Making religiosity person-centered: A latent profile analysis of religiosity and psychological health outcomes

Personality and Individual Differences 88 (2016) 160–169

Contents lists available at ScienceDirect

Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid

Making religiosity person-centered: A latent profile analysis of religiosity and psychological health outcomes Adrian J. Bravo a,⁎, Matthew R. Pearson b, Leah E. Stevens a a b

Department of Psychology, Old Dominion University, 250 Mills Godwin Building, Norfolk, VA 23529-0267, United States Center on Alcoholism, Substance Abuse, & Addictions, University of New Mexico, 2650 Yale Blvd SE MSC 11-6280, Albuquerque, NM 87106, United States

a r t i c l e

i n f o

Article history: Received 19 June 2015 Received in revised form 17 August 2015 Accepted 31 August 2015 Available online xxxx Keywords: Religiosity Emotional health Psychological well-being Psychological flexibility Latent profile analysis Person-centered analysis

a b s t r a c t Although variable-centered analyses predominate the religiosity-health literature, they are limited in that they tend to focus on the (unique) associations between a single facet of religiosity and outcomes. Person-centered analyses allow the identification of distinct subpopulations defined by individuals' full response profiles on facets of religiosity. The present study used latent profile analysis to identify distinct subgroups defined by their scores on the Religious Life Inventory-Revised. Using the Lo–Mendell–Rubin Likelihood Ratio Test, we found that a four-class solution fits optimally in two samples of Christian college students, including questioning (high quest, low intrinsic/extrinsic), intrinsically motivated (high intrinsic), high religiosity (high on all religious orientations), and low religiosity (low on all religious orientations) groups. Across both studies, we found, that the high religiosity, low religiosity and questioning groups reported significantly lower levels of psychological well-being compared to the “Intrinsically Motivated” group. These results corroborate studies suggesting that intrinsic religiosity is a protective factor associated with good psychological well-being among religious students and that personal religious struggles (i.e., quest religiosity) are associated with poorer psychological well-being. Our results point to the utility of person-centered analyses to examine religiosity in unique ways. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Within the psychology of religion, religious orientations reflect different motivations for being religious. Extending the foundational research of Allport and Ross (1967), Batson and Ventis (1982), Batson and Schoenrade (1991a, 1991b) and Batson, Schoenrade, and Ventis (1993) posit that religious orientation, or religiosity, can be understood in terms of three dimensions: intrinsic religiosity, extrinsic religiosity, and quest religiosity. According to Batson et al. (1993), intrinsicoriented people take religion seriously as an end in itself; thus, these individuals have a strong dedication to their religious values, beliefs, and practice. In contrast, extrinsic-oriented people view religion as a useful means to an end; thus, these individuals may use religion as means to other, self-serving ends, like social gains. Finally, quest-oriented people view religion as an interactive way of finding meaning in life and tend to critically question one's religious beliefs. Based on various theoretical models, such as Self-Determination Theory (Deci & Ryan, 1985; Ryan, Rigby, & King, 1993), Religious Coping Theory (Pargament, 1997) and Meanings Systems Framework (Park, 2005, 2007), research has found ⁎ Corresponding author. E-mail addresses: [email protected] (A.J. Bravo), [email protected] (M.R. Pearson), [email protected] (L.E. Stevens).

http://dx.doi.org/10.1016/j.paid.2015.08.049 0191-8869/© 2015 Elsevier Ltd. All rights reserved.

that each of these dimensions differentially predicts various aspects of personal meaning and psychological well-being (see Moreira-Almeida, Lotufo Neto, & Koenig, 2006 for a review). Specifically, among religious college students, intrinsic religiosity has been identified as a protective factor for depressive symptoms, anxiety symptoms, and alcohol-related outcomes (Berry & York, 2011; Jansen, Motley, & Hovey, 2010; Stewart, 2001; Wood & Hebert, 2005). In contrast, quest religiosity and extrinsic religiosity have been linked to poor mental health (Hill & Pargament, 2008; Maltby & Day, 2000; Steger et al., 2010; Steger, Kashdan, Sullivan, & Lorentz, 2008). Notably, a recent longitudinal study found that religious service and activity attendance tends to decrease during the first few semesters of college (Stoppa & Lefkowitz, 2010), suggesting that the college years are a time of transition. Although variable-centered analyses (e.g., multiple regression, structural equation modeling) predominate the psychology of religion literature, they are limited in that they tend to focus on the (unique) associations between a single facet of religiosity and outcomes. This approach may be a serious limitation considering that these religious orientations (i.e., intrinsic, extrinsic, and quest) have never been claimed to be mutually exclusive of each other (Hills, Francis, & Robbins, 2005), meaning that an individual can be high in both intrinsic and quest religiosity, for example. This limitation can be overcome through the use of person-centered analyses.

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

1.1. Person-centered approaches Person-centered analyses can identify subpopulations, or subgroups, of individuals who share particular attributes. An increasing number of studies have utilized person-centered analyses in the examination of religiosity including cluster analysis (Fife et al., 2011; Halama, 2015) and latent class or latent profile analysis (Park et al., 2013; Pearce, Foster, & Hardie, 2013; Salas-Wright, Vaughn, Hodge, & Perron, 2012; Salas-Wright, Vaughn, & Maynard, 2014). Latent class or latent profile analysis (the former typically reserved when using categorical indicators and the latter when using continuous indicators) has several strengths over cluster analytic approaches. Unlike cluster analysis, latent profile analysis assigns class membership probabilistically, which correctly accounts for and quantifies the degree of classification error. Also, the sample size of latent classes is taken into account when assigning probabilistic class membership such that an individual with scores between two classes is noted to more likely be in the larger class than the smaller class. Finally, there is a range of statistical tests and fit indices to determine the ideal number of classes to most parsimoniously explain population heterogeneity. For these reasons, we focus on studies using latent class/profile analysis. In a nationally representative sample of adults (i.e., 18 years or older), Park et al. (2013) found 4 subgroups based on measures of religious service attendance (1 item), prayer (1 item), positive religious coping (3 items), and daily spiritual experiences (6 items), which they described as highly, moderately, somewhat, and minimally religious groups. They found that the highly religious group reported the highest self-perceived health, general happiness, and financial satisfaction, and the lowest psychological distress. Using the National Survey on Drug Use and Health (NSDUH; Substance Abuse and Mental Health Services Administration [SAMSHA], 2009; Salas-Wright et al. (2012) found five subgroups of adolescents based on five religiosity items: religious service attendance, participation in faith-based activities, importance of religious beliefs, degree to which religious beliefs influence decisions, and degree to which it is important that peers share the same religious beliefs. In addition to finding very low (“disengaged”), low (“sporadic”), moderate (“regulars”), and high (“devoted”) groups, they found a “privately religious” group that was low on participation in religious activities but high on the other indicators. Their most consistent findings was that the high religiosity group reported lower likelihood of using several substances (i.e., alcohol, marijuana, cocaine/crack, hallucinogens), and lower likelihood of fighting and stealing. Despite using different indicators of religiosity, Salas-Wright et al. (2014) found four subgroups among emerging adults (i.e., ages 18 to 25) using data from both the NSDUH (SAMSHA, 2011) and the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC; Grant et al., 2003). In NSDUH, indicators included religious service attendance, importance of religious beliefs, degree to which religious beliefs influence decisions, and degree to which it is important that peers share the same religious beliefs. In NESARC, indicators included religious service attendance, religious social engagement, and importance of religious beliefs in daily life. Although they described the two intermediate (i.e., low and moderate) groups differently across the two samples, both datasets found very low, low, moderate, and high groups. The high religiosity groups reported substantially less criminal behaviors ranging from antisocial behaviors (i.e., stealing, selling drugs), substance use behaviors (i.e., tobacco, alcohol, marijuana, etc.), and substance abuse/dependence (i.e., nicotine, alcohol, marijuana, illicit drugs). Across multiple studies examining religiosity-related constructs, four or five class solutions predominate despite different sample sizes and numbers of indicators (Park et al., 2013; Pearce et al., 2013; Salas-Wright et al., 2012; Salas-Wright et al., 2014). As most of the previous studies have come from large, multi-purpose epidemiological studies, they have a strength in being from nationally representative

