Helping Protestant Undergraduates in the United States Manage Their Religious Doubt: The Predictive Role of Facet and Domain Traits in the Five Factor Model

Abstract

In a recent Barna poll, 60% of practicing Christians indicated rarely or never doubting religious beliefs. This puzzling statistic evokes questions. How did this approach to uncertainty benefit the respondents? Were suppression and concealment in their coping strategy? Would an alternative management approach, one derived from a compendium of personality traits in a four-stage doubting process, be of assistance? In a review of religion-personality literature, several gaps were uncovered. Gender effects with these constructs were often unattended; authors utilizing the Five-Factor Model (FFM) frequently didn’t include facet traits; positive and negative sides of religious doubt were usually not discussed together; multiple stages in the doubting process were seldom considered; and researchers rarely employed a multidimensional model of quest religious orientation (QRO) as a measure of doubt phenomena. Consequently, a sample of Protestant Christian undergraduates (n = 335) was recruited and administered the FFM using facet and domain traits along with unidimensional and multidimensional instruments of QRO. Findings from hierarchical regression analyses of constructs representing four stages in the doubting process revealed 11 facet and the five domain traits emerging as predictors. Implications and applications from the predictive associations forming a doubt management strategy were discussed.

1. Introduction

Over a hundred years ago, Snowden (1910), a Presbyterian theologian, wrote, religious doubt is “uncertain belief attending all our knowledge. [It is a] great destroyer of error, a step toward clearer light, [and] should not be dealt with by condemnation and suppression” (pp. 151–52, 155). His quotes underline doubt being an integral element in the belief systems of religious people with the potential to improve their doctrinal clarity and lead them toward a more robust faith (Wallace 2021).

Yet, over the past century, Snowden’s constructive perspective on doubt in faith development has seemingly waned. Fewer American Christians are doubting. A recent poll published by Barna (2023) reported that 60% of practicing Christians rarely or never doubt religious convictions; only 20% percent occasionally or frequently doubt.

These puzzling findings in Barna’s poll evoke questions. Who are practicing Christians? How does their approach to religious uncertainty benefit them? Were suppression and concealment main components in the management strategy? Are these statistics related to religious doubt being a controversial and stigmatized topic among American Protestant Christians? Last, do the Barna results underscore the need for a different doubt management approach; possibly, one derived from a compendium of personality traits in a four-stage doubting process?

The following discussion pursues answers to the previously mentioned queries. Relevant information about the American Christian church’s continuing uncomfortableness with religious doubt and research on religion-personality are explicated. A rationale is defined for hierarchical regression analyses of doubt (i.e., quest religious orientation) and personality (i.e., facet and domain traits in the Five Factor Model) checking for gender and suppression effects with a sample of Protestant Christian undergraduates. Further, applications derived from the predictive associations, form a practical management strategy useful for American Protestant Christian college students journeying through four stages of the doubting process and for their mentors (i.e., pastors, counselors) who travel with them.

1.1. The Persistent Disquiet with Religious Doubt in the American Church

One of the startling features in the previously mentioned Barna (2023) poll is that practicing Christians matched themselves to frequency adverbs underscoring that doubt almost never or never emerged in their religious belief system. According to Barna’s (2018) definition, this group of people self-identified as Christian, indicated religion was important to them, and attended at least one religious service a month. Further, Mattingly (2021) described them as religiously sincere, engaged, and genuine indicating Christianity has had a positive and meaningful effect on their lifestyle.

The nagging question is how religious uncertainty became ‘almost non-existent to non-existent’ for practicing Christians. Perhaps the American Christian church’s persistent disquiet with religious doubt is a starting point for explaining the outcome. Over the past few decades, a variety of authors have mentioned and discussed numerous negative views on religious doubt along with suggestions for how perceptions perpetuate. Puffer et al. (2008) mentioned a controversial cloud surrounding religious doubt. Their conclusions derive from the kinds of depiction offered by authors. For instance, Lucado (1989) labeled doubt a nosey, obnoxious pest; Buchanan (2000) deemed it a cancer; and Darmani (2002) considered it a demonic weapon. Strobel (2000) reported doubt as a traumatic and shameful phenomenon; Moreland and Issler (2008) shared it was a subject lacking safety. Almost ten years later, Puffer (2017), mentioning more negative portrayals of religious doubt, dubbed it an uncomfortable and misunderstood subject. For example, Patton (2010) regarded it as being scary while Kinnaman (2011) wrote about doubting being considered sinful.

In recent literature (≥2020), negativity has persisted. Religious doubt is still deemed uncomfortable (Piper 2020) and sinful (Dickinson 2022); it is referred to as being bad, frightening, and an enemy of the faith (P. Long 2023). Some suggested religious doubt results from character flaws (Wallace 2021) leaving Christians to frown upon it, condemn it (Piper 2020), and stigmatize it (Dickinson 2022). Further, Christians who doubt are probably doing so ‘behind closed doors’ (Moldovan 2022) and must socially present themselves with a cheerful veneer (P. Long 2023).

Puffer (2013) proffered three mechanisms aiding the perpetuation of anxiety with and awkwardness toward religious doubt. He suggested influential interpretative approaches and actions contributing to the re-indoctrination of doubt negativity. First were lopsided characterizations. These are extreme negative representations related to doubt (e.g., demonic, bad, frightening) which omit possible positive and beneficial information (Darmani 2002; P. Long 2023). Second were inappropriate transpositions. These construe religious doubt to be something that it isn’t (e.g., unbelief, contrarism, skepticism) (Buchanan 2000; Wallace 2021). Third were subtle and blatant suppressive tactics. This behavioral mechanism entails church leaders or adult caregivers squelching contrary ideas or refusing to discuss doctrinal critiques or complaints which in turn can encourage doubters to conceal their pressing questions (Dark 2009; Moldovan 2022; Sentell 2021; Snowden 1910).

Other mechanisms have emerged. Ponce (2022) underscored internet misinformation bedeviling Christians’ personal knowledge base. In religious people’s struggle to ascertain truthful knowledge, they often encounter and must sort through a vast amount of false information on the world wide web such as fake news, rumors, hearsay, tampered videos, conspiracy ideologies, etc. (Davis 2021). If unchecked, this easily obtained material can render Christians misdirected, overwhelmed, and anxious which in turn can complicate the process of forming accurate and balanced opinions about religious doubting.

Further, Reynoso (2020) pointed to deconstructionism, a movement growing in popularity among Christians (Gryboski 2023; Guhu 2023). In general, this mechanism refers to religious people jettisoning church traditions and purging cultural-political baggage attached to their Christianity (Loyd 2021; Moore 2022). The assumed goal is a pure, unadulterated biblical faith. Snowden (1910) referred to religious doubt as a deconstructionistic force destroying error. Yet, he also considered it constructionistic to pave a way for spiritual growth. The contemporary deconstructionistic movement doesn’t always encourage a reconstruction of faith components after a jettison/purge and has been unfortunately aligned with religious doubt (Szterszky 2022).

1.2. Movement Beyond Dichotomy and Toward Rapprochement

In both religious and psychological literature, positivistic voices about religious doubt are evident. Snowden (1910) considered doubt a lasting component in life that grows as people develop. Allport (1950) maintained the “workshop of doubt” shapes the religious sentiment (p. 73). Halfaer (1972) mentioned doubt is intricately woven into the human condition. Tillich (1957) as did Fowler (1996) regarded doubt as a constituent of faith and a vital aspect in the faith developmental process.

Recent articles have also depicted doubt as normative. Authors have regarded it as a common phenomenon (Wallace 2021) that is unavoidable in a fallen world (Ponce 2022) and a vital piece of religious people’s faith structure (P. Long 2023). Christians need to anticipate and positively embrace their religious uncertainty (Wallace 2021).

Yet, discussions on religious doubt must move beyond dichotomy and toward rapprochement. Dichotomous discussions are problematic by being imbalanced (Krause and Ellison 2009). The authors of these articles only prepare readers to encounter one side of doubt. For instance, how will Christians be ready for a good doubt experience that turns into a knotty and awkward journey such as a dark night of the soul (see Appendix A)?

Instead of painting doubt as an “either-or” phenomenon, it is time to conceptualize it as an “and” phenomenon. For example, religious doubt is a complex process providing positive insights that can empower Christians to cope with an imperfect world and it can be an experience that becomes problematic, uncomfortable, and stressful needing skillful coping responses. What is needed in the religion-personality literature is a study embracing reproachment and investigating both sides of religious doubt together.

1.3. Overlooked Features of Religious Doubt

Often overlooked are the conversations about religious doubt underlining two kinds of doubt. Piper (2020) divided the cognitive and emotional phenomenon as believing and unbelieving doubt; Osborne (2011) referred to the two dimensions as honest and dishonest doubt. Believing/honest doubt tends to seek a deconstruction of religious beliefs followed by a reconstruction; the disbelieving/dishonest usually pursues deconstruction. The former embodies humility and genuine inquisitiveness; the latter tends to be dismissive of reality and borders on becoming moribund and dull (P. Long 2023).

Additionally, both dimensions of religious doubt tend to question, self-dispute, waver, and swim in uncertainty. But religious people with honest/believing doubt remain anchored to God and trust in Him; their allegiance continues as steely and singular. Persons with dishonest/disbelieving doubt are more tethered to human wisdom and quasi-trusting God; their loyalty to God is chance-dependent and half-hearted (Piper 2020). Osborne (2011) argued that the message of the Apostle James exposed and condemned this kind of dividedness or double mindedness.

Furthermore, investigators of religious doubt tend to center on one feature in the overall doubt process, the cognitive/emotional state, overlooking other facets. Krause and Ellison (2009) expanded the view on the doubt process and delineated its complexity with four inter-related stages. These include precipitant[s], the cognitive/emotional experience, coping response[s], and possible outcome[s]. They argued religious doubt is triggered by something (e.g., new religious information, social conflicts with religious friends, etc.); it includes a state of uncertainty; and there are after-effects—both immediate (e.g., coping efforts to resolve or ameliorate doubtfulness) and long-term (e.g., outcomes such as health difficulties).

Puffer (2017) implemented the model by Krause and Ellison (2009) with a few different nuances. He operationalized three of their four stages using doubt phenomena from the Multidimensional Quest Orientation Scale (MQOS; Beck and Jessup 2004) and sought to discover personality traits that predicted each stage utilizing the California Psychological Inventory (CPI; Gough 1987). However, Puffer (2017) only found three social personality traits (i.e., intolerance, dependence, and non-conformity) predicting doubt phenomena, only two of the four stages (i.e., the cognitive/emotional state [stage 2] and coping response [stage 3]) had obtained predictors, and the range of explained variance was 2–8%.

What is also needed in the literature of religion-personality is an expansion of the study by Puffer (2017). More specifically, three important additions are needed. First is the operationalization of all four stages in the model by Krause and Ellison (2009). Second is the employment of a different personality model (i.e., the Five Factor Model) which has potential to improve both the number of predictors and the amount of explained variance. Third is the development of a more extensive doubt management strategy based on the personality traits obtained at each stage.

1.4. Research on Psychology of Religion and Personality

The relationship between the psychology of religion and personality psychology has been longstanding and increasingly constructive over time (Aten et al. 2013; McCullough et al. 2003; Schnitker and Emmons 2021). One of the hallmarks of the former psychological specialty has offered scientists a wide and thorough understanding of diverse religious phenomena. The latter has provided a comprehensive picture of the intricate details of human personality traits. Over the last three decades, hundreds of researchers have combined the two hallmarks to ascertain how people, religious and non-religious, see, interpret, and react to religious/spiritual constructs (Brooks 2023; Entringer et al. 2021).

In the religion–personality literature, authors have operationalized religiosity in a variety of ways. Two notable approaches include classic or traditional specific constructs and author-developed generalized concepts. For the latter, Paunomen (1998) invited his sample to indicate via one question whether they were religious in a behavioral sense (i.e., attendance at a religious gathering). Entringer et al. (2021) asked participants one religious identity question (i.e., perception of self as very religious). Other tactics include a composite of questions about religious upbringing (McCullough et al. 2003), faith in God (Löckenhoff et al. 2009); emotional religion (Saroglou and Muñoz-García 2008), and faith as a sense of security (Unterrainer et al. 2014).

For the classic or traditional approach, religious motivational variables such as intrinsic (IRO), extrinsic (ERO), and quest (QRO) religious orientation, were frequently utilized. Lace et al. (2020) operationalized IRO via three items in the Duke University Religion Index (DUREL; Koenig et al. 1997). Most other researchers have utilized the Allportian (i.e., IRO and ERO) and Batsonian (QRO) measures or forms of them (Barrett and Roesch 2009; Henningsgaard and Arnau 2008; Kosek 1999; Piedmont 2005; Robbins et al. 2010; Taylor and MacDonald 1999). Puffer (2013, 2017) employed a multidimensional quest religiosity scale.

Several investigations operationalized personality with one of various measures of the global or domain traits of the Five Factor model (FFM). Among the studies using the generalized, author developed religious constructs, Agreeableness and Conscientiousness consistently emerged in a positive albeit small (0.10–0.30) relationship (McCullough et al. 2003; Miller and Worthington 2012; Saroglou and Muñoz-García 2008; Löckenhoff et al. 2009). Religious individuals tend to be agreeable (i.e., compliant, cooperative) and conscientious (i.e., self-restrained, disciplined) (Howard and Howard 2010).

Paunomen (1998) found religiosity positively related to Neuroticism while Löckenhoff et al. (2009) reported it negatively correlated with religiousness. Further, there were mixed findings with Openness; some authors found this domain positively associated (Löckenhoff et al. 2009; McCullough et al. 2003; Unterrainer et al. 2014) and non-significant with religiosity (Paunomen 1998; Saroglou and Muñoz-García 2008). Extraversion was consistently less relevant to understanding individual differences in religion (Löckenhoff et al. 2009; McCullough et al. 2003; Saroglou and Muñoz-García 2008; Unterrainer et al. 2014).

Concerning FFM outcomes with religious motivational concepts, Agreeableness and Conscientiousness were positively related to IRO (0.10–0.40) in a consistent manner (Henningsgaard and Arnau 2008; Kosek 1999; Lace et al. 2020; Miller and Worthington 2012; Piedmont 2005; Taylor and MacDonald 1999). ERO was positively related to Neuroticism (0.10–0.20) (Barrett and Roesch 2009; Piedmont 2005; Taylor and MacDonald 1999).

As for associations between QRO and FFM domains, authors found mixed results. Agreeableness (+0.10–0.30; Barrett and Roesch 2009; Kosek 1999), Neuroticism (+0.10–0.20; Barrett and Roesch 2009; Henningsgaard and Arnau 2008), and Openness (+0.20; Barrett and Roesch 2009) were positively related. But Conscientiousness was both positively and negatively correlated with quest religiosity (−0.20 (Henningsgaard and Arnau 2008) and +0.30 (Kosek 1999)).

Using the California Psychological Inventory (Gough 1987), Puffer (2013) reported that the Quest Scale (Batson 1976) was positively associated with empathy and social sensitivity while negatively related to independence and conformity. He also noted that exploration (i.e., pursuit of religious knowledge), a subscale of the Multidimensional Quest Orientation Scale (MQOS; Beck and Jessup 2004), was positively related to empathy, independence, conformity, dominance, capacity for status, and sociability. Furthermore, Puffer (2017) reported that complexity (i.e., preference for non-simple religious ideas) of the MQOS was predicted by non-conformity; existential motives (i.e., finding purpose and meaning) of the MQOS was predicted by dependence and intolerance; and change (i.e., openness to evaluate and alter religious views) of the MQOS was predicted by non-conformity (Beck and Jessup 2004).

