press of America, Pp. 237-254.
Kirkpatrick, L. A. (1988). A psychometric analysis of the AllportRoss and Feagin measures of intrinsic-extrinsic religious orientation. In D. O. Moberg & M. L. Lynn (Eds.), Research in the social scientific study of religion (Vol. 1). Greenwich, CT: JAI Press.
Kirkpatrick, L. A., & Hood, it W. (1990). Intrinsic-extrinsic religious orientation: The boon or bane of contemporary psychology of religion? Journal for the Scientific Study of Religion, 29, 442-462.
Methuen, L. K., & Methuen, B. 0. (2004). Mplus users guide (3rd ed.). Los Angeles: Methuen & Methuen.
McFarland, S. G. (1989). Religious orientations and the targets of discrimination. Journal for the Scientific Study of Religion, 28, 324-336.
Moberg, D. 0. (1971). Spiritual well-being: Background issues. Washington, DC: White House Conference on Aging.
Moberg, D. 0. (1984). Subjective measures of spiritual well-being. Review of Religious Research, 25, 351-364.
Palouttian, R. F. (1996). Invitation to the psychology of religion (2nd ed.). Needham Heights, MA: Ailyn and Bacon.
Paloutzian, R. F., & Ellison, C. W. Loneliness, spiritual well-being and quality of life. In U A. Peplau and D. Perlman (Eds.), Loneliness: A sourcebook of current theory, research and therapy. New York: Wiley 1nterscience, 1982.
Pargament, K. 1. (1999). The psychology of religion and spirituality? Yes and no. The International Journal for the Psychology of Religion, 9, 3-16.
Pargament, K. 1. (1992). Of means and ends: Religion and the search for significance. The International Journal for the Psychology of Religion, 2, 201-229.
Rummel, It L. (1970). Applied factor analysis. Evanston, IL: Northwestern University Press.
Slater, W. E. (1999, August). Defining and measuring spiritual health: An investigation of the psychometric properties of measures of spiritual well-being and spiritual maturity.
Paper presented at 107th Annual Convention of American Psychological Association, Boston, MA.
Slater, W. E. (1998, March). Measures of spiritual maturity: An investigation of their psychometric properties. Paper presented at the Coalition for Christian Colleges and Universities National Assessment Conference, Cleveland, TN.
Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25, 173 -- 180.
Strahan, R. & Gerbasi, K. C. (1972). Short, homogeneous versions of the Marlowe-C:rowne Social Desirability Scale. Journal of Clinical Psychology, 28, 191-193.
Ventis, W. L (1995). The relationship between religion and mental health. Journal of Social Issues, 51, 33-48.
Watson, P. J., Morris, R. J., & Hood, it W., Jr. (1989). Interactional factor correlations with Means and End religiousness. Journal for the Scientific Study of Religion, 28, 337-347.
Yeasts, J. R., & Asher, W. (1982). Factor analysis of religious variables: Some methodological considerations. Review of Religious Research. 24, 49-54.
KEITH J. Edwards and Todd W. Hall Rosemead School of Psychology, Biola University
Will Slater Bluffton College
Jonathan Hill Calvin College
EDWARDS, KEITH, J. Address: 13800 Biola Ave., Rosemead School of Psychology, Biola University, La Mirada, CA 90639. Degrees: BEd, Mathematics, University of Wisconsin-Whitewater; MS, PhD, Quantitative Methods, New Mexico State University; PhD, Social and Clinical Psychology, University of Southern California. Title: Professor of Psychology. Areas of Specialization: quantitative methods, individual and couple emotion focused therapy, psychology of religion, interpersonal affective neuroscience of attachment, missionary member care.
HALL, TODD. Address: 13800 Biola Ave., Rosemead School of Psychology, Biola University, La Mirada, CA 90639. Titles: Professor of Psychology, Director, Institute for Research on Psychology and Spirituality. Degree: Ph.D. Areas of specialization: Christian spirituality, spiritual development, attachment theory, relational psychoanalysis.
SLATER, WILL. Address: Bluffton University, 1 University Drive Bluffton, OH 45817. Titles: Professor of Psychology, Chair, Social Sciences Department. Degree: Ph.D. Areas of specialization: Christian spirituality, mindfulness, positive psychology.
HILL, JONATHAN, P. Address: Department of Sociology and Social Work, Calvin College, 3201 Burton SE, Grand Rapids, MI 49546. Titles: Assistant Professor of Sociology. Degree: Ph.D. Areas of specialization: Sociology of religion, quantitative methods, higher education.
|Printer friendly Cite/link Email Feedback|
|Author:||Edwards, Keith J.; Hall, Todd W.|
|Publication:||Journal of Psychology and Theology|
|Date:||Jun 22, 2011|
|Previous Article:||Books received.|
|Next Article:||Biblical Foundationalism and Religious Reflection: polarization of faith and intellect oriented epistemologies within a Christian ideological...|