161

samples, but a weakness in the ability to comprehensively assess religiosity (studies above used 4–12 items). They have also focused on aspects of intrinsic and extrinsic religiosity, and have not examined quest religiosity. 1.2. Study 1 purpose The purpose of the present study was to identify subpopulations of individuals defined by the three religious orientations described above: intrinsic, extrinsic, and quest. Specifically, we use latent profile analysis to determine the number of distinct religiosity subpopulations in our sample of Christian college students. Next, we examine how these distinct groups differ on a host of religiosity-related constructs (i.e., personal religious struggle, religious commitment, positive religious coping, negative religious coping, and purpose in life) and psychological health outcomes (i.e., depressive symptoms, anxiety symptoms, rumination, alcohol consumption, and alcohol-related problems), which have been found to be linked to religiosity among religious college students. 2. Study 1 method 2.1. Participants and procedure Participants were undergraduate students recruited from a Psychology Department participant pool at a large, southeastern university in the United States to complete an online survey for research participation credit. To have access to the participant pool, students had to be at least 18 years old and enrolled in a psychology course. From 772 total participants, we used the data from 530 students who selfidentified as Christian. Most participants were female (n = 398, 75.1%), identified as being either White (n = 246, 46.4%) or AfricanAmerican (n = 219, 41.3%), and reported a mean age of 21.75 (SD = 5.35) years. With regards to Christian denomination, most participants identified as either Baptist (n = 216, 40.8%) or Catholic (n = 90, 17%). The study was approved by the institutional review board at the participating institution. 2.2. Measures For all measures, composite scores were created by averaging items and reverse-coding items when appropriate such that higher scores indicate higher levels of the construct. The bivariate correlations, descriptive statistics, and internal consistency measures for all variables in Study 1 are shown in Table 1. 2.2.1. Religiosity Religiosity was assessed using the 24-item Revised Religious Life Inventory (RLI-R; Hills et al., 2005), which is measured on a 9-point response scale ranging from 1 (Strongly disagree) to 9 (Strongly agree). The RLI-R assesses the extrinsic (7 items; e.g., “The church is most important as a place to formulate good social relationships”; α = .84), intrinsic (9 items; e.g., “I try hard to carry my religion over into all my other dealings in life”; α = .93), and quest (8 items; e.g., “I am constantly questioning my religious beliefs”; α = .89) orientations of religiosity. The RLI-R has shown good to excellent reliability and convergent validity has been demonstrated by correlations with the original Religious Life Inventory (Batson & Schoenrade, 1991b) r = .89 for intrinsic and r = .96 for extrinsic and quest scales (Hills et al., 2005). 2.2.2. Personal religious struggle Religious commitment and religious struggle were assessed using the College Student's Beliefs and Values Survey (CSBV; Astin et al., 2011). The CSBV consists of 12 “scales” that assess student's spiritual and religious orientations. The present study only examined the religious commitment and religious struggle subscales. Religious

162

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

Table 1 Bivariate correlations and descriptive statistics among all study variables in Study 1. 1 1. Extrinsic 2. Intrinsic 3. Quest 4. Personal religious struggle 5. Religious commitment 6. Positive religious coping 7. Negative religious coping 8. Purpose in life 9. Depressive symptoms 10. Anxiety symptoms 11. Rumination 12. Alcohol consumption 13. Alcohol-related problems

.84 .51 .30 −.15 .37 .36 .08 .15 .05 .05 .11 −.07 .02

2 .93 .18 −.29 .74 .64 −.07 .20 .07 .02 .09 −.23 −.12

3

.89 .44 −.02 .01 .28 −.07 .23 .14 .19 .01 .03

4

.79 −.32 −.24 .46 −.36 .35 .22 .17 .04 .15

5

.95 .73 −.16 .25 −.01 .04 .14 −.21 −.14

6

.95 .04 .22 .03 .11 .22 −.22 −.09

7

.93 −.33 .26 .08 .11 .02 .10

8

.83 −.40 −.17 −.02 .01 −.12

9

.93 .41 .36 −.01 .21

10

.94 .60 −.07 .06

11

.96 −.00 .07

12

.94 .56

13

M

SD

.94

4.66 5.30 4.17 1.65 3.09 3.72 1.83 4.03 0.70 3.21 4.60 0.77 0.13

1.74 2.10 1.85 0.45 0.66 1.09 0.99 0.69 0.58 0.93 1.20 1.11 0.19

Note. Significant correlations (p b .05) are bolded for emphasis. Cronbach's alphas are underlined and shown on the diagonal.

commitment has 11 items and assesses the degree to which an individual has a strong dedication to their religion and follows their creed fully (e.g., “My religious beliefs are one of the most important things in my life”; α = .95). Religious struggle has 7 items and assesses the degree to which an individual is struggling with their current religion (e.g., “I am feeling unsettled about religious matters”; α = .79). An examination of the psychometric properties of the measure revealed that the CSBV exhibited good psychometric properties and is a valid measure of spirituality and religiosity (Astin et al., 2011). 2.2.3. Religious coping Religious coping was measured using the Brief RCOPE (Pargament et al., 1998). The Brief RCOPE is a self-report measure that consists of 14 items that assesses positive and negative coping strategies and uses a Likert-type, 5-point scale ranging from 1 (not at all) to 5 (a great deal). The participants were instructed to indicate how typically they use the coping response when faced with stressful or negative events. The positive subscale consists of seven items reflecting seven coping strategies, such as collaborative religious coping, (e.g., “tried to put my plans into action together with God”; α = .95). The negative subscale consists of seven items reflecting seven coping strategies, such as demonic reappraisals, (e.g., “decided the devil made this happen”; α = .93). Pargament et al. (2011) conducted a meta-analysis of the psychometrics of the measure and found the measure has demonstrated good predictive validity, incremental validity, and concurrent validity since its creation roughly 15 years ago. 2.2.4. Purpose in life Purpose in life was assessed using the 6-item Life Engagement Test (LET; Scheier et al., 2006), which is measured on a 5-point response scale (1 = Strongly Disagree, 5 = Strongly Agree). Example items include “To me, the things I do are all worthwhile” and “I value my activities a lot” (α = .83). Test-retest correlations ranged from .61 to .76 and demonstrated satisfactory convergent reliability with other constructs similar to purpose in life (Scheier et al., 2006). 2.2.5. Depressive symptoms Depressive symptoms were assessed using the 20-item Center for Epidemiological Studies Depression-Revised (CESD-R; Van Dam & Earleywine, 2011) measured on a 5-point response scale (1 = Not at all or Less than 1 day, 2 = 1–2 days, 3 = 3–4 days, 4 = 5–7 days, 5 = Nearly Every day for 2 weeks). As advised by Van Dam and Earlywine (2011), the ‘5–7 days’ and ‘nearly every day…’ were collapsed into the same value. Example items include, “Nothing made me happy” and “I could not get going” (α = .93). An examination of the psychometric properties of the measure revealed that the CESD-R exhibited good psychometric properties and is an accurate and valid measure of depression (Van Dam & Earleywine, 2011).