Religion-personality findings with the facets of the FFM have been limited (Entringer et al. 2021). Paunomen (1998) added 21 facets from the Jackson Personality Inventory (Jackson 1984) to the five domains of the FFM. He found religiosity was positively related to several Neuroticism facets. Saroglou and Muñoz-García (2008) reported two Neuroticism facets (i.e., n1 [anxiety] and n6 [vulnerable]) were positively associated with their religion construct; three Conscientiousness facets (i.e., c1 [competence], c4 [achievement] and c5 [self-discipline]) were positively related; all Agreeableness facets positively correlated; and the Openness facets (i.e., o4 [actions]; o5 [ideas]; and o6 [values]) were either negatively related or not significant. They concluded that more comprehensive information is found with facets along with slight increases in explained variance were obtained.

The facet correlations from Löckenhoff et al. (2009) uniquely contrast the previous results. They indicated five of the six Neuroticism facets (i.e., n1 [anxiety], n2 [angry], n3 [depression], n5 [impulsive], and n6 [vulnerable]) were negatively correlated with religiousness. Four Extraversion facets (i.e., e1 [warmth], e2 [gregariousness], e3 [assertive], and e6 [positive emotions]) were positively related and all Conscientiousness facets were positively associated. Only four of the Agreeableness facets (i.e., a1 [trust], a3 [altruism], a4 [compliance], and a6 [tendermindedness]) were positively related while only two Openness facets (i.e., o5 [ideas] and o2 [aesthetics]) were positively associated with their religiosity construct.

Entringer et al. (2021) studied religion-personality in 96 countries (n > 2 million) with 10 facets from the FFM. Using a Likert scale of 1 (strongly disagree) to 5 (strongly agree) with the one religious question (i.e., I see myself as very religious) to obtain participants’ religiosity and country-level religiosity, they reported Nigeria had a 3.71 result while Sweden had 1.67. Associations between religiosity and the 10 FFM facets were small, 0.00–0.26. In general, they found the religion-personality correlations overwhelmingly positive for the Agreeableness (a3 [altruism] at 100%, and a4 [compliance] at 76%) and Conscientiousness (c2 [order] at 73%, and c5 [self-discipline] at 94%). There were mixed results for Extraversion (e3 [assertiveness] at 85% negative and e4 [activity] at 98% positive), Openness (o2 [aesthetics] at 76% positive and o5 [ideas] at 88% negative), and Neuroticism (n1 [anxiety] at 87% positive and n3 [depression] at 52% negative). Hence, Entringer et al. concluded that FFM facets do explain more variance in religiosity but only if the sample is predominately religious. For example, a3 (altruism), a4 (compliance) and c5 (self-discipline) positively predicted religiosity in religious cultures while o5 (ideas) negatively projects religiosity in religious contexts.

Further, among the previously mentioned studies, there are two other noticeable issues—insensitivity to gender differences and a reluctance to employ a multidimensional measure of quest religious orientation (QRO). Gender differences in religion and personality have been clearly established. For instance, Rassoulian et al. (2021) concluded religion offers more inner strength for women than men and Holman and Podrazik (2018) wrote that for 50 years, women as compared to men, express higher levels of most forms of religiosity. Lace et al. (2020) reported that female participants reported more significant differences among the FFM than males. Yet less than half of the previously mentioned articles attended to gender differences (Entringer et al. 2021; Lace et al. 2020; Löckenhoff et al. 2009; Paunomen 1998; Robbins et al. 2010; Saroglou and Muñoz-García 2008; Taylor and MacDonald 1999).

For the issue with QRO, there has been criticism of Batson’s (1976) unidimensional measure. Some have argued the twelve items are insufficient (i.e., decreased discriminant validity; problematic predictive validity) to capture the three main dimensions comprising quest (Batson et al. 1993; Beck et al. 2001; Crosby 2013; Leak 2011). When comparing the previously mentioned investigations, less than half examined quest religiosity (Barrett and Roesch 2009; Henningsgaard and Arnau 2008; Kosek 1999; Puffer 2013, 2017; Robbins et al. 2010) and among those, only two studies (Puffer 2013, 2017) employed a multidimensional model of QRO.

Therefore, the present study sought to respond to the previously mentioned gaps by extending the study by Puffer (2017). The intention was to assess the predictive relationship between religious doubt and personality over four stages of the doubt process while attending to gender and suppression effects. Towards that end, this investigation would include the FFM facets along with the domains, discuss both sides of doubt—positive and negative, and employ both unidimensional and multidimensional models of quest religiosity. Consequently, two research questions were pursued:

• For Protestant Christian undergraduates in the United States, what combination of personality variables (i.e., facet and domain traits in the FFM) predicts religious doubt variables (i.e., both unidimensional and multidimensional models of quest religious orientation) in the four-stage doubt process (i.e., precipitant, cognitive state, coping, and outcome) by Krause and Ellison (2009) when checking for gender and suppression effects?
• What doubt management ideas emerge from obtained predictors at each of the four stages from the hierarchical regression analyses that would benefit American Protestant Christian college students engaging both the positive or negative sides of doubt and that also profit the mentors (i.e., pastoral counselors, pastors, and mental health therapists) who support the undergraduates in their religious pilgrimage.

In the present investigation, the following outcomes were expected.

• From a theoretical vantage, using the expressive (i.e., authenticity) and sociocultural (i.e., social sway) perspectives as to why personality traits anticipate religious behavior, it is expected Openness and Neuroticism will be positive predictors of quest religion (Gebauer et al. 2020). Doubt positivity is not normative in the American Christian church culture (Dickinson 2022). Further, Agreeableness, Conscientiousness, and Extraversion will be negative predictors; the process of religious doubt doesn’t always comport with trust, compliance, traditionalism, and positive emotions (McCrae and Costa 2010).
• Yet, from an empirical vantage, studies in religion-personality literature prompt a few nuances. These older findings lead to the anticipation that facet and global traits of Openness, Neuroticism, and Agreeableness will be positive predictors for quest religiosity (Barrett and Roesch 2009; Henningsgaard and Arnau 2008; Kosek 1999) while the facet and domain traits of Conscientious will be negative predictors for quest religiosity (Henningsgaard and Arnau 2008). Further, the facet and domain traits of Extraversion will be unrelated to quest religiosity (Barrett and Roesch 2009).

2. Methods and Materials

2.1. Participants

Recruitment for participants followed a few specific selection criteria. These preferences included persons in attendance at a university, aged 18–22, self-described as religious (i.e., a high importance for religion), identified from a diverse ethnic background, and enrolled in a psychology course that offers extra credit for research participation or has a battery of personality and religious instruments incorporated in the curriculum. Further, a power analysis indicated at least 150 participants were needed to detect a significant increase in R² of 0.15, with an alpha level of 0.05 and a power of 0.80, for a multiple (hierarchical) regression model with about 20 predictors. Therefore, a private Protestant Christian institution advertising the importance of diversity was viewed as an appropriate place to recruit persons possessing most if not all previously mentioned conditions.

Participants were college students (n = 335) enrolled in one of two classes, General Psychology or Theories of Personality, at a private Protestant Christian university located in the Midwest of the United States. The students comprised 231 females (69%) and 104 males (31%) with a mean age of 19.6 years. Regarding religious landscape, 82% indicated religion was very important, 14% indicated religion is somewhat important, 2.0% indicated religion is not important, and 2.0% indicated not sure. Religious association included 59% identified as Protestant, 39% as other or undesignated, and 2% as Catholic. Pertaining to ethnic affiliations, 90% were European American, 4% Hispanic American, 3% other or undesignated, 2% Asian American, and 1% African American. Further, the undergraduates’ affirmations for major entailed: Psychology—27%, Education—9%, Addiction Counseling—8%, Business—8%, Pre-Declared—7%, Christian Ministries—6%, Nursing—6%, Social Work—4%, Criminal Justice—4%, Pre-Art Therapy—4%, Biology—4%, Physical Education—2%, Art + Design—1.6%, Communication—1.6%, Exercise Science—1.3%, Health Science—1.1%, Music—0.9%, and the remaining 4.5% included three other majors (0.2–0.7%). Hence, these college students were appropriate candidates for this study, meeting most of the selection criteria and making a ‘purposive sampling’ appropriate for the aims in this investigation (Tie et al. 2019).

2.2. Procedures

The duration of the present study lasted six years; the protocol of the Institutional Review Board of the university (approval IRB ID Number—1890.23) was honored. Data collection took place in two psychology courses in both fall (i.e., September–December) and spring semesters (i.e., January–April). The first course, General Psychology (PSY 150), is a general educational requirement for most majors and offered in the fall and spring; eight sections (n = 204) were recruited over the six years. Its curriculum includes extra credit rewards for participation in psychological research projects. The second course, Personality Psychology (PSY 370), is an elective course for several majors and offered only in the spring; eight sections (n = 131) were involved. On the first day of class, a Personality Assessment Battery, the first assignment which includes the Five Factor Model (FFM) measure and consent form, was explained. Religious measures were offered as an extra credit opportunity later in the semester.

Although the FFM and religious measures are assignments or an extra credit opportunity, the undergraduates were carefully informed in both PSY 150 and 370 about the project, the measures, the consent form, possible risks, the proviso of being free to withdraw their scores from the study without penalty, incentives (i.e., extra credits or participation points), and completion expectations (i.e., one week). All participants heard these details of the study verbally and were also directed to re-read through the consent form again and submit it with their signature along with the completed measures.

All participants’ measures were scored by trained research assistants. They also entered the obtained scores in an Excel (Version 2501) file and then transferred them to a SPSS (Version 20) file. Interpretative information for participants’ results was distributed at scheduled times according to the professors’ preference in the fall or spring semesters.

2.3. Measures

To assess the religious doubt-personality relationship, three measures were employed; the doubt instruments were derived from two different models of quest religious orientation (QRO)—a unidimensional and a multidimensional measure and the personality instrument was the Five Factor Model (FFM). The former were selected because QRO is unlike other religious motivation constructs; it examines religious motives/functioning related to uncovering new religious knowledge, grappling with painful circumstances due to an imperfect world, questioning/challenging doctrinal tenets, and tweaking/altering a belief system. The latter was selected because it provides a concise and comprehensive picture of human personality. The following discussion details pertinent information for the threesome.

2.3.1. Religion Variables

Quest Scale (QS). This is the one-dimensional measure of QRO created by Batson (1976). The QS comprises 12 items using a nine-point Likert response format; the questions are subdivided to cover his three components of quest as religion. The first facet, responsiveness, refers to a willingness to face existential challenges and complex questions in life. For instance, one of the four items in this facet states: “My life experiences have led me to rethink my religious convictions” (Batson et al. 1993, p. 170). The second facet, doubt positivism, indicates a constructive and optimistic attitude toward uncertainty. For example, one of the four questions in this facet states: “I value my religious doubts and uncertainties” (Batson et al. 1993, p. 170). The third facet, openness, refers to an on-going search for new religious knowledge that can be accompanied with belief alterations. For instance, one of the four item states: “As I grow and change, I expect my religion also to grow and change” (Batson et al. 1993, p. 170). As for reliability data, alpha coefficients of 0.80 and 0.82 for the composite score have been reported (Batson 1976; Mikani et al. 2022).

Multidimensional Quest Orientation Scale (MQOS). The MQOS is the multi-dimensional measure of QRO created by Beck and Jessup (2004). The authors questioned the predictive validity of Batson’s composite score and developed a 62-item instrument using a seven-point Likert response format. Nine subscales were generated including existential motives (EXM), complexity (COM), tentativeness (TEN), exploration (EXP), change (CHAN), religious angst (RA), ecumenism (EC), universality (UNV), and moral interpretation (MI). As for reliability data, alpha coefficients across the nine subscales range from 0.68–0.90 (Beck and Jessup 2004) and 0.69–0.89 (Crosby 2013). Regarding validity specifications, the principal axis factor analysis by Crosby (2013) corroborated the nine subscales of the MQOS.

In the present study, six of the nine MQOS subscales were utilized and applied to the four stages in the doubt process espoused by Krause and Ellison (2009). Following Puffer’s (2017) ideas for operationalization, Stage 1 (Precipitant[s]) points to the starting point of the doubt process (Allport 1950; Krause and Ellison 2009). EXP, exploration, was applied. This subscale assesses people’s effort to investigate and analyze religious and spiritual information (Beck and Jessup 2004). Further, EXP embodies a part of the Batson’s openness facet in the QRO—searching for more light and absorbing new knowledge (Batson et al. 1993).

Stage 2 (Cognitive/Emotional State) entails the well-known state of uncertainty where religious people question, hesitate, self-dispute, and waver. EXM, existential motives, and COM, complexity, were applied here. The former measures motivational level for pursuing answers to existential questions and the latter evaluates the preference level for complex (i.e., not simple) views and answers to religious issues and inquiries (Beck and Jessup 2004). Further, the two criterion measures represent Batson’s (1976) responsiveness, a facet of the QRO, to attend and courageously face challenging complex questions and even tragic circumstances in everyday life situations.

Stage 3 (Coping Response[s]) involves the first of the after-effects, that is a response to doubtfulness and has a short-term timeframe. This refers to efforts that swiftly move to resolve or eliminate doubtfulness. CHAN, change, is applied here because it is an evaluation of a person’s willingness to inspect his or hers religious or spiritual perspectives and possibly alter them (Beck and Jessup 2004). Further, this instrument embodies the other part of Batson’s facet of openness in the QRO—possible belief modification (Batson et al. 1993).

Stage 4 (Outcome[s]) points to the second of the after-effects but this one, timewise, is more long-term in nature. Krause and Ellison (2009) studied health difficulties (e.g., distress) related to doubt. Two MQOS subscales, TEN, tentativeness, and RA, religious angsts, were applied to illustrate both positive and negative outcomes in the doubt process. TEN is an assessment of how much a person views doubt as a beneficial and positive experience (Beck and Jessup 2004); it embodies Batson’s facet of doubt positivism, the degree of comfortability with doubt. RA measures religious people’s experience with the possible negative side of doubt, its uncomfortable and frightening dimensions such as doubtful experiences, negative emotions (e.g., anxiety, despair), and a sense of isolation in their faith pilgrimage (Beck and Jessup 2004; P. Long 2023; Piper 2020).

2.3.2. Personality Variables

The NEO Personality Inventory—Revised (NEO PI-R: McCrae and Costa 2010) was utilized to measure the Five Factor Model (FFM). This instrument comprises 240 items with a five-point Likert response format. It covers five major domains of human personality with six corresponding facets defining each domain. The dimensions of the NEO include Neuroticism (i.e., the emotional style indicating affect stability), Extraversion (i.e., interpersonal style noting sociability), Openness (i.e., the experiential style pointing to the level of curiosity), Agreeableness (i.e., the attitudinal style—the level of sociocentrism), and Conscientiousness (i.e., the motivational style—the level of intrapersonal restraint) (Howard and Howard 2010; McCrae and Costa 2010). The thirty facets (e.g., n1 = anxiety; o5 = ideas, etc.) comprise more specific information about each domain offering a wide amount of relevant cognitive, affective, and behavioral insights.

Reported psychometrics for the NEO PI-R indicate that the range of alpha coefficients for the domains is 0.86–0.92. With the facets, the ranges vary some. For those with Neuroticism, the range is 0.68–0.81; for the Extraversion facets, it is 0.63–0.77; for the Openness facets, it is 0.58–0.80; for the Agreeableness facets, it is 0.56–0.79; and for the Conscientiousness facets, it is 0.75–0.90 (McCrae and Costa 2010). Validity data is abundant, indicating the NEO measures the concepts it indicates. For instance, bipolar adjective scales and adjective check lists have demonstrated the five factors (McCrae and Costa 2010). Further, Markon et al. (2005) compiled correlations of scales from many instruments (e.g., Eysenck) evaluating normal and abnormal personality features with the NEO. All scales fit within a five-factor structure like the FFM.

3. Results

3.1. Research Design

Following this study’s research questions, the aim was to develop a predictive model of religious doubt (i.e., over four stages) using one measure of personality traits, the facet and domain traits of the Five Factor Model (FFM) among religious undergraduates while checking for gender and suppression effects. Hierarchical regression procedures were selected as the analytical strategy. The targeted construct, religious doubt, led to the selection of two measures, the Quest Scale—a unidimensional measure (Batson 1976), and six subscales from the Multidimensional Quest Orientation Scale (Beck and Jessup 2004); collectively, they became the seven criterion variables. The predictor variables included gender as the designated constant in model one and the selected FFM facet and domain traits in model two. Selection of the personality variables had one criterion—partial correlation at the level of significance of p ≥ 0.05 to the doubt criterion variables. Overall, this arrangement permitted the examination of the amount of change in variance explained by the facet and domain traits of the FFM after partialling out variance related to gender (Crowson 2020).

The inclusion of a gender variable in the predictive model was based on the plethora of research in religion-personality literature. For religion, Moon et al. (2022) concluded sex differences in religion are cross culturally common and robust. For personality, Schmitt et al. (2017) concluded gender differences in most aspects of personality (e.g., FFM traits) are noticeably greater in societies with more egalitarian gender roles.

The personality variables were entered simultaneously into model two. There was no specific theoretical justification from religion or personality. Yet, entry order of the variables followed one consistent pattern per criterion variable per regression procedure: facet traits from neuroticism were first, followed by extraversion ones, openness facets were next, agreeableness ones followed, and then the conscientiousness traits. The same arrangement occurred with the domain traits.

3.2. Preliminary Analyses

3.2.1. Assumptions with Hierarchical Regression

An appraisal was conducted for how data met the assumptions for hierarchical regression. First, missing data along with outliers among the scores from the forty-two variables (i.e., seven religion and 35 personality) were examined. For the former, pairwise deletion was employed. Allison (2001) shared that pairwise deletion tends to be efficient due to the usage of more information when compared to listwise deletion; also, pairwise can generate consistent parameter estimates assuming missingness among the data is completely random (MCAR) (Tabachnick and Fidell 2019). Less than 2% of the scores for the religion and personality variables were missing. Using the data from the hierarchical analyses, sample sizes ranged from 335–340 undergraduates. Consequently, MCAR appears to be a suitable depiction of the pattern for the missing data (Tabachnick and Fidell 2019). As for outliers, the range of possible scores published by the authors of the measures guided eliminations. Three outliers (i.e., a facet with Neuroticism, ethnicity, and Religious Angst) were detected and removed. Most likely, the outliers were input errors by research assistants. Further, several age variables which did not lie within the range of 18–22 were also eliminated; seven were under 18 years of age and twelve were over 22.

Second, multicollinearity was checked with variance inflation factor (VIF) calculations in SPSS and a visual inspection of the correlation matrix. All VIF values ranged from 1.01 to 2.81, marking them under the conservative threshold of 3.0; no correlation coefficients exceeded 0.70, suggesting that the data does not show multicollinearity (Gaskin 2020).

Third, checks for normality, homoscedasticity, and linearity were conducted. Employment of P-P plots and the skewness statistic for normality and scatterplots for homoscedasticity revealed no violations. The Test of Linearity in SPSS identified three predictor variables that were not linear with criterion variables; that was c3 (dutifulness) with exploration (EXP) along with a2 (straightforwardness) and a3 (altruism) with religious angst (RA). All three predictors were appropriately correlated to the criterion variables and had normal distributions. Yet, it was determined that removing the three from the hierarchical regression analyses with the criterion variables (i.e., EXP, RA) was the best response to the linearity violations. No major effects relative to the removal were noted (Choueiry 2023; David 2019; Tabachnick and Fidell 2019).

Fourth, independence of observations was examined using the Durbin-Watson (DW) statistic. Values ranged from 1.75 to 2.14 for the religious and personality variables; these are in the middle of the 0–4 range, indicating little to no autocorrelation (CFI Team 2022).

3.2.2. Descriptive Statistics and Correlational Findings

In

Table 1. Comparison of the Means of Religious Doubt and FFM Variables Between Male and Female Undergraduates.

	Males		Females		t-Test	p-Value	ES
	Means	SD	Means	SD
QS	55.54	16.21	57.40	15.38	−0.982	0.328	0.12
EXM	19.67	6.23	19.02	5.44	0.908	0.365	0.12
COM	34.52	6.70	34.45	6.99	0.090	0.929	0.01
TEN	47.38	9.00	48.21	8.38	−0.793	0.429	0.10
EXP	25.17	6.81	24.31	7.22	1.047	0.296	0.12
CHAN	25.93	10.97	24.80	10.47	0.885	0.377	0.11
RA	22.50	8.75	23.82	8.57	−1.279	0.203	0.15
N	51.62	9.94	51.91	11.03	−0.247	0.805	0.03
E	48.90	11.97	48.55	12.97	0.250	0.803	0.03
O	47.12	9.94	50.90	11.60	−3.114	0.002	0.33 @
A	53.83	10.84	55.87	11.48	−1.597	0.112	0.18
C	48.89	10.00	50.57	11.20	−1.400	0.163	0.15
N1	52.13	11.61	54.84	12.33	−1.972	0.050	0.22 @
E1	50.86	11.68	51.28	11.30	−0.317	0.752	0.04
O1	46.58	10.15	51.23	13.01	−3.605	<0.001	0.36 @
A1	50.89	11.76	50.35	12.09	0.396	0.692	0.05
C1	47.63	10.44	48.29	11.31	−0.523	0.602	0.06
N2	45.86	9.97	45.61	10.48	0.209	0.835	0.02
E2	49.19	11.82	48.00	13.85	0.824	0.411	0.09
O2	47.43	10.14	52.27	11.17	−3.967	<0.001	0.43 @
A2	51.70	9.16	52.76	9.73	−0.970	0.333	0.11
C2	49.94	10.27	49.87	11.77	0.059	0.953	0.01
N3	54.74	10.54	54.71	10.60	0.029	0.977	0.00
E3	50.49	11.11	47.54	11.23	2.269	0.024	0.26 @
O3	49.95	10.83	52.75	12.37	−2.118	0.035	0.23 @
A3	50.02	10.73	54.38	10.68	−3.489	<0.001	0.41 @
C3	50.10	10.49	52.66	10.17	−2.115	0.036	0.25 @
N4	51.13	9.97	52.71	11.20	−1.317	0.189	0.14
E4	48.25	10.63	46.47	11.48	1.402	0.162	0.16
O4	47.45	10.19	46.00	11.67	1.167	0.245	0.12
A4	53.06	11.59	55.46	10.82	−1.813	0.071	0.22
C4	49.71	10.36	51.40	10.03	−1.409	0.160	0.17
N5	51.39	10.33	48.56	11.03	2.308	0.022	0.26 @
E5	46.42	12.84	47.47	12.13	−0.711	0.478	0.09
O5	52.48	10.34	52.69	12.24	−0.164	0.870	0.02
A5	55.64	9.41	56.38	10.44	−0.651	0.516	0.07
C5	46.37	10.58	47.05	12.08	−0.530	0.597	0.06
N6	51.90	11.54	50.66	11.36	0.928	0.355	0.11
E6	50.33	12.23	52.69	12.95	−1.625	0.106	0.18
O6	44.52	9.77	47.01	11.06	−2.102	0.037	0.23 @
A6	53.08	11.29	58.51	11.50	−4.100	<0.001	0.47 #
C6	51.42	10.42	52.75	11.67	−1.053	0.294	0.11

Note: n = 335; QS = Batson’s Quest Scale; EXM = Existential Motives; COM = Complexity; TEN = Tentativeness; EXP = Exploration; CHAN = Change; RA = Religious Angst; N1–6 = facet traits of Neuroticism; E1–6 = facet traits of Extraversion; O1–6 = facet traits of Openness; A1–6 = facet traits of Agreeableness; C1–6 = facet traits of Conscientiousness; N = global trait of Neuroticism; E = global trait of Extraversion; O = global trait of Openness; A = global trait of Agreeableness; C = global trait of Conscientiousness; @ = small ES, and # = medium ES.

, descriptive statistics for the religious and personality variables distinguished by gender are presented. The statistical check for gender differences was computed using an independent samples t-test. Unexpectedly, no religious variables exhibited gender distinctions; there were eleven personality variables (i.e., facet and domain traits) revealing gender differentiations. Ten were small effect sizes (ES) including O (Openness), n1, o1, o2, e3, o3, a3, c3, n5, and o6 (i.e., 0.2–0.4); there was one medium ES (i.e., 0.5) with a6. Specifically, female participants were more open (O) although both men and women’s means were still in the average range which is also regarded as moderate (Howard and Howard 1995, 2010). Further, women were more apprehensive (n1), imaginative (o1), appreciative of art and beauty (o2), receptive to their emotional life (o3), willing to reassess their value systems (e.g., spiritual ones) (o6), interpersonally concerned (altruistic) (a3), tender-minded (a6), and likely to adhere to their personal ethics (c3). Male undergraduates were less likely to have restraint with their appetites and passions (n5) and more socially dominant (e3) (McCrae and Costa 2010). Some of the gender distinctions in this study are supported in the literature. For instance, Kajonius and Johnson (2018) noted that they found women means in n1 (apprehension), o3 (receptive to emotional life), and a3 (altruism) greater that men’s means. Mac Giolla and Kajonius (2019) reported females had higher domain scores than males. Last, Gilmartin et al. (2022) stated women scores in o2 (appreciation for art and beauty), o6 (willingness to reassess value system), and a6 (tendermindedness) were larger than the men with these facets.

Because of the established gender differences in religious and personality phenomena, partial correlations were conducted for religious and personality variables. The matrix is presented in

Table 2. Partial Correlations between Religious Doubt and Five Factor Model Variables.

	QS	EXM	COM	TEN	EXP	CHAN	RA
N1	0.19 **	0.10	0.18 **	−0.04	−0.02	0.13 *	0.34 **
N2	0.15 **	0.12 *	0.02	−0.19 **	−0.13 *	0.18 **	0.26 **
N3	0.22 **	0.27 **	0.21 **	−0.03	−0.03	0.25 **	0.49 **
N4	0.11 *	0.12 *	0.24 **	−0.06	−0.02	0.14 *	0.33 **
N5	0.19 **	0.19 **	0.08	−0.08	0.01	0.14 **	0.18 **
N6	0.08	0.24 **	0.15 **	−0.06	−0.11	0.20 **	0.36 **
E1	−0.08	−0.13 *	−0.04	0.06	0.20 **	−0.15 **	−0.23 **
E2	−0.01	−0.03	−0.15 **	0.02	0.08	−0.06	−0.12 *
E3	−0.03	−0.09	−0.20 **	0.00	0.06	−0.07	−0.14 *
E4	0.04	0.07	−0.06	0.04	0.05	−0.05	−0.06
E5	0.11 *	0.11 *	0.04	−0.00	0.02	0.02	0.05
E6	−0.11 *	−0.07	−0.05	0.06	0.06	−0.19 **	−0.25 **
O1	0.19 **	0.07	0.10	0.13 *	0.18 **	0.10	−0.01
O2	0.17 **	0.14 **	0.29 **	0.12 *	0.25 **	0.18 **	0.09
O3	0.13 *	0.01	0.11 *	0.12 *	0.13 *	0.07	−0.01
O4	0.17 **	0.13 *	0.09	0.09	0.22 **	0.10	−0.07
O5	0.26 **	0.09	0.37 **	0.19 **	0.35 **	0.25 **	0.05
O6	0.32 **	0.18 **	0.27 **	0.22 **	0.04	0.31 **	0.19 **
A1	−0.20 **	−0.13 *	−0.12 *	0.14	0.01	−0.24 **	−0.34 **
A2	−0.11 *	−0.09	−0.09	0.06	−0.06	−0.17 **	−0.11 *
A3	−0.10	−0.09	0.00	0.08	0.14 *	−0.14 *	−0.12 *
A4	−0.15 **	−0.06	0.05	0.06	0.07	−0.12 *	−0.17 **
A5	0.06	0.09	0.13 *	0.06	0.04	−0.04	0.09
A6	0.07	−0.01	0.13 *	0.17 **	0.21 **	0.02	0.08
C1	−0.11 *	−0.21 **	−0.14 *	0.06	0.14 **	−0.12 *	−0.20 **
C2	−0.12 *	−0.19 **	−0.14 **	−0.08	−0.01	−0.15 **	−0.11 *
C3	−0.09	−0.16 **	−0.04	0.04	0.11 *	−0.16 **	−0.17 **
C4	−0.06	−0.14 **	−0.08	0.10	0.14 *	−0.15 **	−0.09
C5	−0.14 **	−0.22 **	−0.13 *	0.03	0.11	−0.19 **	−0.18 **
C6	−0.10	−0.13 *	−0.05	0.02	0.01	−0.07	−0.13 *
N	0.26 **	0.26 **	0.21 **	−0.07	−0.07	0.26 **	0.47 **
E	−0.00	−0.03	−0.12 *	0.04	0.12 *	−0.09	−0.16 **
O	0.31 **	0.15 **	0.31 **	0.25 **	0.30 **	0.25 **	0.06
A	−0.12 *	−0.07	0.02	0.11 *	0.07	−0.20 **	−0.16 **
C	−0.15 **	−0.21 **	−0.12 *	0.01	0.08	−0.20 **	−0.19 **

Note: N = 335; QS = Batson’s Quest Scale; EXM = Existential Motives; COM = Complexity; TEN = Tentativeness; EXP = Exploration; CHAN = Change; RA = Religious Angst; N1–6 = facet traits of Neuroticism; E1–6 = facet traits of Extraversion; O1–6 = facet traits of Openness; A1–6 = facet traits of Agreeableness; C1–6 = facet traits of Conscientiousness; N = global trait of Neuroticism; E = global trait of Extraversion; O = global trait of Openness; A = global trait of Agreeableness; C = global trait of Conscientiousness; * = p ≤ 0.05 ** (2-tailed); and ** = p ≤ 0.01 (2-tailed).

. Due to the large possible associations (i.e., >1500) between doubt-personality, the table does not include doubt-doubt or personality-personality correlations.

In general, the directions of the coefficients were expected. Most correlations were significant at ≤0.05 level (i.e., 53% for facets; 66% for global traits), small in size (96% for facet and global traits; 0.1–0.3), and positive in direction (58% for facets; 57% for global). When analyzing only the coefficients (i.e., significant at ≤0.05) in Table 2 with a personality lens, traits from Neuroticism and Openness were largely positive; specifically, 93% of the associations with the neurotic facets, 100% for the neurotic domain trait, and 100% for the Openness facet and domain traits were positive. Conscientiousness facet and domain traits were negative; in particular, 88% of the correlations with the facets and the domains were 100%. Extraversion traits were predominately unrelated at 69% of the coefficients.