2.2.6. Anxiety symptoms Anxiety symptoms (i.e., worry) were assessed using the 16-item Penn State Worry Questionnaire (PSWQ; Meyer et al., 1990) measured on a 5-point response scale (1 = not at all typical of me, 5 = very typical of me). Example items include, “Many situations make me worry” and “I have been a worrier all my life” (α = .94). The PSWQ has demonstrated good discriminant, divergent, and convergent validity (Molina & Borkovec, 1994). 2.2.7. Rumination Rumination was assessed using the 20-item Ruminative Thought Style Questionnaire (RTSQ; Brinker, & Dozois, 2009) measured on a 7-point response scale (1 = Not All Descriptive of Me, 7 = Describes Me Very Well). Example items include, “I find that my mind often goes over things again and again” and “When I am looking forward to an exciting event, thoughts of it interfere with what I am working on” (α = .96). 2.2.8. Alcohol consumption Alcohol consumption was measured using a modified version of the Daily Drinking Questionnaire (Collins et al., 1985), which uses a sevenitem (Monday through Sunday) grid to assess number of standard drinks consumed during a typical drinking week in the past 30 days. We summed number of standard drinks consumed on each day of the typical drinking week (α = .94). 2.2.9. Alcohol-related problems Alcohol-related problems were assessed using the 48-item Young Adult Alcohol Consequences Questionnaire (YAACQ; Read et al., 2006). Individuals responded on a checklist response format indicating whether they experienced a specific consequence in the past month. Non-drinkers were given a score of ‘0’ to reflect the absence of alcohol-related problems. Example items include, “I have had a hangover (headache, sick stomach) the morning after I had been drinking” and “I have passed out from drinking” (α = .94). 3. Study 1 results 3.1. Class solution As recommended by previous research (Marsh et al., 2009; Henson et al., 2007), we relied on goodness-of-fit indexes, such as the Akaike Information Criterion (AIC; Akaike, 1973, 1974; Sakamoto et al., 1986) and Bayesian Information Criterion (BIC; Schwarz, 1978), as well as tests of statistical significance to settle upon the number of latent classes. Specifically, to determine the number of latent classes in our sample based on the pattern of means of the three subscales (i.e., extrinsic, intrinsic, and quest religiosity) of the RLIR, we used the

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

Lo-Mendell-Rubin Adjusted Likelihood Ratio Test (Lo et al., 2001; Vuong, 1989), which compares whether a k class solution fits better than an k − 1 class solution. The Likelihood Ratio Test suggests that a 2-class solution fit better than a 1-class solution (p = .0001), a 3-class solution fit better than a 2-class solution (p = .0027), a 4-class solution fit better than a 3-class solution (p = .0144), but a 5-class solution did not fit significantly better than a 4-class solution (p = .3162), however a 6-class solution did fit better than a 5-class solution (p = .0136). A 7class solution did not fit better than a 6-class solution (p = .1172). Further, the AIC continued to improve (i.e., decrease) from 1 through 7 class solutions, the BIC and adjusted BIC decreased from 1 through 6 class solutions and increased for a 7 class solution (see Table 2). Though the Likelihood Ratio Test may indicate either a 4-class or 6-class solution, Nagin (2005) suggest that the smallest group should not contain around or less than 5% of the sample. Compared to a 6-class solution (7.35%), the smallest class from a 4-class solution encompasses a much larger proportion of individuals from the sample (13.21%). Moreover, researchers recommend selecting the number of classes based on theory, previous research, and interpretation of the results (Marsh et al., 2009; Nylund et al., 2007). Thus, based on previous research (Park et al., 2013; Salas-Wright et al., 2014), more individuals in the smallest group, and the interpretation of the results, we settled on the 4-class solution. The entropy value of .750 indicates that it is estimated that threefourths of subjects were correctly classified in the appropriate latent class, which approaches a level of entropy that is considered high (i.e., 80, Clark & Muthén, 2009). Fig. 1 depicts the pattern of means across the latent classes. Class 1 comprised 15.99% of the sample (N = 84.74), and we tentatively label this class the “low religiosity” group as they were the moderately low on all three subscales of religiosity. Class 2 comprised 25.44% of the sample (N = 134.81), and we label this class the “high religiosity” group as they were relatively high on all three subscales of religiosity. Class 3 comprised 14.11% of the sample (N = 74.79), and we label this class the “intrinsically motivated” group as they were high on intrinsic religiosity, but relatively low on extrinsic and quest religiosity. Finally, the largest group, Class 4, comprised 44.46% of the sample (N = 235.66), and we label this class the “questioning” group as they were high on quest religiosity, but relatively low on extrinsic and intrinsic religiosity. 3.2. Equality of means Upon settling on a 4-class solution, we then tested the equality of means across latent classes on religiosity-related constructs (i.e., personal religious struggle, religious commitment, positive religious coping, negative religious coping, and purpose in life)

163

and psychological health outcomes (i.e., depressive symptoms, anxiety symptoms, rumination, alcohol consumption, and alcoholrelated problems) using pseudo-class-based multiple imputations (Asparouhov and Muthén, 2007). Rather than assigning individuals to the latent class where their membership has the highest probability and conducting traditional techniques like analysis of variance (ANOVA), this method accounts for the probabilistic nature of class membership, and both global and pairwise comparisons can be conducted using Wald tests (see Table 3). To maximize interpretability of mean differences, all variables were converted to z-scores. Thus, a mean difference of one indicates a one standard deviation difference. 3.2.1. Religiosity-related constructs Compared to the “low religiosity” and “questioning” groups, we found that the “intrinsically motivated” and “high religiosity” groups had higher religious commitment, religious coping, and purpose in life, but lower personal religious struggle and negative religious coping. Further, the “high religiosity” and “intrinsically motivated” groups did not significantly differ from each other on any of these constructs; nor were there any statistical differences between the “questioning” and “low religiosity” groups on most of these constructs (see Table 3). 3.2.2. Psychological health outcomes Surprisingly, we did not find any significant differences across the groups on depressive symptoms and anxiety symptoms (see Table 3). With regards to rumination, the only significant difference was found between the “low religiosity” and “high religiosity” groups. Specifically, the “low religiosity” group had a more adaptive profile (i.e., lower rumination) compared to the “high religiosity” group (i.e., higher rumination). However, we did find discrepancies between the groups on alcohol outcomes. We found that the “intrinsically motivated” group had significantly lower alcohol consumption than all other groups and significantly lower alcohol-related problems than the “low religiosity” and “questioning” groups. Further, the “high religiosity” group had significantly lower alcohol consumption than the “low religiosity” and “questioning” groups. 4. Study 1 discussion The present study aimed to examine the extent to which personcentered analyses (i.e., latent profile analysis) could identify distinct subpopulations of college students based on their religious orientations. In our sample of Christian college students, we found four subgroups with conceptually distinct profiles: Questioning (high on quest religiosity, low on intrinsic and extrinsic religiosity), High Religiosity (high on

Table 2 Fit statistics for 1 through 7 class solutions for Latent Profile Analysis (LPA) across both studies. Number of classes — Study 1 Fit statistics

1

2

3

4

5

6

7

AIC BIC Adjusted BIC Entropy Smallest n

6497.73 6523.37 6504.32 – 530

6328.03 6370.75 6339.01 0.711 131

6255.07 6314.89 6270.45 0.759 84

6203.31 6280.23 6223.09 0.750 70

6184.39 6278.40 6208.56 0.740 48

6158.26 6269.36 6186.83 0.819 39

6154.06 6282.25 6187.02 0.760 1

Number of classes — Study 2 Fit statistics

1

2

3

4

5

6

7

AIC BIC Adjusted BIC Entropy Smallest n

5720.47 5745.32 5726.28 – 465

5596.63 5638.05 5606.31 0.604 183

5517.61 5575.60 5531.17 0.771 53

5451.58 5526.14 5469.01 0.771 59

5431.08 5522.20 5452.38 0.810 11

5415.14 5522.83 5440.32 0.751 23

5404.71 5528.97 5433.76 0.782 9

Note. AIC = Akaike Information Criterion, BIC = Bayesian Information Criterion.