When analyzing the correlations (i.e., significant at ≤0.05) through a religious doubt vantage, the coefficients with the Quest Scale (QS) were 100% positive with Openness; both change (CHAN) and religious angst (RA) were 100% positive with Neuroticism. Associations with the existential motives (EXM) were 100% negative with Conscientiousness. Further, tentativeness (TEN) had the least significant coefficients among facet traits (i.e., 7 out of 30); for instance, TEN had 0% significant coefficients with both Extraversion and Conscientiousness.

3.3. Hierarchical Regression Analyses with Religious Doubt, Gender, and Personality Variables

Results are presented from the hierarchical regression procedures of the influence of personality on religious doubt variables controlling for gender in

Table 3. Hierarchical Regression Analyses for the Influence of Personality (FFM—Facet and Domain traits) and Gender on Religious Doubt Phenomenon.

Facets	Doubt—Exploration (EXP) (n = 339)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	0.758	0.835	−0.049	−1.587	0.818	−0.103
N2				−0.006	0.040	−0.009
E1				0.072	0.040	0.116
O1				0.050	0.036	0.086
O2				0.066	0.040	0.105
O3				−0.043	0.038	−0.074
O4				0.069	0.034	0.110 *
O5				0.138	0.034	0.228 **
A3				−0.056	0.046	−0.086
A6				0.079	0.038	0.131 *
C1				0.076	0.040	0.120
C4				0.072	0.043	0.104
F	0.823			7.223
Sig.	0.365			<0.001
R2 change	0.002			0.207
R2	0.002			0.210 (total)
VIF				1.171–2.109
DW				2.037
Domains	Doubt—Exploration (EXP) (n = 337)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−0.758	0.838	−0.049	−1.450	0.812	−0.094
E				0.031	0.030	0.055
O				0.180	0.034	0.290 **
F	0.818			11.583
Sig.	0.366			<0.001
R2 change	0.002			0.092
R2	0.002			0.094 (total)
VIF				1.028–1.079
DW				1.921
Facets	Doubt—Existential Motives (EXM) (n = 338)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−0.639	0.662	−0.053	−0.703	0.658	−0.058
N2				−0.032	0.035	−0.059
N3				0.098	0.041	0.185 *
N4				−0.045	0.038	−0.086
N5				0.021	0.033	0.041
N6				0.063	0.034	0.131
E1				−0.093	0.032	−0.190 **
E5				0.045	0.028	0.101
O2				0.023	0.030	0.047
O4				0.054	0.029	0.108
O6				0.061	0.029	0.118 *
A1				−0.004	0.029	−0.008
C1				−0.025	0.036	−0.051
C2				−0.007	0.034	−0.015
C3				0.047	0.038	0.088
C4				0.020	0.041	0.036
C5				−0.054	0.041	−0.112
C6				−0.015	0.032	−0.031
F	0.932			3.762
Sig.	0.335			<0.001
R2 change	0.003			0.172
R2	0.003			0.175 (total)
VIF				1.136–2.812
DW				1.917
Domains	Doubt—Existential Motives (EXM) (n = 337)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−0.639	0.662	−0.053	−0.758	0.647	−0.062
N				0.105	0.029	0.201 **
O				0.048	0.027	0.097
C				−0.059	0.029	−0.117 *
F	0.929			8.453
Sig.	0.336			<0.001
R2 change	0.003			0.089
R2	0.003			0.092 (total)
VIF				1.037–1.186
DW				1.828
Facets	Doubt—Complexity (COM) (n = 336)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−0.086	0.824	−0.006	−1.081	0.775	−0.072
N1				0.017	0.039	0.030
N3				−0.009	0.050	−0.014
N4				0.050	0.046	0.078
N6				0.043	0.040	0.072
E2				−0.005	0.030	−0.009
E3				−0.090	0.039	−0.146 *
O2				0.068	0.038	0.110
O3				−0.023	0.034	−0.040
O5				0.185	0.033	0.313 **
O6				0.102	0.034	0.158 **
A1				−0.040	0.033	−0.070
A5				−0.005	0.038	−0.007
A6				0.003	0.036	0.005
F	0.011			70.828
Sig.	0.917			<0.001
R2 change	0.000			0.254
R2	0.000			0.254 (total)
VIF				1.141–2.477
DW				1.941
Domains	Doubt—Complexity (COM) (n = 337)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−0.086	0.822	−0.006	−0.943	0.778	−0.063
N				0.088	0.036	0.137 *
E				−0.088	0.029	−0.163 **
O				0.204	0.033	0.335 **
C				0.007	0.035	0.012
F	0.011			11.690
Sig.	0.917			<0.001
R2 change	0.000			0.150
R2	0.000			0.150 (total)
VIF				1.040–1.213
DW				1.908
Facets	Doubt—Change (CHAN) (n = 336)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−1.023	1.253	−0.045	−1.298	1.208	−0.057
N1				−0.063	0.060	−0.073
N2				0.040	0.072	0.039
N3				0.101	0.078	0.100
N4				−0.065	0.068	−0.066
N5				−0.030	0.060	−0.032
N6				0.134	0.065	0.147 *
E1				−0.070	0.067	−0.075
E6				−0.045	0.059	−0.055
O2				0.031	0.056	0.033
O5				0.191	0.052	0.212 **
O6				0.275	0.052	0.281 **
A1				−0.091	0.053	−0.105
A2				−0.105	0.065	−0.095
A3				−0.054	0.072	−0.056
A4				−0.008	0.065	−0.008
C1				0.029	0.065	0.031
C2				−0.006	0.060	−0.006
C3				0.102	0.072	0.102
C4				−0.039	0.075	−0.038
C5				−0.101	0.074	−0.111
F	0.667			5.377
Sig.	0.415			<0.001
R2 change	0.002			0.262
R2	0.002			0.264 (total)
VIF				1.187–2.814
DW				2.017
Domains	Doubt—Change (CHAN) (n = 337)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	−1.023	1.251	−0.045	−1.541	1.173	−0.067
N				0.158	0.054	0.160 **
O				0.249	0.050	0.268 **
A				−0.198	0.050	−0.212 **
C				−0.058	0.052	−0.061
F	0.669			13.040
Sig.	0.414			<0.001
R2 change	0.002			0.162
R2	0.002			0.164 (total)
VIF				1.038–1.197
DW				1.857
Facets	Doubt—Tentativeness (TEN) (n = 338)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	0.818	1.018	0.044	0.114	1.017	0.006
N2				−0.140	0.047	−0.167 **
O1				0.023	0.044	0.032
O2				−0.025	0.050	−0.033
O3				0.041	0.046	0.058
O5				0.087	0.044	0.119 *
O6				0.132	0.046	0.166 **
A6				0.028	0.046	0.039
F	0.645			4.884
Sig.	0.422			<0.001
R2 change	0.002			0.104
R2	0.002			0.106 (total)
VIF				1.090–1.563
DW				1.789
Domains	Doubt—Tentativeness (TEN) (n = 336)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	0.818	1.021	0.044	0.021	1.004	0.001
O				0.180	0.041	0.238 **
A				0.047	0.041	0.062
F	0.641			7.991
Sig.	0.424			<0.001
R2 change	0.002			0.065
R2	0.002			0.067 (total)
VIF				1.028–1.074
DW				1.763
Facets	Doubt—Religious Angst (RA) (n = 335)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	1.299	1.027	0.069	1.008	0.904	0.054
N1				0.048	0.047	0.068
N2				−0.054	0.055	−0.064
N3				0.238	0.059	0.290 **
N4				0.056	0.055	0.070
N5				−0.045	0.046	−0.057
N6				0.100	0.050	0.133 *
E1				−0.119	0.055	−0.156 *
E2				0.049	0.043	0.076
E3				0.019	0.048	0.025
E6				−0.024	0.044	−0.035
O6				0.131	0.039	0.163 **
A1				−0.116	0.040	−0.164 **
A4				−0.031	0.051	−0.040
C1				0.012	0.050	0.016
C2				0.028	0.047	0.036
C3				0.049	0.053	0.059
C5				0.001	0.053	0.002
C6				−0.133	0.045	−0.175 **
F	1.601			8.717
Sig.	0.207			<0.001
R2 change	0.005			0.339
R2	0.005			0.344 (total)
VIF				1.112–2.527
DW				2.044
Domains	Doubt—Religious Angst (RA) (n = 336)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	1.299	1.025	0.069	1.322	0.910	0.070
N				0.353	0.043	0.437 **
E				−0.038	0.033	−0.056
A				−0.050	0.038	−0.066
C				−0.019	0.040	−0.025
F	1.606			20.360
Sig.	0.206			<0.001
R2 change	0.005			0.230
R2	0.005			0.235 (total)
VIF				1.013–1.229
DW				2.137
Facets	Doubt—Quest Scale (QS) (n = 335)
	Model 1			Model 2
	B	SE B	β	B	SE B	Β
Gender	1.862	1.828	0.056	1.305	1.786	0.039
N1				0.120	0.088	0.095
N2				−0.024	0.106	−0.016
N3				0.055	0.114	0.038
N4				−0.049	0.104	−0.035
N5				0.054	0.089	0.039
E5				0.047	0.078	0.039
E6				−0.135	0.081	−0.112
O1				0.062	0.083	0.049
O2				−0.022	0.088	−0.016
O3				0.015	0.086	0.012
O4				0.185	0.079	0.135 *
O5				0.216	0.077	0.165 **
O6				0.334	0.078	0.234 **
A1				−0.173	0.078	−0.137 *
A2				0.007	0.093	0.004
A4				−0.121	0.096	−0.087
C1				−0.039	0.089	−0.028
C2				0.053	0.091	0.039
C5				−0.042	0.099	−0.032
F	1.038			4.843
Sig.	0.309			<0.001
R2 change	0.003			0.232
R2	0.003			0.235 (total)
VIF				1.174–2.498
DW				1.957
Domains	Doubt—Quest Scale (QS) (n = 336)
	Model 1			Model 2
	B	SE B	Β	B	SE B	Β
Gender	1.862	1.825	0.056	0.500	1.710	0.015
N				0.269	0.079	0.187 **
O				0.435	0.073	0.322 **
A				−0.202	0.072	−0.149 **
C				0.001	0.076	0.001
F	1.041			13.330
Sig.	0.308			<0.001
R2 change	0.003			0.165
R2	0.003			0.168 (total)
VIF				1.038–1.197
DW				1.923

Note: B = unstandardized beta; SE B = standard error of beta; β = standardized beta; N1–6 = facet traits of Neuroticism; E1–6 = facet traits of Extraversion; O1–6 = facet traits of Openness; A1–6 = facet traits of Agreeableness; C1–6 = facet traits of Conscientiousness; N = domain trait of Neuroticism; E = domain trait of Extraversion; O = domain trait of Openness; A = domain trait of Agreeableness; C = domain trait of Conscientiousness; VIF = variance inflation factor; DW = Durbin Watson; * = p ≤ 0.05 *; and ** = p ≤ 0.01.

.

Eleven different facet and five domain traits emerged. There were two consistent predictors among the facets. Across the six criterion variables (excluding Quest Scale [QS]), values (o6) appeared as a predictor five times while ideas (o5) appeared four times. Collectively, Openness facets (i.e., o4, o5, o6) appeared 10 times, Neuroticism traits (n2, n3, n6)—five times, Extraversion (e1, e3)—three times, Agreeableness (a1, a6)—2 times, and Conscientiousness (c6)—one time. Regarding domain traits, all five surfaced. Amid the six religious doubt criterion measures (excluding QS), Openness and Neuroticism materialized four times as predictors, while Agreeableness, Extraversion, and Conscientiousness, one time each.

For religious doubt (Stage 1—Precipitant[s]) as measured by the exploration scale (EXP) of the Multidimensional Quest Orientation Scale (MQOS; Beck and Jessup 2004), gender in model one was a non-contributing trait to the predictive model, F (1, 338) = 0.823, p = 0.365. The placement of 11 facet traits and gender in model two was significant, F (12, 327) = 7.223, p < 0.001. Three predictors emerged; actions (o4), ideas (o5) and tendermindedness (a6) were positive and explained 21.0% of the variance (McCrae and Costa 2010). These findings were partially expected outcomes. From an empirical vantage, agreeableness facets were expected to be positive. But a theoretical perspective anticipated agreeableness facets to be negative. Regarding the prediction of EXP with domain traits, gender in model one was not a contributor, F (1, 335) = 0.818, p = 0.366. Placement of two domain traits and gender in model two was significant, F (3, 334) =11.583, p < 0.001. Only Openness materialized as a positive predictor explaining 9.4% of the variance in EXP. This was an anticipated finding.

For religious doubt (Stage 2—Cognitive/Emotional State) as measured by the existential motive scale (EXM) of the MQOS, the outcome indicated gender in model one did not contribute to the regression model, F (1, 337) = 0.932, p = 0.335. Introduction of the 17 facet traits and gender in model two was significant, F (18, 320) = 3.762, p < 0.001. Three predictors surfaced, explaining 17.5% of the variance. Depression (n3) and values (o6) were positive in direction and expected while warmth (e1) was negative, a partially expected outcome (McCrae and Costa 2010). From a theoretical vantage, −e1 was anticipated; empirically it was expected to be non-significant. Concerning the prediction of EXM with domain traits, gender in model one did not contribute to the regression model, F (1, 336) = 0.929, p = 0.336. In model two, placement of three domain traits and gender was significant, F (4, 333) = 8.453, p < 0.001. Two obtained domain traits emerged and explained 9.2% of the variation in EXM. Neuroticism as a positive predictor and Conscientiousness as a negative one were anticipated outcomes.

Regarding religious doubt (Stage 2—Cognitive/Emotional State) as measured by the complexity scale (COM) of the MQOS, gender in model one was not a contributor to the predictive model, F (1, 335) = 0.011, p = 0.917. The entry of 13 facet traits and gender in model two was significant, F (14, 322) = 7.828, p < 0.001. Three predictors emerged; both ideas (o5) and values (o6) were positive while assertiveness (e3) was negative explaining 25.4% of the variance (McCrae and Costa 2010). The former two were expected from a theoretical and empirical vantage; the last one was theoretically but not empirically expected. For the prediction of COM with domain traits, gender in model one was again a non-contributor to the model, F (1, 336) = 0.011, p = 0.917. Placement of four domain traits and gender in model two was significant, F (5, 332) = 11.690, p < 0.001. The obtained predictors, Neuroticism and Openness, as positive predictors were anticipated results. Extraversion as a negative predictor was expected only from a theoretical vantage, not an empirical one. Collectively, the threesome explained 15% of the variance in COM.

For religious doubt (Stage 3—Coping Response[s]) as measured by the change scale (CHAN) of the MQOS, gender in model one was a non-contributing trait to the predictive model, F (1, 335) = 0.667, p = 0.415. Entry of 20 facet traits and gender in model two was significant, F (21, 315) = 5.377, p < 0.001. Three predictors emerged; vulnerability (n6), ideas (o5) and values (o6) were positive and projected results (McCrae and Costa 2010). Collectively, the three explained 26.4% of the variance. Concerning the prediction of CHAN with domain traits, gender in model one was discovered to not be a contributor to the model, F (1, 336) = 0.669, p = 0.414. Introduction of four domain traits and gender in model two was significant, F (5, 332) = 13.040, p < 0.001. The emergent predictors, Neuroticism and Openness, surfaced as positive predictors and Agreeableness was a negative one. The former two were anticipated while the latter was expected only from a theoretical view not an empirical one. Collectively, they explained 16.4% of the variance in CHAN.

Concerning religious doubt (Stage 4—Outcome[s]) as measured by the tentativeness scale (TEN) of the MQOS, gender in model one was also a non-contributor to the predictive model, F (1, 337) = 0.645, p = 0.422. The entry of seven facet traits and gender in model two was significant, F (8, 330) = 4.884, p < 0.001. Three predictors emerged; both ideas (o5.) and values (o6) were positive while angry hostility (n2) was negative (McCrae and Costa 2010). The former two were expected and the last one was not. Jointly, the trio explained 10.6% of the variance. Regarding the prediction of TEN with domain traits, gender in model one was not a contributor to the model, F (1, 335) = 0.641, p = 0.424. Placement of two domain traits and gender in model two was significant, F (3, 333) = 7.991, p < 0.001. Only, Openness materialized as a positive predictor explaining 6.7% of the variance in TEN. This too was a foreseeable finding.

For religious doubt (Stage 4—Outcome[s]) as measured by the religious angst scale (RA) of the MQOS, gender in model one was a non-contributing trait to the predictive model, F (1, 334) = 1.601, p = 0.207. Entry of 18 facet traits and gender in model two was significant, F (19, 316) = 8.717, p < 0.001. Six predictors materialized: depression (n3), vulnerability (n6), and values (o6) as positive and deliberation (c6) as negative predictors were projected results. Reserved (−e1) and mistrust (−a1) were negative and partially unexpected (McCrae and Costa 2010). Theoretically, the two were anticipated; empirically speaking, facets from Extroversion were projected to be non-significant and facets from Agreeableness were expected to be positive. Together, the six explained 34.4% of the variance. Pertaining to the prediction of RA with domain traits, gender in model one was found as a non-contributor to the model, F (1, 335) = 1.606, p = 0.206. The entry of four domain traits and gender in model two was significant, F (5, 331) = 20.360, p < 0.001. Only Neuroticism surfaced as a positive predictor—an assumed outcome; it explained 23.5% of the variance in RA.

Pertaining to religious doubt as measured by the Quest Scale (QS; Batson 1976), the findings revealed gender in model one not contributing to the regression model, F (1, 334) = 1.038, p = 0.309. The introduction of the personality variables—19 facet traits, and gender in model two was significant, F (20, 315) = 4.843, p < 0.001. The obtained predictors explained 23.5% of the variation in QS. Although gender was not a predictor, actions (o4), ideas (o5), and values (o6) were positive predictors and estimated results, whereas mistrust (−a1) was anticipated from a theoretical view not an empirical one. Concerning the prediction of QS with domain traits, gender in model one did not contribute to the regression model, F (1, 335) = 1.041, p < 0.308. The introduction of the domain traits and gender in model two was significant, F (5, 331) = 13.330, p < 0.001 and explained 16.8% of the variance with QS; both Openness and Neuroticism were positive predictors and anticipated. Agreeableness was a negative one and was an awaited result from a theoretical vantage but not from an empirical one.

3.4. Additional Regression Analyses with Religious Doubt, Gender, and Personality Variables

Table 4. Additional Regression Analyses for the Influence of Personality. (FFM—Obtained Facet and Domain traits) on Religious Doubt Phenomenon.

Facets	Doubt—Exploration (EXP) (n = 340)
	B	SE B	Β	R2 change @
O5	0.180	0.032	0.299 **	0.092
O4	0.080	0.033	0.126 *	0.050
A6	0.076	0.031	0.121 **	0.014
F	20.787
Sig.	<0.001
R2 total				0.156
VIF	1.043–1.091
DW	2.049
Domains	Doubt—Exploration (EXP) (n = 337)
	B	SE B	β	R2 change @
O	0.178	0.032	0.287 **	0.082
F	30.087
Sig.	<0.001
R2 total				0.082
VIF	1.000
DW	1.997
Facets	Doubt—Existential Motives (EXM) (n = 338)
	B	SE B	Β	R2 change @
N3	0.120	0.028	0.226 **	0.072
E1	−0.053	0.026	−0.107 *	0.022
O6	0.079	0.028	0.153 **	0.006
F	12.487
Sig.	<0.001
R2 total				0.101
VIF	1.056–1.073
DW	1.968
Domains	Doubt—Existential Motives (EXM) (n = 337)
	B	SE B	Β	R2 change @
N	0.107	0.029	0.205 **	0.065
C	−0.070	0.028	−0.138 *	0.017
F	14.838
Sig.	<0.001
R2 total				0.081
VIF	1.142
DW	1.855
Facets	Doubt—Complexity (COM) (n = 339)
	B	SE B	Β	R2 change @
O5	0.208	0.030	0.352 **	0.153
E3	−0.135	0.030	−0.219 **	0.037
O6	0.107	0.032	0.167 **	0.026
F	30.963
Sig.	<0.001
R2 total				0.217
VIF	1.020–1.079
DW	1.985
Domains	Doubt—Complexity (COM) (n = 337)
	B	SE B	Β	R2 change @
O	0.197	0.032	0.323 **	0.096
N	0.087	0.034	0.134 *	0.045
E	−0.086	0.029	−0.159 **	0.005
F	19.022
Sig.	<0.001
R2 total				0.146
VIF	1.088–1.127
DW	1.881
Facets	Doubt—Change (CHAN) (n = 340)
	B	SE B	Β	R2 change @
O5	0.197	0.047	0.218 **	0.076
O6	0.231	0.050	0.236 *	0.052
N6	0.203	0.046	0.223 **	0.042
F	22.904
Sig.	<0.001
R2 total				0.169
VIF	1.021–1.083
DW	2.018
Domains	Doubt—Change (CHAN) (n = 337)
	B	SE B	Β	R2 change @
N	0.176	0.052	00.179 **	0.067
O	0.250	0.049	00.269 **	0.045
A	−0.206	0.049	−0.222 **	0.044
F	20.556
Sig.	<0.001
R2 total				0.156
VIF	1.080–1.117
DW	1.915
Facets	Doubt—Tentativeness (TEN) (n = 338)
	B	SE B	Β	R2 change @
N2	−0.142	0.044	0.169 **	0.035
O6	0.152	0.043	0.190 **	0.034
O5	0.094	0.040	0.128 *	0.030
F	12.233
Sig.	<0.001
R2 total			0.099
VIF	1.015–1.078
DW	1.838
Domains	Doubt—Tentativeness (TEN) (n = 336)
	B	SE B	Β	R2 change @
O	0.190	0.040	0.252 **	0.063
F	22.692
Sig.	<0.001
R2 total				0.063
VIF	1.000
DW	1.799
Facets	Doubt—Religious Angst (RA) (n = 337)
	B	SE B	β	R2 change @
N3	0.263	0.047	0.321 **	0.237
O6	0.129	0.038	0.161 **	0.026
A1	−0.115	0.038	−0.162 **	0.021
E1	−0.088	0.039	−0.116 *	0.016
C6	−0.083	0.035	−0.109 *	0.011
N6	0.079	0.041	0.105	0.010
F	26.105
Sig.	<0.001
R2 total				0.321
VIF	1.045–1.612
DW	2.073
Domains	Doubt—Religious Angst (RA) (n = 337)
	B	SE B	β	R2 change @
N	0.382	0.039	0.472 **	0.223
F	96.448
Sig.	<0.001
R2 total				0.223
VIF	1.000
DW	2.109
Facets	Doubt—Quest Scale (QS) (n = 338)
	B	SE B	Β	R2 change @
O6	0.384	0.073	0.269 **	0.064
A1	−0.299	0.063	−0.236 **	0.054
O5	0.220	0.068	00.168 **	0.050
O4	0.150	0.072	0.109 *	0.027
F	20.267
Sig.	<0.001
R2 total				0.195
VIF	1.031–1.124
DW	1.970
Domains	Doubt—Quest Scale (QS) (n = 336)
	B	SE B	Β	R2 change @
O	0.437	0.071	0.324 **	0.080
N	0.269	0.075	0.187 **	0.067
A	−0.201	0.072	−0.148 **	0.020
F	22.316
Sig.	<0.001
R2 total				0.167
VIF	1.080–1.117
DW	1.899

Note: B = unstandardized beta; SE B = standard error of beta; β = standardized beta; N2,3,6 = facet traits of Neuroticism; E1,3 = facet traits of Extraversion; O4,5,6 = facet traits of Openness; A1,6 = facet traits of Agreeableness; C6 = facet traits of Conscientiousness; N = domain trait of Neuroticism E = domain trait of Extraversion; O = domain trait of Openness; A = domain trait of Agreeableness; C = domain trait of Conscientiousness; VIF = variance inflation factor; DW = Durbin Watson; @ = stepwise hierarchical regression was employed; * = p ≤ 0.05; and ** = p ≤ 0.01.

shows outcomes from additional regression analyses (i.e., standard multiple regression [SMR] and stepwise regression) with only obtained predictors per criterion variable. The purposes were to ascertain how each variable significantly contributed to the model without the non-significant facets, domains, or gender and to estimate the explained variance per predictor. As in Table 3, two tables are dedicated to each predictor—facets followed by the domains.

In the SMR model predicting religious doubt (Stage 1—Precipitant[s]) via exploration (EXP) of the MQOS (Beck and Jessup 2004) with only the emergent predictors noted in Table 3, the three facets surfaced again. Ideas (+o5), actions (+o4), and tendermindedness (+a6) explicated 15.6% of the variance as compared to 21% noted in Table 3 (McCrae and Costa 2010). In addition, the estimates of the R² changes per predictor, 9.2% (ideas), 5% (actions), and 1.4% (tendermindedness), disclosed how each one contributed uniquely to the overall R². Regarding the predictive model of religious doubt with EXP involving domain traits, the one variable, Openness (+O), re-materialized explaining 8.2% of the variance of EXP as compared to 9.4% in Table 3.

In the SMR model predicting religious doubt (Stage 2—Cognitive/Emotional State) with existential motives (EXM) of the MQOS, the three obtained facets, depression (+n3), reserved (−e1), and values (+o6) re-appeared (McCrae and Costa 2010). Collectively, the R² total of the trio was 0.101 relative to the R² of 0.175 in Table 3. Further, the estimates of R² change, 0.072 (+n3), 0.022 (−e1), and 0.006 (+o6), revealed the unique contribution from each predictor to the total explained variance. Pertaining to the model of religious doubt with EXM including emergent domain traits, Neuroticism (+N) and Conscientiousness (−C) ascended again as predictors. The duo explained 8.1% of the variance as compared to 9.2% in Table 3. In addition, the R² change estimates, 6.5% (+N) and 1.7% (−C), noted the unique contributions from each respective predictor.

In the SMR predicting religious doubt (Stage 2—Cognitive/Emotional State) with complexity (COM) of the MQOS, the obtained predictors, ideas (+o5), non-assertive (−e3), and values (+o6) surfaced again explaining 21.7% of the explained variance relative to the 25.4% displayed in Table 3 (McCrae and Costa 2010). And values for R² change for each facet predictor, 15.3% (+o5), 3.7% (−e3), and 2.6% (+o6), underlined the possible estimate and unique contribution for each variable. Regarding the predictive model of religious doubt with COM and domain traits, the trio of emergent predictors, Openness (+O), Neuroticism (+N), and Extraversion (−E) clarified 14.6% of explained variance relative to the 15.0% in Table 3. Also, the estimated R² change values, 0.096 (+O), 0.045 (+N), and 0.005 (−E) indicate the unique contribution to explained variance per predictor.

In the SMR predictive model of doubt (Stage 3—Coping Response[s]) via change (CHAN) of the MQOS, the re-emergent facet variables, ideas (+o5), values (+o6), and vulnerability (+n6) explained 16.9% of the variance in CHAN as compared to 26.4% in Table 3 (McCrae and Costa 2010). Further, the R² change estimates for each predictor, 7.6% (+o5), 5.2% (+o6), and 4.2% (+n6), display the unique contribution per variable to the total explained variance. Concerning the model for the prediction of CHAN with domain traits, the three variables re-materialized having a R² of 0.156 relative to 0.164 in Table 3. The estimated R² change values of Neuroticism (+N), Openness (+O), and Agreeableness (−A), were 0.067, 0.045, and 0.044, respectively, and underlined the unique influence each predictor offered to the total explained variance.

In the SMR model of doubt (Stage 4—Outcome[s]) predicting tentativeness (TEN) of the MQOS, slow to anger (−n2), values (+o6), and ideas (+o5) resurfaced explaining 9.9% of the explained variance in TEN relative to 10.6% displayed in Table 3 (McCrae and Costa 2010). The predictors’ R² change estimates, 0.035 (−n2), 0.034 (+o6), and 0.030 (+o5), pointed to the unique contribution from each variable. Regarding the predictive model of TEN with obtained domain traits, Table 4 revealed Openness (+O) re-materializing and explaining 6.3% of the total explained variance as compared to 6.7% in Table 3.

In the SMR predictive model of doubt (Stage 4—Outcome[s]) with religious angst (RA) of the MQOS, five of the six emergent predictors re-surfaced; vulnerability (n6) was eliminated. Collectively, depression (+n3), values (+o6), mistrust (−a1), reserved (−e1), and impulsive (−c6) had a R² total of 0.321 relative to 0.344 in Table 3 (McCrae and Costa 2010). In addition, the predictors’ estimated R² change values, 0.237 (+n3), 0.026 (+o6), 0.021 (−a1), 0.016 (−e1), and 0.011 (−c6), indicated that each of the fivesome offer a unique portion of the overall R² total. Concerning the prediction of RA with domain traits, the only obtained predictor, Neuroticism (+N) surfaced again explaining 22.3% of the explained variance as compared to 23.5% in Table 3.

Last, in the SMR model predicting the Quest Scale (QS; Batson 1976) with only the ascertained facet predictor variables, values (+o6), mistrust (−a1), ideas (+o5), and actions (+o4) re-emerged explaining 19.5% of the variance relative to the 23.5% in the model with 19 facets in Table 3 (McCrae and Costa 2010). Further, as can be seen in Table 4, the stepwise regression with the four predictors displayed estimates of each predictor’s unique contribution, 6.4% (+o6), 5.4% (−a1), 5.0% (+o5), and 2.7% (+o4), to the overall explained variance. Pertaining to the predictive model of QS with only the obtained domain variables, the three predictors—Openness (+O), Neuroticism (+N), and Agreeableness (−A) materialized again explaining 16.7% of the variance relative to 16.8% in Table 3. In addition, the stepwise procedure revealed estimates of explained variance pointing to how each of threesome, 8.0% (+O), 6.7% (+N), and 2.0% (−A), are contributing uniquely to the total.

Therefore, in general from the additional regression analyses, this study discovered 11 different facet traits and five domain traits as predictors of doubt phenomena (excluding QS). Values (+o6) and ideas (+o5) were still the consistent and most frequent facet predictors (McCrae and Costa 2010); the former emerged five times while the latter surfaced four times. Facets from Openness materialized 10 times and from Neuroticism, facets emerged four times which is one less than the report from Table 3. Further, the range of total explained variance was 9.9% to 32.1%, which is again slightly different from the range in Table 3, 10.6% to 34.4%.