164

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

Fig. 1. Depiction of the four latent classes defined by pattern of standardized means on three religiosity subscales in Study 1.

all three religious orientations), Low Religiosity (moderately low on all three religious orientations), and Intrinsically Motivated (high on intrinsic religiosity, relatively low on extrinsic and quest religiosity). The largest subgroup was the Questioning group and the smallest subgroup was the Intrinsically Motivated group, which aligns with previous research showing that behavioral aspects of religiosity decline during the early college years (Stoppa & Leftkowitz, 2010). On religiosity-related constructs, the High Religiosity and Intrinsically Motivated groups were differentiated from the Low Religiosity and Questioning groups as they were higher in religious commitment, positive religious coping, and purpose in life; further, they were lower in personal religious struggle and negative religious coping. On emotional health outcomes, we found no significant differences across these four groups. The High Religiosity group reported significantly more rumination than the Low Religiosity group. Finally, the Intrinsically Motivated group reported lower alcohol use and problems than the Low Religiosity, High Religiosity, and Questioning groups. Thus, our results suggest that the Intrinsically Motivated group demonstrated the most adaptive outcomes on most measures, though we did not find any differences on symptoms of depression or anxiety (i.e., worry).

One of the greatest strengths of latent profile analysis (as with all person-centered analyses) is its ability to identify population heterogeneity in a purely data-driven way. Thus, it is inherently an exploratory data analytic technique, thus the number of classes identified in one sample may not replicate in another sample. As the first study to use LPA examining these three religious orientations among college students, it is unknown whether the 4-class solution that emerged in our study is likely or unlikely to replicate in an independent sample. Further, we were largely unable to distinguish between the Questioning and Low Religiosity groups or between the Intrinsically Motivated and High Religiosity groups on outcomes, suggesting that additional variables are needed to explain differences across these groups. 5. Study 2 purpose The purpose of Study 2 was to determine whether we could replicate the 4-class solution found in Study 1 in an independent sample of Christian college students. Further, we wanted to examine a wider range of constructs that may be able to distinguish between the groups observed in Study 1. Thus, in addition to all the outcomes assessed in Study 1, we also examined other psychological health factors (i.e., distress tolerance,

Table 3 Mean comparisons between latent classes on religiosity subscales, religiosity-related constructs, and psychological health outcomes in Study 1. Standardized scores (z-scores) Class 1: Low religiosity

Class 2: High religiosity

Class 3: Intrinsically motivated

Class 4: Questioning

−0.531c 1.034b −0.320c

0.005d −0.335c 0.241b

Religiosity subscales Extrinsic Intrinsic Quest

−1.257a −1.455a −0.794a

Religiosity-related constructs Personal religious struggle Religious commitment Positive religious coping Negative religious coping Purpose in life

0.208a −1.106a −1.062a 0.045ab −0.317a

−0.317b 0.661b 0.604b −0.037a 0.206a

−0.389b 0.777b 0.587b −0.289ab 0.193a

0.240a −0.242c −0.199c 0.100b −0.066ab

Psychological health outcomes Depressive symptoms Anxiety symptoms (worry) Rumination Alcohol consumption Alcohol-related problems

−0.210a −0.165a −0.226a 0.238a 0.090a

0.007a 0.019a 0.115b −0.147b −0.042ab

0.048a −0.062a −0.010ab −0.377c −0.264b

0.056a 0.067a 0.018ab 0.117ad 0.075a

1.072b 0.930b 0.238b

Note. Means sharing a subscript in a row indicate means that are not significantly different from each other. Mean comparisons of the raw scores are available from the authors upon request.

16

17

.94 .56 −.46 .94 .47 −.25 .94 .11 .03 −.13 .94 .53 −.09 .07 −.04

18

19

Participants were recruited using the same procedures as Study 1. From 699 total participants, we used the data from 465 students who self-identified as Christian. Most participants were White (n = 221, 47.5%) or African-American (n = 208, 44.7%), female (n = 333, 71.6%), and reported a mean age of 22.18 (SD = 6.54) years. With regards to Christian denomination, most participants identified as either Baptist (n = 185, 39.8%) or Catholic (n = 87, 18.7%). The study was approved by the institutional review board at the participating institution.

.96 −.12 .09 .46 −.31 −.14

20

6.1. Participants and procedure

.93 .49 −.05 .12 .43 −.46 −.20

SD M 22 21

6. Study 2 method

4.87 5.48 4.32 1.84 3.26 3.93 2.46 3.96 4.14 3.95 4.39 4.35 4.42 4.21 0.65 3.05 4.23 0.57 0.14 2.50 3.82 3.75

self-regulation, and psychological flexibility) and psychological wellbeing (i.e., autonomy, environmental mastery, personal growth, positive relationships, purpose in life, self-acceptance), which have been found to be linked to religiosity (see McCullough & Willoughby, 2009 for a review).

165

1.81 2.12 1.86 0.72 0.73 1.32 1.61 0.76 0.80 0.60 0.83 0.83 0.83 0.92 0.58 0.89 1.18 0.94 0.20 0.90 0.58 0.57

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

.93 .38 .31 .05 .17 .40 −.39 −.41 .81 −.42 −.29 −.17 .01 −.12 −.31 .53 .56 .75 .72 −.32 −.08 −.01 −.07 −.11 −.18 .41 .60 .72 .71 .75 −.31 −.16 −.05 .03 −.06 −.22 .45 .53 .73 .72 .81 .67 −.26 −.11 −.03 −.03 −.11 −.19 .43 .58 .37 .60 .66 .65 .74 −.32 −.24 −.17 .03 −.11 −.28 .53 .58 .67 .61 .64 .61 .62 .69 −.30 −.31 −.15 −.03 −.09 −.29 .46 .59 .84 .58 .58 .65 .64 .69 .72 −.38 −.18 −.04 −.06 −.11 −.18 .42 .60 .96 −.32 −.30 −.25 −.37 −.38 −.40 −.31 .15 .02 .06 .04 .03 .13 −.16 −.26 .96 .47 .14 .08 .06 .01 .04 .04 .11 −.01 −.03 .15 −.05 −.08 .12 .05 .05 .94 .65 .22 .16 .09 .10 .08 .09 .12 .14 .02 .01 .10 −.01 −.07 .12 .05 .09 .87 .24 .28 .60 −.32 −.32 −.22 −.28 −.33 −.32 −.31 .23 .15 .13 .04 .06 .18 −.13 −.23 .87 .28 .07 .09 .21 −.14 −.18 −.23 −.08 −.14 −.16 −.20 .22 .08 .19 .01 .02 .20 −.10 −.12

8 7 6 5 4 3 2

Note. Significant correlations (p b .05) are bolded for emphasis. Cronbach's alphas are underlined and shown on the diagonal.

6.2.5. Psychological well being Psychological Well Being was assessed using the 42-item Psychological Well-Being Questionnaire (PWB; Ryff, 1989) measured on a 6-point response scale (1 = Strongly Disagree, 6 = Strongly Agree). The measure assesses six subscales of psychological well-being: autonomy (e.g., “I judge myself by what I think is important, not by the values of what others think is important”; α = .67), environmental mastery (e.g., “In

1

6.2.4. Self-regulation Self-Regulation was assessed using the 31-item Short Self-Regulation Questionnaire (Carey et al., 2004) measured on a 5-point response scale (1 = Strongly Disagree, 5 = Strongly Agree). Example items include, “Once I have a goal, I can usually plan how to reach it” and “I tend to keep doing the same thing, even when it doesn't work” (α = .94).