Concerning domain traits, the frequency of obtained predictors was like Table 3. Specifically, Openness and Neuroticism appeared four times, while Agreeableness, Extraversion, and Conscientiousness appeared one-time. The range of total R² was slightly different. In Table 4, the range for the domains was 6.3% to 22.3% while, in Table 3, the range for the domains was 6.7% to 23.5%.

3.5. Suppression Effects

An investigation of suppressive effects was conducted with facet and domain variables. Suppression in multiple regression investigations can be unnoticed and even disregarded; yet its influence can impair the interpretation of predictive models (Guinn 2019; Martinez Gutierrez and Cribbie 2021). For instance, a suppressive variable can uniquely alter the relationship in both positive and negative directions between criterion and predictor variables. Further, when predictive models have several predictive variables (e.g., >2), detection can become a challenge.

Two procedures were executed to detect and verify the existence of suppressors. According to Tabachnick and Fidell (2019), the first procedure was to perform comparisons between the partial correlation of the predictor and criterion variable with the standardized beta weight for predictor variables. Suspected suppressors are distinguished when the beta weight of the predictor variable is “significantly different from zero”, and the beta weight of the predictor variable is larger than the “absolute value of the correlation coefficient” between the predictor and criterion variables (Tabachnick and Fidell 2019, p. 133). Among the facets, three of twenty-five comparisons met the criteria; with the domains, three of the fourteen comparisons fulfilled the criteria.

Table 5. Comparisons of Correlation Coefficients between Predictor and Criterion Variable with the Beta Weights of Predictor Variables.

Predictor Variable	Criterion Variable	Absolute Value of the Partial Correlation Coefficient Between Predictor and Criterion Variables	Beta Weight of Predictor Variable
−E3	COM	0.20	−0.219
+N6	CHAN	0.20	0.223
−A1	QS	0.20	−0.236
+O	COM	0.31	0.323
+O	CHAN	0.25	0.269
+O	QS	0.31	0.324

Note: −E3 (not assertive) = facet trait 3 of Extraversion; N6 (vulnerability) = facet trait 6 of Neuroticism; −A1 (mistrust) = facet trait 1 of Agreeableness; O = domain trait; Openness; COM = complexity from the Multidimensional Quest Scale (MQOS) of Beck and Jessup; CHAN = change from the MQOS; and QS = Batson’s Quest Scale.

presents these findings.

The second procedure entailed removing possible suppressor variables from the original predictive model and noting any alterations in the regression outcomes. In particular, changes in explained variance, obtained predictors, and standardized beta weights relative to the original model need to be observed. According to Tabachnick and Fidell (2019), one of the effects of a suppressor variable is to “suppress variance that is irrelevant to the prediction of the [criterion variable]” (p. 133). Outcomes from this second procedure were displayed in

Table 6. Comparisons of Standard Regression Models for the Detection of Personality Suppression Variables in the Regression of Religious Doubt Phenomenon.

Facets	Doubt	Complexity (COM)	n = 339	Facets	Doubt	Complexity (COM)	n = 340
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
O5	0.208	0.030	0.352 **	+O5	0.192	0.030	0.324 **
E3	−0.135	0.030	−0.219 **	−E3	---	---	---
O6	0.107	0.032	0.167 **	O6	0.120	0.033	0.187 **
F	30.963			F	34.467
Sig.	<0.001			Sig.	<0.001
R2 total	0.217			R2 total	0.169
VIF	1.020–1.079			VIF	1.062
DW	1.985			DW	1.975
Facets	Doubt	Change (CHAN)	n = 340	Facets	Doubt	Change (CHAN)	n = 340
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
O5	0.197	0.047	0.218 **	O5	0.168	0.047	0.186 **
O6	0.231	0.050	0.236 **	O6	0.247	0.051	0.252 **
N6	0.203	0.046	0.223 **	N6	---	---	---
F	22.904			F	23.199
Sig.	<0.001			Sig0.	<0.001
R2 total	0.169			R2 total	0.121
VIF	1.021–1.083			VIF	1.062
DW	2.018			DW	2.077
Facets	Doubt	Quest Scale (QS)	n = 338	Facets	Doubt	Quest Scale (QS)	n = 339
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
O6	0.384	0.073	0.269 **	O6	0.337	0.075	0.264 **
A1	−0.299	0.063	−0.236 **	A1	---	---	---
O5	0.220	0.068	0.168 **	O5	0.223	0.070	0.177 **
O4	0.150	0.072	0.109 *	O4	0.093	0.073	0.067
F	20.267			F	18.396
Sig.	<0.001			Sig.	<0.001
R2 total	0.195			R2 total	0.141
VIF	1.031–1.124			VIF	1.082–1.119
DW	1.970			DW	1.921
Domains	Doubt	Complexity (COM)	n = 337	Domains	Doubt	Complexity (COM)	n = 338
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
O	0.197	0.032	0.323 **	O	---	---	---
N	0.087	0.034	0.134 *	N	0.126	0.035	0.195 **
E	−0.086	0.029	−0.159 **	E	−0.041	0.030	−0.075
F	19.022			F	8.885
Sig.	<0.001			Sig.	<0.001
R2 total	0.146			R2 total	0.050
IF	1.088–1.127			VIF	1.053
DW	1.881			DW	1.879
Domains	Doubt	Change (CHAN)	n = 337	Domains	Doubt	Change (CHAN)	n = 338
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
N	0.176	0.052	0.179 **	N	0.225	0.053	0.228 **
O	0.250	0.049	0.269 **	O	---	---	---
A	−0.206	0.049	−0.222 **	A	−0.141	0.050	−0.151 **
F	20.556			F	16.403
Sig.	<0.001			Sig0.	<0.001
R2 total	0.156			R2 total	0.089
VIF	1.080–1.117			VIF	1.043
DW	1.915			DW	1.932
Domains	Doubt	Quest Scale (QS)	n = 336	Domains	Doubt	Quest Scale (QS)	n = 336
Model 1a	B	SE B	Β	Model 1b	B	SE B	Β
O	0.437	0.071	0.324 **	O	---	---	---
N	0.269	0.075	0.187 **	N	0.355	0.077	0.247 **
A	−0.201	0.072	−0.148 **	A	−0.086	0.073	−0.064
F	22.316			F	12.814
Sig.	<0.001			Sig0.	<0.001
R2 total	0.167			R2 total	0.071
VIF	1.080–1.117			VIF	1.043
DW	1.899			DW	1.911

Note: B = unstandardized beta; SE B = standard error of beta; β = standardized beta; Model 1a = original standard regression model with possible suppressor variable; Model 1b = alternative standard regression model without possible suppressor variable; N6 (vulnerability) = facet trait of Neuroticism; E3 = (not assertive) facet trait of Extraversion; O4–6 = facet traits of Openness; A1 (mistrust) = facet trait of Agreeableness; N = domain trait of Neuroticism; E = domain trait of extraversion; O = domain trait of Openness; A = domain trait of Agreeableness; VIF = variance inflation factor; DW = Durbin Watson; * = p ≤ 0.05; and ** = p ≤ 0.01.

.

When non-assertiveness (−e3) was removed from the standard multiple regression (SMG) equation for the prediction of complexity (COM) of the MQOS (Beck and Jessup 2004), there were a few changes. The total explained variance decreased; there were slight changes in the variables’ beta weight, and no alterations in other predictors. Specifically, R² in the original equation was 0.217, while R² in the alternative equation was 0.169. This was a 4.8% reduction in the explained variance; this decrease in R² was slightly larger than the estimate of R² change (3.7%) in Table 4. That means more R² was lost without non-assertiveness. Comparisons of standardized beta weights revealed, one variable slightly decreased (ideas; +o5) and one slightly increased (values; +o6) (McCrae and Costa 2010). Last, there were no changes in the emergent predictors. Hence, these findings pointed to non-assertiveness (−e3) functioning as a suppressor that was enhancing the regression outcomes (Tabachnick and Fidell 2019).

When vulnerability (+n6) was eliminated from the SMR equation for the prediction of change (CHAN) of the MQOS, there were some alterations like the removal of non-assertiveness (−e3) with COM. The total explained variance decreased from 0.169 to 0.121, another 4.8% reduction; this alteration in R² was also slightly more than the estimated R² change (i.e., 4.2%) in Table 4. Standardized beta weights of the predictors decreased (ideas; +o5) and increased (values; +o6). And there were no changes in emergent predictors. Thus, these outcomes indicated that vulnerability (+n6) was a suppressor enhancing the SMR outcomes (McCrae and Costa 2010; Tabachnick and Fidell 2019).

When mistrust (−a1) was removed from the SMR equation for the prediction of the Quest Scale (QS), there were notable changes. The total explained variance decreased from 0.195 to 0.141; this was a 5.4% decrease. For the standardized beta weights of the predictors, there were two decreases (values [+o6] and actions [+o4]) and one increase (ideas [+o5]) (McCrae and Costa 2010). Further, a predictor, actions (o4), dropped out of the set. Collectively, these results pointed to mistrust (−a1) being a suppressor enhancing SMR outcomes with its presence (Tabachnick and Fidell 2019).

An examination of the domain predictors, Openness (+O) emerged three times as a possible suppressor with different criterion variables. When Openness (+O) was eliminated from the SMR equation for the prediction of complexity (COM) of the MQOS, there were a few important changes. Total explained variance dropped from 0.146 to 0.050 which was a 9.6% reduction. It is important to note that this loss was equal to the estimated R² change contribution in Table 4. Changes with the standardized beta weights entailed one increase (+Neuroticism) and one decrease (−Extraversion). Yet, the outstanding change was that −Extraversion was eliminated from the set of emergent predictors. Consequently, Openness appeared to be a suppressor enhancing the SMR outcomes (Tabachnick and Fidell 2019).

When Openness (+O) was removed from the SMR equation for the prediction of change (CHAN) of the MQOS, there were a few important alterations. The total R² dropped from 0.156 to 0.089; this was a 6.7% decrease. Based on Table 4, the estimated R² change for CHAN was 0.045; thus, the loss in R² was greater than the estimated contribution. Modifications with the standardized beta weights for the emergent predictors in Table 6 split; there were both an increase (i.e., +Neuroticism was altered from 0.179 to 0.228) and a decrease (i.e., −Agreeableness changed from −0.222 to −0.151). Further, there were no alterations with the emergent variables. Hence, these outcomes pointed to Openness (+O) operating as a suppressor benefitting the other predictors in the predictive set (Tabachnick and Fidell 2019).

When Openness (+O) was removed from the SMR equation for the prediction of the QS, there were notable changes. First, the total explained variance fell from 16.7% to 7.1%; this was a 9.6% decline. It is important to note that the estimated contribution of the R² change for Openness was 8.0% in Table 4; the loss appeared to be greater than the estimated contribution. Second, changes in the standardized beta weight values split; Neuroticism (+N) increased from 0.187 to 0.247 while −Agreeableness (−A) decreased from −0.148 to −0.064. Third, a predictor (−A) was eliminated. Collectively, these findings indicated Openness (+O) functioned as a suppressor enhancing the influence of the other variables in the predictive set (Tabachnick and Fidell 2019).

4. Discussion

Whether religious doubt is perceived positively or negatively, religious people still need a management strategy for coping. Snowden (1910) remarked, “[Christians should not] think less [about their doubt] until [they] relapse into mental stagnation”, [instead they need] to “think more [working their way toward] clearer light and larger truth” (p. 154). Toward that end, a sample of Protestant Christian undergraduates were recruited for the intention of ascertaining personality traits predicting each of the four stages in the doubt process developed by Krause and Ellison (2009). Obtained personality predictors provided insight into the formation of a doubt management strategy per stage benefiting religious college students and their mentors (i.e., pastoral counselors, etc.) journeying with them. Important implications and applications of these findings warrant further discussion.

4.1. Stage 1—Predictors for the Precipitant Phase in the Doubt Process

Based on Krause and Ellison’s (2009) model, religious doubt has a beginning. A trigger initiates the process. Not all precipitants are the same. Those mentioned or studied in the literature include new religious information (Allport 1950; Batson 1976), stressful circumstances (Ingersoll-Dayton et al. 2002), and social conflicts (Krause and Ellison 2009).

For this study, new religious knowledge was the focused trigger. It was operationalized with the construct, exploration (EXP), a subscale of the Multidimensional Quest Orientation Scale (MQOS; Beck and Jessup 2004). This criterion variable evaluates the level of effort exerted to search for and analyze new religious/theological knowledge. Allport (1950), in defining doubt, remarked that it was the product of a collision—new evidence colliding with present-day belief. Clark (1958) mentioned “true believers [can] absorb new points of view into [their] truth system” (p. 247). Further, Batson et al. (1993) noted that religious maturity is evidenced by a “continual search for more light on religious questions” (p. 166).

Two predictive models for EXP were discovered. Between them, four different predictors emerged—one domain and three facets. For the former and speaking generally, theological exploration was projected by Openness (+O; Explorer). Thus, Protestant religious undergraduates exert a high degree of effort to investigate and analyze religious doctrine and information (EXP) when they have an openness to life experiences (+O) which is visible via college students’ active creativity, aesthetic alertness, intrapersonal awareness, and strong preferences for originality and independent thinking (Beck and Jessup 2004; McCrae and Costa 2010).

With the facets and more specifically, theological exploration was predicted by ideas (+o5), actions (+o4), and tendermindedness (+a6). These results reveal Protestant religious college students actively pursue and examine religious knowledge when they gravitate toward intellectually curious subjects (+o5), behaviorally chase novelty and variety (+o4) and possess a noticeable sensitivity to human needs and injustices (+a6) (Beck and Jessup 2004; McCrae and Costa 2010).

4.1.1. Implications

A few implications materialized with the two models. First, the percentages of explained variance were noticeably distinct (see Table 4). The three facet traits explained almost twice (15.6%) the amount of variance relative to the domain trait (8.2%). Second, the facets provided informational specificity as compared to the domain predictor (Saroglou and Muñoz-García 2008). The two Openness facets, +o4 (actions) and +o5 (ideas) clarify which of the features in the Openness domain are germane to exploration (EXP) as Entringer et al. (2021) expected in a religious cultural context.

Third, unexpectedly gender effects were not a factor in the regression analyses for Stage 1 which was also the outcome for all the doubt stages. This also occurred in the descriptive outcomes (i.e., the means) with the religious variables (see Section 3.2.2). Possibly, gender effects are not as pronounced with doubt phenomena as the effects are with other religious behavior (e.g., attendance to church) or these participants were socialized in environments where egalitarian parental roles were modeled (Adams 2020).

Fourth, there is little research using the Five Factor Model (FFM) with the MQOS that would afford comparisons with this study’s outcomes. Yet, the study by Puffer (2013) reported EXP being predicted by two interpersonal personality traits in the California Personality Inventory (CPI; Gough 1987). They were dominance (+DO; assertive social leadership) and empathy (+EM; a kind of social intelligence) (McAllister 1988). The present study corroborated the association of social sensitivity (+EM) and EXP by the emergence of a6 (tendermindedness). But it also expanded understanding of the personality predictors for EXP with important intrapersonal personality traits—+o4 (actions—a readiness to pursue novelty) and o5 (ideas—comfortability with unconventionality) which can be motivational forces for exploratory efforts.

Last, Entringer et al. (2021) mentioned four facets being predictive of religiosity—altruism (+a3), compliance (+a4), self-discipline (+c5), and traditional ideas (−o5) with a religious sample. Yet, the present study challenged their conclusion. It appears that highly religious college students can swim against the religious cultural tide and remain religiously committed. Theological exploration (EXP), a religious construct, was predicted by unconventional ideas (+o5), only one Agreeableness facet (a6—tendermindedness), and no Conscientiousness facets using a religious sample. Possibly, operationalization was the key distinction. The present study operationalized religiosity using a classic or traditional construct (i.e., religious motivation—quest religious orientation) instead of a generalized, author developed concept (i.e., do I see myself as a very religious person).

4.1.2. Applications

Practically speaking, there are several applications germane to these results (see Appendix B for a summary). Suggestions are focused on helping religious college students manage this first stage of doubt and equipping their mentors with guidance ideas. First is an important perspective. The modern American church needs undergraduates’ creativity, fresh ideas, and new solutions. The three Openness facets and domain trait elicit the tendency for divergent thinking allowing students to think outside of the box (McCrae and Costa 2010). The statistics on how many church members have left mainline Protestant denominations (Burge 2023; Kinnaman 2011) and how many Americans are labeling themselves “nones” (Smith 2023) are difficult to ignore; these and other trends point to the pressing need for new solutions. The imaginative proclivities of Openness (+O), actions (+o4), and ideas (+o5) can translate exploratory outcomes into new and efficient church-life developments. For instance, new ideas are needed for how the church can relate and minister to disenfranchised groups (e.g., immigrants, the queer community) or how to tweak church practices such as the worship service, offering, baptism, and communion. Further, mentors of religious undergraduates can remind students to address pragmatic and feasible features (e.g., finances, recruitment of volunteers) during the generation of new ideas and solutions and mentors can also support the real-life implementation of the undergraduates’ ideas/solutions.

Second is empowerment for the exploratory journey. Collectively, the three obtained facet predictors can aptly energize religious undergraduates for the exploration of religious knowledge. Every day, humans think, act, and feel—basic functions for survival; each facet corresponds to one of the functions. Ideas (+o5) animate cognitive functioning; it elicits the tendency to be intellectually curious. People with this personality trait enjoy debates, brain teasing puzzles or riddles and willingly pursue idiosyncratic ideas. Further, they tend to be perceptive, innovative, and visionary (McCrae and Costa 2010). Actions (+o4) rouse behavioral functioning; it elicits the tendency to behaviorally engage in novel games, locations, and cuisines. Persons with this trait are adventurous, prefer the unfamiliar, and eschew routines (McCrae and Costa 2010). Tendermindedness (+a6) evokes emotive functioning; it elicits the tendency to be sympathetic and empathic. Individuals with this personality trait excel at being interpersonally attentive; they exude warmth, kindness, and gentleness along with being protective of social injustices (McCrae and Costa 2010). Together, these tendencies can empower religious college students to help the American Protestant church address other important concerns such as diversity, biblical literacy, congruent discipleship, commitment avoidance, autonomy idolization, etc. (Hilder 2024; McCracken 2016). Further, mentors of religious students can educate mentees on how to navigate the local church’s political structure in bringing the aforementioned concerns to the attention of the church community.

Third, there are a few cautions when undergraduates explore religious information. These include the need to: (1) prepare for misunderstanding. Exploration is inherently challenging the status quo—present day doctrine and church practices. Some church leaders may interpret the exploratory pursuits as a personal affront to them or to mean all presently held beliefs and practices must be forsaken and completely replaced; (2) explore new religious information with trusted friends, avoid going alone. Bounce ideas off several safe, wise, and responsible people—peers and mentors; (3) depend on multiple sources of information, not just one. Proverbs 11:14 notes how there is safety in relying on multiple wise counselors (e.g., authors); (4) develop the habit of double-checking resources—avoid taking one opinion at face-value. Analyze the new information by cross-checking with several other authors; and (5) ready the heart for possible condemnation. Nay-sayers and status quo protectors will challenge anything different, bristle at questions and scoff at the mention of the need for improvements. Such narrow-minded religious congregants and threatened church leaders will work hard at diminishing challengers’ optimism, adventurousness (+04), idealism (+o5), friendliness and gentleness (+n6). Therefore, it would be prudent for undergraduates and their spiritual advisers to anticipate disputes, prepare positive, respectful retorts (e.g., I understand your position; here is where we disagree), and remain resolute to move forward.

4.2. Stage 2—Predictors for the Cognitive/Emotional Phase in the Doubt Process

Based on the model by Krause and Ellison (2009), one or a few precipitants initiate Stage 2, the cognitive/emotional phase (Wallace 2021). Relative to the other three, this stage has garnered the most attention from authors and researchers for well over a half of a century. Allport (1950) considered doubt as an “unstable or hesitant reaction” (p. 100). In a state of uncertainty, religious persons are tentative and indecisive while questioning and self-disputing their religious beliefs (Hunsberger et al. 1993; Puffer 2017).

For this study, religious doubt was operationalized with two constructs from the MQOS—existential motives (EXM) and complexity (COM). The former criterion variable measures the amount of concern for existential matters (e.g., identity, freedom, death) and meaning making (Beck and Jessup 2004). Batson et al. (1993) described the quest motive as dealing “openly and honestly with matters central to all existence” (p. 160). Mature religious individuals pursue the meaning and purpose in life and reflect on their tension with current world problems with religious questions.

COM assesses the level of importance and preference for complex/philosophical perspectives on religious topics instead of naïve ones. Batson et al. (1993) discussed Allport’s view on religious maturity being evidenced when people face their “complex problems [in life (e.g., personal calling, relational breakup, death of a loved one)] … without reducing their complexity” (p. 166). Further, among the COM items in the MQOS, participants are asked if “religious questions involve complex answers” (Beck and Jessup 2004, p. 292).

Two predictive models surfaced per EXM and COM. Five different predictors materialized for EXM while six emerged for COM. For the former and speaking generally, existential motives (EXM) were projected by two domains, Neuroticism (+N; Reactive) and Conscientiousness (−C; Flexible) (Howard and Howard 2010). Religious undergraduates inquire about existential concerns when they have the proclivity to be bothered by a greater number of environmental provocations triggering a variety of negative emotions (+N) and when they possess a diffused focus related to juggling several goals and being spontaneous and easily distracted (−C) (Beck and Jessup 2004; Howard and Howard 1995).

With the three facets and more specifically, existential motives (EXM) were predicted by depression (+n3), social aloofness (−e1), and values (+o6). Religious individuals are also energized to investigate existential and purpose-type of questions when they possess penchants for experiencing depressive affect including guilt and sadness (+n3), for being socially distant, formal, and uncompassionate (−e1), and for lacking dogmatism (+o6)—a willingness to re-assess religious, social, and political views or values (Beck and Jessup 2004; McCrae and Costa 2010).

Three domain traits predicted complexity—Openness (+O; Explorer), Neuroticism (+N; Reactive), and Extraversion (−E; Introvert). Generally speaking, religious students are drawn toward complex religious issues and choose to face them head-on (COM), when they tend to embrace divergent thinking along with an active imagination, aesthetic discernment, and intrapersonal sensitivity (+O), when they have the tendency to be sensitive to environment stimuli generating several negative affect (+N), and when they prefer an introverted social posture with skeptical, serious, and quiet proclivities (−E) (Beck and Jessup 2004; Howard and Howard 2010; McCrae and Costa 2010).

Last, three facet traits also projected complexity—ideas (+o5), non-assertiveness (−e3), and values (+o6). More specifically, religious undergraduates possess a high preference for complex religious thinking when they have an open-minded, intellectual curiosity (+o5), when they tend to be shy, reserved, and non-assertive (−e3), and when they possess a willingness to re-examine personal values such as religious convictions and political views (+o6).

4.2.1. Implications

A few implications surfaced with the four models. First, there were important distinctions among the percentages of explained variance (EV) (see Table 4). In general, facets explained more variance than domain traits as expected by Entringer et al. (2021) in a highly religious sample. Comparisons within each criterion variable reveal that the obtained facets for EXM (10.1%) explained slightly more variance than the emergent domain traits (8.1%) while the obtained facets for COM facets (21.7%) explained notably more than the emergent domains (14.6%). Further, comparisons between the two criterion variables reveal facets and domain traits for COM explaining more variance than EXM; this was an unexpected outcome. Among the facets, COM’s EV was twice (21.7%) the size of EXM’s (10.1%). With domain traits, COM’s EV was almost twice the size (14.6%) of EXM’s EV (8.1%). Apparently, there must be unaccounted predictors for EXM. The authors assumed more similarity between them; both measures represent key features in the cognitive/emotional state of doubt (e.g., mulling over one’s personal purpose in life, re-investigating doctrine, wrestling with complex answers over naïve ones).

Second, comparisons of this study’s religion-personality findings with investigations in the literature are limited. Yet, Puffer (2017) reported that existential motives (EXM) were predicted by two social personality traits in the California Personality Inventory (CPI)—dependence (−IN; clinging, strong reliance on others) and intolerance (−TO; pushing others away with suspicion and fault-finding) (Gough 1987; McAllister 1988). These variables explained 8% of the variance. Puffer (2017) also found complexity (COM) being projected by non-conformity (−GI; a disregard for making a good impression on others) (McAllister 1988). This one variable explained 2% of the variance. When comparing explained variance (EV) between the Five Factor Model (FFM) in the present study and the CPI predictors, the former offers slightly more EV for EXM (10.1% vs. 8%) while there is a considerable gap between the two EVS with COM (21.3% vs. 2%).

Third, a potential purpose for the odd set of domain predictors (i.e., Neuroticism [+N; Reactive] and Conscientiousness [−C; Flexible]) of EXM bear similarities with the possible rationale for the odd pairing of predictors (dependence [−IN] and intolerance [−TO]) of EXM from the CPI in the study by Puffer (2017). Although the two FFM domain predictors are more intrapersonal and the CPI ones are interpersonal in nature, the relationship the four have with EXM can be understood by the same issue—the context. In the milieu of Stage 2, religious undergraduates are questioning and challenging religious beliefs and church practices; discontentment and dissatisfaction potentially churn inside them. If concerns are met with status quo answers, frustration can multiply. In like fashion, undergraduates’ protests with belief/church practices might become an annoyance to the local church’s leadership who are the entrusted protectors of the status quo. College students, as reactives, may bristle at criticism to their concerns and possible solutions and become moody. But, as flexibles, they also have the potential to respond to negative feedback in a more relaxed manner (McCrae and Costa 2010). This edginess (i.e., Reactive) can be balanced with calmness (i.e., Flexible) just as the pushing away habits (i.e., −TO or intolerance) can moderate the clinging tendency (i.e., −IN or dependence) (Howard and Howard 2010; McAllister 1988).

4.2.2. Applications

There are a few real-world applications from these findings (see Appendix B for a summary); each suggestion seeks to expand religious undergraduates’ management strategy at this point in the process of doubt (Stage 2) and to equip their mentors with guidance ideas. First, with the obtained traits for EXM, a few ideas surfaced. The three facet predictors serve undergraduates’ efforts to pursue existential matters and meaning making questions. Values (+o6) elicit an openness to self-examine their values which can reduce dogmatic reactions to their current religious beliefs and practices. The trait infuses a courageousness to consider conventional and unconventional topics (McCrae and Costa 2010). But during re-examination, it is also important to lean on safe friends and mentors to serve as a sounding board for current musing. They need to reflect in community, not in isolation. Further, these musings need to be checked with respected biblical and theological resources—books, articles, blogs, podcasts. Double-checking with trustworthy references can hone the thought process and can prevent unhelpful conclusions. Spiritual advisors can assist in the discovery of reliable and responsible sources of information and the re-checking process.

Depression (+n3) along with Neuroticism (Reactive; +N) evokes a sensitivity to environmental factors leading to negative emotional reactions. People are prone to feel anger, embarrassment, etc. Yet, the sensitive awareness of circumstances and conditions can serve another purpose; it can be a means to alert a church community to possible unhealthy, unbiblical, and unjust church doctrines and practices. Without parishioners’ anger and disgust, exposure of possible injustices (e.g., abusive behaviors) within the church community may be delayed.

It is important to note that the experience of depressive affect (e.g., worry, pessimism, moodiness) is uncomfortable and exhausting. Yet, there are ways to manage it. Students with the assistance of their mentors can become ‘emotionally savvy’; they can learn to understand about the purposes for common negative affect (e.g., disgust, sadness, fear, etc.), distinguish between healthy and unhealthy forms (e.g., good guilt versus destructive false guilt), and discover effective means to manage the negative affect (e.g., having constructive reactions with anger). Further, a licensed therapist can assist in this discovery while spiritual advisers can assist students in the rehearsal of their emotional knowledge.

Reserved (−e1) reveals a social style of undergraduates examining existential matters and challenging church doctrine and practices. According to McCrae and Costa (2010), people who scored low in warmth (i.e., −e1) are not “hostile or lacking compassion;” [they] are “formal, reserved, and distant in manner” (p. 22). This social presentation makes sense; students may want to keep their doubtful musings more private than public, a natural reaction when swimming across the cultural tide (i.e., norms of the church). Reservedness can also help in the resistance to social pressure to conform to the status quo. Undergraduates’ reticence can create enough social distance between church leaders and themselves to avoid sensing an obligation to follow suit or conform to leadership’s preferences/expectations. Further, although students’ social reservation is apparent, their opinions and perspectives still need to be released to someone—somehow and someway; their thoughts are important to the health of the church community. Possibly, their mentors can help generate alternative dissemination practices.

Neuroticism (Reactive; +N) and Conscientiousness (Flexible; −C), as previously mentioned, appear at first glance to be an odd pair of predictors for EXM. Yet, the twosome can assist in the efforts related to this criterion variable. For instance, religious students are questioning theological tenets, a task that can be daunting and vigorous. Reflections on the meaning of their life and place in the world are often messy. Responses to heart breaking news and tragedies can be tiring. Challenges to church doctrine and practices can include fault finding and potentially elicit defensiveness in church leaders. So, the kind of responses which reactives often have to environmental stimuli bring a keen level of sensitivity, alertness, concern, and expression. These can be unsettling to some people around them because reactives often appear stressed, restless, and worried (Howard and Howard 2010). Yet, the domain predictors, Reactive being paired with Flexible, can translate into Flexible tendencies moderating the common social presentation of reactives. For instance, flexibles tend to be calm instead of being restless and more relaxed instead of being stressed; they also tend to take things less personally (Howard and Howard 2010).

Second, with the emergent facet predictors of COM, a few specific applications also materialized. The three facet traits—ideas (+o5), values (+o6), and non-assertive (−e3), serve the preference for non-simple philosophical religious views. Ideas (+o5) elicit an intellectual curiosity and openness to unconventional topics along with an idealistic inventiveness and imagination providing a strong draw toward complexity (McCrae and Costa 2010). But this openness needs to be accompanied by intellectual accountability. This answerability is not a conformity to the status quo, instead it is a willingness to lean on one’s community such as bouncing ideas off safe peers and mentors along with checking with reliable, scholarly resources. It mimics the recommended social interaction of Proverbs 27:17—iron sharpening iron; people in these kinds of social encounters help one another by exposing blind spots and illogical conclusions (Lennox 1998).