Table 4 Bivariate correlations and descriptive statistics among all study variables in Study 2.

6.2.3. Psychological flexibility Psychological Flexibility was assessed using the 16-item Acceptance and Action Questionnaire (AAQ; Hayes et al., 2004) measured on a 7-point response scale (1 = Never True, 7 = Always True). Example items include, “I am able to take action on a problem even if I am uncertain what is the right thing to do” and “It's OK to feel depressed or anxious” (α = .56).

9

6.2.2. Distress intolerance Distress intolerance was assessed using the 15-item Distress Tolerance Scale (DTS, Simons & Gaher, 2005) measured on a 5-point response scale (1 = Strongly Agree, 5 = Strongly Disagree). Example items include, “My feeling of distress are so intense that they completely take over” and “I'll do anything to avoid feeling distressed or upset” (α = .94).

.93 .23 −.16 .51 .43 −.15 .27 .21 .09 .19 .19 .23 .21 .01 −.03 .07 −.11 −.08 .12 .05 .19

14 13 10

11

12

6.2.1. Alcohol-related problems Alcohol-related problems were assessed using the 24-item BriefYoung Adult Alcohol Consequences Questionnaire (B-YAACQ; Kahler et al., 2005). Individuals responded on a checklist response format indicating whether they experienced a specific consequence in the past month. Non-drinkers were given a score of ‘0’ to reflect the absence of alcohol-related problems (α = .94).

.84 .52 .30 −.01 .20 .22 .06 .08 −.05 −.10 −.02 −.03 .00 .20 −.01 .02 .10 −.07 .02 .19 −.10 .05

Most measures in Study 2 were the same as in Study 1. We describe only measures unique to Study 2 below. The bivariate correlations, descriptive statistics, and internal consistency measures for all variables in Study 2 are shown in Table 4. It is important to note that the reliabilities from measures used in Study 1 were similar in strength in study 2.

1. Extrinsic 2. Intrinsic 3. Quest 4. Personal religious struggle 5. Religious commitment 6. Positive religious coping 7. Negative religious coping 8. Purpose in life 9. Autonomy 10. Environment mastery 11. Personal growth 12. Positive relations 13. Purpose in life 14. Self-acceptance 15. Depressive symptoms 16. Anxiety symptoms 17. Rumination 18. Alcohol consumption 19. Alcohol-related problems 20. Distress intolerance 21. Psyc. flexibility 22. Self-regulation

15

6.2. Measures

166

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

Fig. 2. Depiction of the four latent classes defined by pattern of standardized means on three religiosity subscales in Study 2.

general, I feel I am in charge of the situation in which I live”; α = .37), personal growth (e.g., “I have the sense that I have developed a lot as a person over time”; α = .73), positive relations (e.g., I enjoy personal and mutual conversations with family members or friends”; α = .72), purpose in life (e.g., “I have a sense of direction and purpose in life”; α = .75), and self-acceptance (e.g., “In general, I feel confident and positive about myself”; α = .81). 7. Study 2 results 7.1. Class solution The Lo-Mendell-Rubin Adjusted Likelihood Ratio Test (Lo et al., 2001; Vuong, 1989) suggests that a 2-class solution fit better than a 1-

class solution (p = .0058), a 3-class solution fit better than a 2-class solution (p = .0054), a 4-class solution fit better than a 3-class solution (p = .0068), but a 5-class solution did not fit significantly better than a 4-class solution (p = .1003), however a 6-class solution did fit better than a 5-class solution (p = .0101). A 7-class solution did not fit better than a 6-class solution (p = .0728). Further, similar to Study 1, the AIC and adjusted BIC decreased from 1 through 7 class solutions, and the BIC decreased from 1 through 6 class solutions, but increasing for a 7 class solution (see Table 2). Once again, though the Likelihood Ratio Test may indicate either a 4-class or 6-class solution, Nagin (2005) suggest that the smallest group should not contain around or less than 5% of the sample. Compared to a 6-class solution (4.95%), the smallest class from a 4-class solution encompasses a much larger proportion of individuals from the sample (12.69%). Moreover, researchers recommend

Table 5 Mean comparisons between latent classes on religiosity subscales, religiosity-related constructs, psychological well-being, and psychological health outcomes in Study 2. Standardized scores (z-scores) Class 1: Low religiosity

Class 2: High religiosity

Class 3: Intrinsically motivated

Class 4: Questioning

Religiosity subscales Extrinsic Intrinsic Quest

−1.115a −1.467a −1.018a

1.114b 1.085b 1.397b

0.307c 0.927b −0.636c

−0.138d −0.341c 0.238d

Religiosity-related constructs Personal religious struggle Religious commitment Positive religious coping Negative religious coping Purpose in life

−0.018a −0.803a −0.626a 0.054a −0.247a

0.193a 0.383b 0.433b 0.214a −0.022a

−0.415b 0.545b 0.437b −0.414b 0.503b

0.166a −0.150c −0.160c 0.142a −0.183a

Psychological well-being Autonomy Environmental mastery Personal growth Positive relations Purpose in life Self-acceptance

−0.136a 0.045a −0.264a −0.157a −0.273a −0.150a

−0.210a −0.368b −0.184a −0.170a −0.266a −0.169a

0.536b 0.442c 0.442b 0.455b 0.577b 0.490b

Psychological health outcomes Depressive symptoms Anxiety symptoms (worry) Rumination Alcohol consumption Alcohol-related problems Distress intolerance Psychological flexibility Self-regulation

−0.236a −0.104a −0.348a 0.293a 0.119a −0.446a −0.068a −0.206a

0.290b 0.186a 0.361b 0.021ab 0.180a 0.434b −0.307a −0.165a

−0.197a −0.097a −0.061ab −0.176b −0.142a −0.082c 0.219b 0.466b

−0.181a −0.142ab −0.102a −0.143a −0.147a −0.164a 0.091b 0.031a 0.361b −0.001ab −0.011a 0.057cd −0.010a −0.134a

Note. Means sharing a subscript in a row indicate means that are not significantly different from each other. Mean comparisons of the raw scores are available from the authors upon request.

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

selecting the number of classes based on theory, previous research, and interpretation of the results (Marsh et al., 2009; Nylund et al., 2007). Thus, based on previous research (Salas-Wright et al., 2014; Park et al., 2013), more individuals in the smallest group, and the interpretation of the results, we again settled on the 4-class solution. The entropy value of .771 indicates that it is estimated that about three-fourths of subjects were correctly classified in the appropriate latent class, which approaches a level of entropy that is considered high (i.e., 80, Clark & Muthén, 2009). The profiles of each class were markedly similar to Study 1 (see Fig. 2). Class 1 comprised 14.21% of the sample (N = 66.05), and we tentatively label this class the “low religiosity” group as they were the moderately low on all three subscales of religiosity. Class 2 comprised 12.97% of the sample (N = 60.32), and we label this class the “high religiosity” group as they were relatively high on all three subscales of religiosity. Class 3 comprised 24.53% of the sample (N = 114.07), and we label this class the “intrinsically motivated” group as they were high on intrinsic religiosity, but relatively low on extrinsic and quest religiosity. Finally, the largest group, Class 4, comprised 48.29% of the sample (N = 224.56), and we label this class the “questioning” group as they were high on quest religiosity, but relatively low on extrinsic and intrinsic religiosity. 7.2. Equality of means In an attempt to replicate our findings from Study 1, we once again tested the equality of means across the 4-latent classes on religiosityrelated constructs and psychological health outcomes using pseudoclass-based multiple imputations (see Table 5). In furthering our findings from Study 1, we also tested the equality of means across the 4-latent classes on psychological well-being outcomes (i.e., autonomy, environmental mastery, personal growth, positive relationships, purpose in life, self-acceptance), and other psychological health outcomes (i.e., distress intolerance, psychological flexibility, and self-regulation). 7.2.1. Religiosity-related constructs Similar to Study 1, we found that the “intrinsically motivated” group had statistically higher religious commitment, religious coping, and purpose in life, but lower personal religious struggle and negative religious coping compared to the “low religiosity” and “questioning” groups. Further, “questioning” and “low religiosity” groups did not differ on most of these constructs (i.e., both being lower in purpose in life, lower in religious commitment, and higher in negative religious coping). However, divergent findings arose with the “high religiosity” group. Specifically, compared to Study 1, the “high religiosity” group had high negative religious coping, high personal religious struggle, and low purpose in life; although still high religious commitment and positive religious coping.

167

higher psychological flexibility than the “high religiosity” group. Finally, with regards to distress tolerance, we found that the “low religiosity” group had significantly lower distress intolerance than all other groups. 7.2.3. Psychological well-being Across each psychological well-being outcome, we found that the “intrinsically motivated” group had the most adaptive profile (i.e., higher autonomy, environmental mastery, personal growth, positive relations, purpose in life, and self-acceptance) and was significantly different than all other groups, which did not significantly differ from each other on most of these outcomes (see Table 5). 8. Study 2 discussion In an attempt to replicate and extend the findings from Study 1, we conducted latent profile analysis in an independent sample of Christian college students. We replicated a conceptually meaningful 4-class solution that very closely matched the four classes observed in Study 1: Questioning, High Religiosity, Low Religiosity, and Intrinsically Motivated. We again found Questioning to be the largest group and within this study, with a broader set of outcome measures, we were able to further distinguish between these groups. The “Intrinsically Motivated” group reported high levels of intrinsic religiosity, moderately high levels of extrinsic religiosity, and low levels of quest religiosity. The “Intrinsically Motivated” group showed the most adaptive profile on most measures including high positive religious coping, high purpose in life, high psychological well-being (i.e., across all indices), high self-regulation abilities, high psychological flexibility, and low rumination. This group is highly committed to and has low personal struggles with their religion, which may help explain their relatively high psychological well-being. The “High Religiosity” group reported the highest levels of intrinsic, extrinsic, and quest religiosity. The “High Religiosity” group appears to be of two minds in relation to religion as they appeared both committed to religion but also struggling with religion; they reported using the highest negative religious coping, but also reported high levels of positive religious coping. They reported similarly low levels of psychological well-being as the “Low Religiosity” and “Questioning” groups compared to the “Intrinsically Motivated” group. This group also had the highest levels of depressive symptom, which was significantly higher than the “Low Religiosity” and “Intrinsically Motivated” groups. The “Low Religiosity” group did not differ from “High Religiosity” and “Questioning” groups on most measures, but had adaptive associations with psychological health and had the lowest distress intolerance compared to all other groups, suggesting that they have the strongest abilities to tolerate distressing thoughts and emotions. 9. Overall discussion

7.2.2. Psychological health outcomes We once again did not find any significant differences across the groups on anxiety symptoms. However, divergent findings arose on alcohol outcomes, depressive symptoms, and rumination. Unlike in Study 1, we did not find any significant differences across the groups on alcohol-related problems. Further, with regards to alcohol consumption, the only significant difference was found between the “low religiosity” and “intrinsically motivated” groups such that the “intrinsically motivated” group had lower alcohol consumption compared to the “low religiosity” group. Across depressive symptoms and rumination, we found that the “intrinsically motivated” and the “low religiosity” groups had lower depressive symptoms and rumination, and did not significantly differ from each other on either outcome. With regards to self-regulation we found that the “intrinsically motivated” group had significantly higher self-regulation that all other groups. With regards to psychological flexibility, the only significant difference was found between the “intrinsically motivated” and “high religiosity” groups such that “intrinsically motivated” group had a

In terms of examining the construct of religiosity, traditional variable-centered approaches in religious college-students have found (unique) associations between single facets of religiosity (i.e., intrinsic, extrinsic, and quest religiosity) and outcomes (Maltby & Day, 2000; Steger et al., 2010). However, this approach does not take into consideration that these religious orientations are not mutually exclusive (Hills et al., 2005) and one of the limitations of variable-centered analyses is that they assume that all participants have been sampled from a single population (i.e., population homogeneity assumption, Collins & Lanza, 2010). Within two independent studies on Christian college students, we conducted latent profile analyses (LPAs) to determine the number of distinct religiosity subpopulations based on their religious orientations (intrinsic, extrinsic, and quest). Across both studies, we selected and replicated a 4-class solution to parsimoniously describe groups of Christian college students. Although each class seems very similar across studies, the most notable difference is that the “High Religiosity” group in Study 1 had

168

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169

higher intrinsic and extrinsic religiosity compared to quest religiosity whereas the “High Religiosity” group in Study 2 had higher quest religiosity compared to intrinsic and extrinsic religiosity. In Study 1, the “High Religiosity” group showed a similarly adaptive profile on most outcome measures as the “Intrinsically Motivated” group; however, in Study 2, the “High Religiosity” group showed the most maladaptive profile on most outcome measures. These findings corroborate studies suggesting that higher quest religiosity or personal religious struggle tends to be associated with poor psychological health (Bravo et al., 2015; Hill & Pargament, 2008; Steger et al., 2008; Steger et al., 2010). Consistently across both studies, the “Intrinsically Motivated” group appeared to exhibit the most psychologically healthy profile on a wide range of outcome variables, corroborating studies suggesting that intrinsic religiosity is a protective factor associated with good psychological health (Berry & York, 2011; Jansen et al., 2010; Stewart, 2001; Wood & Herbert, 2005). This preliminary investigation begins to show the promise of LPA in separating groups of individuals based on their profiles of religious orientations. As recommended by Hill and Pargament (2008), “Research designs and measures are needed that better capture the dynamic qualities of religion and spirituality—the possibility of change, growth, deterioration, or stability in religious and spiritual life across time and situations” (p. 12). Additional work examining longitudinal personcentered methods is needed to better examine how changes in religiosity over time relate to important outcomes. Although some work has been done using latent class growth curve analysis (Eriksson et al., 2015) and growth mixture modeling (McCullough et al., 2005), we are unaware of any studies examining changes in religiosity over time using repeated-measures latent profile analysis or latent transition analysis. For example, using ecological momentary assessment (EMA) methods, one could examine how religious orientations wax and wane over several days or weeks to better categorize individuals into groups defined by specific trajectories of “change, growth, deterioration, and stability” in religious orientations, an approach that has been applied successfully in the addictions field (McCarthy et al., 2015). 9.1. Limitations Despite our ability to replicate a very similar four-class solution across two independent samples, there are a number of limitations of the present studies. First, as our data were cross-sectional, we are unable to demonstrate temporal precedence, which is requisite for making causal inferences. Second, we do not wish to suggest that there are exactly four classes of individuals who differ in their religiosity scores in the population. This study used two independent samples of Christian college students from the same university and additional work with large samples from distinct populations (i.e., different age groups and religious groups) is needed to determine the number of classes in the broader population. Finally, although there have been connections between religiosity and mental health outcomes (see Moreira-Almeida et al., 2006 for a review), other outcomes including physical health have been associated with religiosity (Ellison & Levin, 1998; Hill & Pargament, 2008), but were not examined in the present study. Thus, it is important to examine how these latent classes differ on a broader range of physical and mental health variables. Additional work is needed to tease apart the potential mechanisms through which these latent profiles have effects on psychologically meaningful outcomes. 9.2. Conclusion Despite the limitations of the present studies, we were able to distinguish between four subgroups of individuals based on their religious orientations profiles, and found that the Intrinsically Motivated group (i.e., high on intrinsic religiosity, but relatively low on extrinsic and quest religiosity) had the most adaptive psychological outcomes. Specifically, individuals in the High Religiosity, Low Religiosity and Questioning

groups reported significantly lower levels of psychological well-being compared to the Intrinsically Motivated group. These results corroborate studies suggesting that intrinsic religiosity is a protective factor associated with good psychological health among religious college students and that personal religious struggles (i.e., quest religiosity) is associated with poorer health outcomes. Furthermore, we found support for the notion that religious orientations are not mutually exclusive, pointing to the utility of person-centered analyses to examine religiosity in unique ways. Acknowledgments Matthew R. Pearson is supported by a career development grant (K01-AA023233) from the National Institute on Alcohol Abuse and Alcoholism (NIAAA). References Akaike, H. (1973). Information theory and an extension of the maximum likelihood principle. In B. N. Petrov, & F. Casaki (Eds.), Second international symposium on information theory (pp. 267–281). Budapest: Academiai Kiado. Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19, 716–723. http://dx.doi.org/10.1109/TAC.1974.1100705. Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5, 432–443. http://dx.doi.org/10.1037/h0021212. Asparouhov, T., & Muthén, B. (2007). Computationally efficient estimation of multilevel high-dimensional latent variable models. Proceedings of the 2007 Joint Statistical Meetings, Section on Statistics in Epidemiology (pp. 2531–2535). Alexandria, VA: American Statistical Association. Astin, A. W., Astin, H. S., & Lindholm, J. A. (2011). Assessing students' spiritual and religious qualities. Journal of College Student Development, 52, 39–61. http://dx.doi.org/ 10.1353/csd.2011.0009. Batson, C. D., & Schoenrade, P. A. (1991a). Journal for the Scientific Study of Religion, 30, 416–430. http://dx.doi.org/10.2307/1387277. Batson, C. D., & Schoenrade, P. A. (1991b). Measuring religion as quest: (2) Reliability concerns. Journal for the Scientific Study of Religion, 30, 430–447. http://dx.doi.org/10. 2307/1387277. Batson, C. D., & Ventis, W. L. (1982). The religious experience. New York, NY: Oxford University Press. Batson, C. D., Schoenrade, P., & Ventis, W. L. (1993). Religion and the individual: A socialpsychological perspective. New York, NY: Oxford University Press. Berry, D. M., & York, K. (2011). Depression and religiosity and/or spirituality in college: A longitudinal survey of students in the USA. Nursing & Health Sciences, 13, 76–83. http://dx.doi.org/10.1111/j.1442-2018.2011.00584.x. Bravo, A. J., Pearson, M. R., Stevens, L. E., & Henson, J. M. (2015). Disentangling the relationship between personal religious struggle and mental health: An examination of rumination and purpose in life in a Christian college-student population. (manuscript submitted for publication). Brinker, J. K., & Dozois, D. J. A. (2009). Ruminative thought style and depressed mood. Journal of Clinical Psychology, 65, 1–19. http://dx.doi.org/10.1002/jclp.20542. Carey, K. B., Neal, D. J., & Collins, S. E. (2004). A psychometric analysis of the selfregulation questionnaire. Addictive Behaviors, 29, 253–260. http://dx.doi.org/10. 1016/j.addbeh.2007.05.004. Clark, S. L., & Muthén, B. (2009). Relating latent class analysis results to variables not included in the analysis. (submitted for publication). Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis with applications in the social, behavioural, and health sciences. Hoboken, NJ: John Wiley & Sons, Inc. Collins, R. L., Parks, G. A., & Marlatt, G. A. (1985). Social determinants of alcohol consumption: The effects of social interaction and model status on the self-administration of alcohol. Journal of Consulting and Clinical Psychology, 53, 189–200. http://dx.doi.org/ 10.1037//0022-006X.53.2.189. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. San Diego, CA: Academic Press. Ellison, C. G., & Levin, J. S. (1998). The religion-health connection: Evidence, theory, and future directions. Health Education & Behavior, 25, 700–720. http://dx.doi.org/10. 1177/109019819802500603. Eriksson, C. B., Holland, J. M., Currier, J. M., Snider, L. M., Ager, A. K., Kaiser, R. E., & Simon, W. S. (2015). Trajectories of spiritual change among expatriate humanitarian aid workers: A prospective longitudinal study. Psychology of Religion and Spirituality, 7, 13–23. http://dx.doi.org/10.1037/a0037703. Fife, J. E., Sayles, H. R., Adegoke, A. A., McCoy, J., Stovall, M., & Verdant, C. (2011). Religious typologies and health risk behaviors of African American college students. North American Journal of Psychology, 13, 313–330. Grant, B. F., Dawson, D. A., Stinson, F. S., Chou, P. S., Kay, W., & Pickering, R. (2003). The Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV (AUDADISIV): Reliability of alcohol consumption, tobacco use, family history of depression and psychiatric diagnostic modules in a general population sample. Drug and Alcohol Dependence, 71, 7–16. http://dx.doi.org/10.1016/S0376-8716(03)00070-X. Halama, P. (2015). Empirical approach to typology of religious conversion. Pastoral Psychology, 64, 185–194. http://dx.doi.org/10.1007/s11089-013-0592-y.

A.J. Bravo et al. / Personality and Individual Differences 88 (2016) 160–169 Hayes, S. C., Strosahl, K. D., Wilson, K. G., Bissett, R. T., Pistorello, J., Toarmino, D., et al. (2004). Measuring experiential avoidance: A preliminary test of a working model. The Psychological Record, 54, 553–578. Henson, J. M., Reise, S. P., & Kim, K. H. (2007). Detecting mixtures from structural model differences using latent variable mixture modeling: A comparison of relative model fit statistics. Structural Equation Modeling, 14, 202–226. Hill, P. C., & Pargament, K. I. (2008). Advances in the conceptualization and measurement of religion and spirituality: Implications for physical and mental health research. Psychology of Religion and Spirituality, 1, 3–17. http://dx.doi.org/10.1037/1941-1022. S.1.3. Hills, P., Francis, L. J., & Robbins, M. (2005). The development of the Revised Religious Life Inventory (RLI-R) by exploratory and confirmatory factor analysis. Personality and Individual Differences, 38, 1389–1399. http://dx.doi.org/10.1016/j.paid.2004.09.006. Jansen, K. L., Motley, R., & Hovey, J. (2010). Anxiety, depression and students' religiosity. Mental Health, Religion and Culture, 13, 267–271. http://dx.doi.org/10.1080/ 13674670903352837. Kahler, C. W., Strong, D. R., & Read, J. P. (2005). Toward efficient and comprehensive measurement of the alcohol problems continuum in college students: The brief young adult alcohol consequences questionnaire. Alcoholism, Clinical and Experimental Research, 29, 1180–1189. http://dx.doi.org/10.1097/01.ALC.0000171940.95813.A5. Lo, Y., Mendell, N., & Rubin, D. (2001). Testing the number of components in a normal mixture. Biometrika, 88, 767–778. http://dx.doi.org/10.1093/biomet/88.3.767. Maltby, J., & Day, L. (2000). Depressive symptoms and religious orientation: Examining the relationship between religiosity and depression within the context of other correlates of depression. Personality and Individual Differences, 28, 383–393. http://dx. doi.org/10.1016/S0191-8869(99)00108-7. Marsh, H. W., Lüdtke, O., Trautwein, U., & Morin, A. J. (2009). Classical latent profile analysis of academic self-concept dimensions: Synergy of person- and variable-centered approaches to theoretical models of self-concept. Structural Equation Modeling, 16, 191–225. McCarthy, D. E., Ebssa, L., Witkiewitz, K., & Shiffman, S. (2015). Paths to tobacco abstinence: A repeated-measures latent class analysis. Journal of Consulting and Clinical Psychologyhttp://dx.doi.org/10.1037/ccp0000017http://dx.doi.org/10.1037/ ccp0000017 McCullough, M. E., & Willoughby, B. L. (2009). Religion, self-regulation, and self-control: Associations, explanations, and implications. Psychological Bulletin, 135, 69–93. http://dx.doi.org/10.1037/a0014213. McCullough, M. E., Enders, C. K., Brion, S. L., & Jain, A. R. (2005). The varieties of religious development in adulthood: A longitudinal investigation of religion and rational choice. Journal of Personality and Social Psychology, 89, 78–89. http://dx.doi.org/10. 1037/0022-3514.89.1.78. Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State worry questionnaire. Behaviour Research and Therapy, 28, 487–495. http://dx.doi.org/10.1016/0005-7967(90)90135-6. Molina, S., & Borkovec, T. D. (1994). The Penn State Worry Questionnaire: Psychometric properties and associated characteristics. In G. C. L. Davey, & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment, and treatment (pp. 265–283). Oxford, England: John Wiley & Sons. Moreira-Almeida, A., Lotufo Neto, F., & Koenig, H. G. (2006). Religiousness and mental health: A review. Revista Brasileira de Psiquiatria, 28, 242–250. http://dx.doi.org/10. 1590/S1516-44462006005000006. Nagin, D. S. (2005). Group-based modeling of development. Cambridge, MA: Harvard University Press. Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14, 535–569. http:// dx.doi.org/10.1080/10705510701575396. Pargament, K. I. (1997). The psychology of religion and coping: Theory, research, practice. New York: Guilford Press. Pargament, K. I., Smith, B. W., Koenig, H. G., & Perez, L. (1998). Patterns of positive and negative religious coping with major life stressors. Journal for the Scientific Study of Religion, 37, 710–724. http://dx.doi.org/10.2307/1388152. Pargament, K., Feuille, M., & Burdzy, D. (2011). The Brief RCOPE: Current psychometric status of a short measure of religious coping. Religion, 2, 51–76. http://dx.doi.org/ 10.3390/rel2010051. Park, C. L. (2005). Religion as a meaning‐making framework in coping with life stress. Journal of Social Issues, 61, 707–729. http://dx.doi.org/10.1111/j.1540-4560.2005. 00428.x.

169

Park, C. L. (2007). Religiousness/spirituality and health: A meaning systems perspective. Journal of Behavioral Medicine, 30, 319–328. http://dx.doi.org/10.1007/s10865-0079111-x. Park, N. S., Lee, B. S., Sun, F., Klemmack, D. L., Roff, L. L., & Koenig, H. G. (2013). Typologies of religiousness/spirituality: Implications for health and well-being. Journal of Religion and Health, 52, 828–839. http://dx.doi.org/10.1007/s10943-011-9520-6. Pearce, L. D., Foster, E. M., & Hardie, J. H. (2013). A person‐centered examination of adolescent religiosity using latent class analysis. Journal for the Scientific Study of Religion, 52, 57–79. http://dx.doi.org/10.1111/jssr.12001. Read, J. P., Kahler, C. W., Strong, D., & Colder, C. R. (2006). Development and preliminary validation of the Young Adult Alcohol Consequences Questionnaire. Journal of Studies on Alcohol and Drugs, 67, 169–177. http://dx.doi.org/10.15288/jsa.2006.67.169. Ryan, R. M., Rigby, S., & King, K. (1993). Two types of religious internalization and their relations to religious orientations and mental health. Journal of Personality and Social Psychology, 65, 586–596. http://dx.doi.org/10.1037//0022-3514.65.3.586. Ryff, C. (1989). Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57, 1069–1081. Sakamoto, Y., Ishiguro, M., & Kitagawa, G. (1986). Akaike information criterion statistics. Dordrecht, The Netherlands: D. Reidel. Salas-Wright, C. P., Vaughn, M. G., Hodge, D. R., & Perron, B. E. (2012). Religiosity profiles of American youth in relation to substance use, violence, and delinquency. Journal of Youth and Adolescence, 41, 1560–1575. http://dx.doi.org/10.1007/s10964-012-9761z. Salas-Wright, C. P., Vaughn, M. G., & Maynard, B. R. (2014). Profiles of religiosity and their association with risk behavior among emerging adults in the United States. Emerging Adulthood, 1–18. http://dx.doi.org/10.1177/2167696814539327. Scheier, M. F., Wrosch, C., Baum, A., Cohen, S., Martire, L. M., Matthews, K. A., ... Zdaniuk, B. (2006). The life engagement test: Assessing purpose in life. Journal of Behavioral Medicine, 29, 291–298. http://dx.doi.org/10.1007/s10865-005-9044-1. Schwarz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 6, 461–464. http://dx.doi.org/10.1214/aos/1176344136. Simons, J. S., & Gaher, R. M. (2005). The distress tolerance scale: Development and validation of a self-report measure. Motivation and Emotion, 29, 83–102. http://dx.doi.org/ 10.1007/s11031-005-7955-3. Steger, M. F., Kashdan, T. B., Sullivan, B. A., & Lorentz, D. (2008). Understanding the search for meaning in life: Personality, cognitive style, and the dynamic between seeking and experiencing meaning. Journal of Personality, 76, 199–228. http://dx.doi.org/10. 1111/j.1467-6494.2007.00484.x. Steger, M. F., Pickering, N. K., Adams, E., Burnett, J., Shin, J. Y., Dik, B. J., & Stauner, N. (2010). The quest for meaning: Religious affiliation differences in the correlates of religious quest and search for meaning in life. Psychology of Religion and Spirituality, 2, 206–226. http://dx.doi.org/10.1037/a0019122. Stewart, C. (2001). The influence of spirituality on substance use of college students. Journal of Drug Education, 31, 343–351. http://dx.doi.org/10.2190/HEPQ-CR08MGYF-YYLW. Stoppa, T. M., & Lefkowitz, E. S. (2010). Longitudinal changes in religiosity among emerging adult college students. Journal of Research on Adolescence, 20, 23–38. http://dx.doi. org/10.1111/j.1532-7795.2009.00630.x\. Substance Abuse and Mental Health Services Administration (SAHMSA) (2009). Office of Applied Studies. Results from the 2008 National Survey on Drug Use and Health: National Findings. Rockville, MD: Substance Abuse and Mental Health Services Administration. Substance Abuse and Mental Health Services Administration, Office of Applied Studies (2011). Results from the 2010 National Survey on Drug Use and Health: National Findings. Rockville, MD: Substance Abuse and Mental Health Services Administration. Van Dam, N. T., & Earleywine, M. (2011). Validation of the center for epidemiologic studies depression scale—revised (CESD-R): Pragmatic depression assessment in the general population. Psychiatry Research, 186, 128–132. http://dx.doi.org/10.1016/j. psychres.2010.08.018. Vuong, Q. (1989). Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica, 57, 307–333. http://dx.doi.org/10.2307/19125. Wood, R. J., & Hebert, E. (2005). The relationship between spiritual meaning and purpose and drug and alcohol use among college students. American Journal of Health Studies, 20, 72–79.