Values (+o6) evokes an inclination for self-examination, a non-dogmatic openness to reassess personally established religious, social, and political values. Such a willingness creates an internal environment posed to embrace more complex religious ideas. Instead of being closed off, following current traditions, and rushing to make overly simple conclusions, these undergraduates grant themselves freedom to learn more to strengthen their value-positions or reassemble them (McCrae and Costa 2010). Yet, as one learns more, it is recommended to consume complex information pertaining to both sides of any issue. A balanced approach is optimal. For example, the ordination of women has advocates and critics. Understanding both views provides for increased complexity (i.e., avoiding narrow-mindedness) and the opportunity to strengthen or revise a value-laden stance on the topic. Further, mentors can prompt students toward equitable research and synthetic summaries on issues.

Non-assertiveness (−e3) prompts the tendency to be slow to speak while eschewing social forcefulness and domination. Persons with this trait relish the social background letting others to be the first to express ideas and opinion (McCrae and Costa 2010). Such a proclivity provides the needed patience required for preferring complexity over simplicity. The former entails more time, work, and information than the latter. However, this shyness does not exempt undergraduates from sharing their wisdom; it is not a pass for non-involvement. Instead, the preference for complex religious views necessitates wisdom garnered from all participants. This may compel students to share their ideas and knowledge with a more talkative peer or mentor who can join them in disclosing their wisdom.

Concerning the three domain traits, Openness (Explorer; +O), Neuroticism (Reactive; +N), and Extraversion (Introvert; −E), a few applications surfaced with COM (Howard and Howard 2010). In discussions of conventional and unconventional religious topics, there can be social pressure to accept simple, uninformed answers because these are less controversial. Pushing beyond the superficial and seeking an in-depth engagement are necessary for progress, but the amount of effort exerted to accomplish these can become exhausting. The Explorer (+O) evokes an imaginative and intellectual curiosity prone to seek variety not routine, independent decision-making not conformity (Howard and Howard 2010). Hence, this kind of divergent cognitive processing comports with and can energize a college student’s drive toward complexity. Yet, it is vital that undergraduates generate ideas which benefit themselves as well as the whole religious community, no matter the age. Creativity needs to serve all people, not just one generation; mentors can advocate for age impartiality.

The Reactive (+N) elicits a heightened sensitivity to environmental stimuli (e.g., criticism, being ignored by adult leaders) and a proneness to experience negative affect (Howard and Howard 2010). This alertness, concern, and excitability provides the needed energy to resist social pressure and push discussions away from triteness and over-simplification. Yet, criticism will pain the Reactive side of these college students; hence they will need to prepare themselves with the assistance of the spiritual advisors for opposition and choose to lean on and appreciate the optimism of the Explorer (+O) side.

The Introvert (−E) prompts an inward pull toward self that can aid in the processing of information. The tendencies of introverts toward seriousness, quietness, and skepticism can assist in the pursuit of appropriate and helpful resources along with the meticulous consideration and comprehension of complex information (Howard and Howard 2010). Yet, it is important that undergraduates do not restrict the topics being considered, a practice which mentors can speak directly into. Biased selection of subjects encourages narrow-mindedness which can stymie personal cognitive/emotional growth. For example, research on a denomination’s progress with diversity needs to include and address its historical endorsement of racial terrorism.

4.3. Stage 3—Predictors for the Coping Response Phase in the Doubt Process

Krause and Ellison (2009) mentioned two after-effects from Stage 2. The first is a coping response—an initial reaction to doubtfulness often within a short timeframe. Religious people attempt to successfully manage uncomfortable demands (e.g., doubts) that are taxing their resources (Lazarus and Folkman 1985). Two coping responses, elimination (i.e., suppressing doubts) and resolution (i.e., finding spiritual resources/support), were investigated by Krause and Ellison (2009).

Stage 3 was operationalized in this study with the construct, change (CHAN) of the MQOS (Beck and Jessup 2004). Items in the subscale include “I anticipate dramatically changing my religious beliefs in the future” and “I believe that changing one’s religious beliefs is a good sign of spiritual development” (p. 292). Snowden (1910) likened religious doubt as a ruiner of error, leading to the removal of threatening beliefs, a form of change. Further, although change appears to be a common experience, it is not a homogenous construct. One size does not fit everyone. Instead, it is a complex process including phases (i.e., steps) and different classifications (Bachmann 2024). For instance, change might involve a ‘complete overhaul’—the original identity is unrecognizable; someone switches from Calvinism to Arminianism. It can involve minor revisions/tweaks/substitutions; an undergraduate adds a Bible study meeting to her schedule. Change can also be a new understanding of a doctrine or practice; a college student has a new perspective on the Trinity.

Two predictive models emerged for CHAN. Overall, six different predictors materialized between them. In the domain model, there were Neuroticism (+N; Reactive), Openness (+O; Explorer), and Agreeableness (−A; Challenger) and in the facet model, there were ideas (+o5), values (+o6), and vulnerable (+n6). For the former and speaking generally, religious undergraduates are open to altering their convictions and to keep scrutinizing them when they have the tendency to react to several environmental stimuli eliciting negative emotions (+N), when they have the proclivity for divergent thinking leading to curious and introspective reflections, an active imagination, openness to change, and an intrapersonal sensitivity (+O), and when they have the propensity to be suspicious of authorities, disagreeable, skeptical, defended, driven, and argumentative (−A) (Beck and Jessup 2004; Howard and Howard 1995, 2010).

For the three facets and more specifically, religious students possess an openness to vary their religious tenets and practices when they have the tendencies to actively pursue intellectual topics accompanied with an openness to reflect on unconventional subjects (+o5), to lack a dogmatic stance on their values and are ready to reexamine them (+o6), and to possess a state of vulnerability—struggling to manage stress and becoming dependent and possibly hopeless under certain situations (i.e., an emergency) (+n6) (Beck and Jessup 2004; Howard and Howard 2010; McCrae and Costa 2010).

4.3.1. Implications

A few implications surfaced with the two models. First, there was little variation between the percentages of explained variance (EV) of the two models (see Table 4). The obtained facets for CHAN (16.9%) explained slightly more variance than the emergent domain traits (15.6%) as expected by Entringer et al. (2021) in a highly religious sample. Yet, the EVs of both models are twice the size of the EV (8%) mentioned by Henningsgaard and Arnau (2008). The type of QRO measure may explain the difference. These researchers employed a unidimensional model for quest orientation while a multidimensional model was utilized in this study.

Second, as mentioned before, there are limitations when comparing this study’s findings with other research projects. However, Puffer (2017) mentioned that CHAN was projected by one social personality trait in the California Personality Inventory (CPI)—non-conformity (−GI) (Gough 1987). This CPI variable explained 4% of the variance which is one-fourth the EV of either the facet or domain predictors of CHAN in this study.

However, there are several unique similarities between non-conformity (−GI) of the CPI and Challenger (−A) of the FFM. The former tends to be sarcastic, critical, rebellious, mutable, and blunt (McAllister 1988). Presenting well, being cooperative and polite are not high priorities. In like fashion, challengers are known for rudeness, hostility, aloofness, arrogance, and combatant; they are also skeptical and competitive along with being suspicious and guarded with authority figures (Howard and Howard 2010). Certainly, both predictors have less than desirable social presentations; but features in both traits can assist undergraduates in a change experience. For example, in the modification dimension of CHAN; mutability, bluntness, and non-cooperation (FFM), aloofness, suspicious, and guardedness (CPI) can reduce pressure to conform to the status quo or to foreclose with a decision to please someone else instead of freely exploring options and committing without compulsion. In addition, frequent scrutiny of religious beliefs in CHAN is bolstered with habits such as criticalness (FFM) and skepticism (CPI) which can supply motivation for analyses of doctrine and church practices (Puffer 2017).

4.3.2. Applications

Pragmatically speaking, there are few applications from these results (See Appendix B for a summary); each one attempts to assist religious undergraduates’ management of Stage 3 and to equip their mentors with guidance ideas. First are applicative ideas derived from the domains. Neuroticism (+N; Reactive) elicits a sensitivity to environmental factors that include experience with several negative emotions. Often, reactives are perceived by others as being stressed, tense, temperamental, and prone to take things personally (McCrae and Costa 2010). Yet, this sensitivity can be a helpful adjunct to a change process; it can act like a conscience monitoring ethical behavior and alerting danger. Such discernment enables college students to be quick to notice problems in doctrine, issues with church practices, and injustices taking place in the church community.

Further, reactives can learn how to manage their edginess and habit of personalizing feedback or criticism. They can discover the thinking errors attached to fearfulness and find ways to re-structure the irrational beliefs into adaptable rehearsal messages that can bring realistic and long-lasting relief (Corey 2024). In addition, they can work to understand and manage fear in a constructive manner. Fear is considered a response to a threat (Smith and Ellsworth 1985). Understanding the threat and creating an alternative response when facing the threat are a few of the tasks in addressing this powerful negative emotion (Greenberg 2011). Again, a license therapist can assist in these discoveries and mentors can encourage students to become emotionally savvy.

Openness (+O; Explorer) evokes a propensity for imagination, creativity, and divergent thinking. Explorers are drawn to a wide variety of issues and embrace change. Further, they tend to be reflective and introspective, leading them to plan strategically. Some pertinent questions related to the change process and needing consideration include: (1) what aspect of faith needs changing; (2) how will that aspect be altered? Will there be an elimination or tweaking; (3) will the change require several steps, or will it entail a quantum leap; and (4) what resources will be needed to bring about the transformation? Hence, the proclivities of explorers are welcomed accessories to the change process. It is like a dreamer proposing what can be different, itemizing needed steps to achieve change, and sketching out how the end goal would look like. Further, spiritual advisors can operate as a sounding board for processing possible answers to the aforementioned questions.

Agreeableness (−A; Challenger) prompts the tendencies for being stubborn, tenacious, vocal, defended, and driven. Being power based and suspicious of authority, challengers appear to their peers as disrespectful, self-focused, detached, skeptical, and disagreeable (Howard and Howard 1995, 2010). However, this hard-headedness and persistence can assist in the change process of religious undergraduates. These features are like having an enforcer who works to make sure change, at any level, happens and empowering undergraduates when encountering criticism from church leadership about their change proposal and agenda.

Further, challengers can also develop a more approachable social presentation with the assistance of their mentors. They can learn to confront people respectfully, genuinely listen to opposing views, and apply empathy when appropriate. Instead of their typical approach being a part of the problem, a kinder and gentler side can energize a solution.

Second are suggestions with the facet predictors. Ideas (+o5) elicit an intellectual curiosity—a motive for learning new things, imagination, inventiveness, and unconventionality. It prompts a growth mindset empowering people to see the unseen—beyond the status quo and not be restrictive with possibilities (Howard and Howard 2010; McCrae and Costa 2010). Such creativity is a helpful adjunct to the change process. Whatever aspect of faith is being altered, this ingenuity can help people tweak a part of the aspect or completely replace the aspect with something new and different. Yet, when religious undergraduates set out to make alterations, collaboration is recommended. They can collaborate with trusted peers and their mentors about the proposed transformation or replacement; Proverbs 15:22 notes ‘many advisors bring success’ (NLT). Ideating alone is not recommended; it would be prudent to listen to feedback from friends and mentors and even consider multiple friends’ (e.g., older ones) ideas on possible ‘change-proposals.’

Values (+o6) evoke an accessible re-evaluation. It is an openness to reassess values; a possible reforging or expansion of principals is being invited. This non-dogmatism means undergraduates have the freedom to judge their standards and alter them; they are not closed off (Howard and Howard 2010). Hence, whatever feature of faith is possibly being transformed, there is liberty or permission to take modifying actions. However, it is suggested they make sure the proposed change adds to their overall well-being, something a mentor can assist with. Changes need to be thoughtful improvements, not change just for change’s sake.

Vulnerability (+n6) prompts an anxiousness, a fear about being assailed including self-perceptions of not being able to handle stress leading to despair and impairment of coping skills (McCrae and Costa 2010). This neurotic feature can complicate a change process (e.g., slow it down). Yet this trait also brings a sense of realism to the change process. Reformation of religious beliefs and practices, at any level, is not easy and can be perceived as scary. An openness to change inherently means religious dissatisfaction exists; this angst may be difficult to express publicly with church friends and leaders who, in turn, might convey disappointment about the decision. For instance, the choice to switch churches may, when announced, frustrate those friends being left behind.

Furthermore, the anxiety in vulnerability (+n6) is manageable. Often, as mentioned previously, thinking errors are associated with it. Identifying and re-structuring irrational beliefs related to the aspect of faith being changed can help reduce the influence of the mental snares (Corey 2024). In addition, emotions provide a source of information that can aid in decision-making (Greenberg 2011). For instance, fear is considered a response to a threat leading persons to withdraw (Smith and Ellsworth 1985). Helpful questions to ask include: What is the danger connected to the anxiety? How might the threat be appropriately responded to? The ensuing answers can also be shared with a spiritual advisor. In addition, it would also be prudent to disclose feelings and thoughts related to being vulnerable to a trusted friend and mentor. Bottling up the powerful negative affect and thoughts will not help in coping (Krause and Ellison 2009); self-disclosing with community associates has the potential of being cathartic and relieving (Corey 2024).

4.4. Stage 4—Predictors for the Outcome Phase in the Doubt Process

Stage 4 is the outcome phase, the second after-effect in the model of the doubt process by Krause and Ellison (2009). Unlike the coping stage, end results are usually not immediate; these emerge over a more gradual time interval after doubtfulness in Stage 2. Outcomes are also dependent on the kind of coping responses employed in Stage 3. For instance, Krause and Ellison (2009) reported participants’ coping of doubt via suppression related to poor health scores.

In this study, Stage 4 was operationalized by two doubt constructs, tentativeness (TEN) and religious angst (RA) of the MQOS (Beck and Jessup 2004). These subscales measure two possible and different outcomes with doubtfulness. TEN assesses the level of doubt positivism and comfortability with religious questions remaining unanswered. RA resembles the negative side of doubt by evaluating the level of experienced negative affect (e.g., loneliness, worry, abandonment, confusion) along with doubt in persons’ spiritual pilgrimage (Beck and Jessup 2004).

4.4.1. Tentativeness/Doubt Positivism

Some of the items in tentativeness, TEN, portray the essence of this measure of religious doubt. These include: “I believe religious doubts play an important role in spiritual development”, “I believe a central part of spiritual maturity is growing comfortable with doubt”, and “I don’t feel a need to know all the answers to every religious question I may have” (Beck and Jessup 2004, p. 292). This subscale of the MQOS resonates with the messages of pro-doubt authors. For instance, doubt is perceived as an expected, normative, and unavoidable phenomenon (Ponce 2022; Wallace 2021). It is knitted into the fabric of human nature and operates as a sign of maturity and a vital piece in faith development (Halfaer 1972; P. Long 2023; Moldovan 2022).

Further, there is a kinship between doubt positivism and honest/believing doubt; both include the element of stability. The former exhibits a level of comfortability, a relaxed calmness, even when religious questions remain unanswered. The latter remains anchored to God during the stages of doubt. For example, during the processes of discovering an error (e.g., a misunderstanding, heresy) leading to a deconstruction (e.g., jettisoning a poor interpretation of a doctrine) followed by a reconstruction (e.g., increased clarity on a doctrine; a more robust faith), religious persons continue being attached and loyal to God (Osborne 2011; Piper 2020; Snowden 1910; Wallace 2021).

Two predictive models surfaced for TEN. Collectively, four different predictors emerged between them: in the domain model, there was Openness (+O; Explorer) and in the facet model, there were easy-going (−n2), values (+o6), and ideas (+o5). For the former and speaking generally, religious college students view doubt positively and can value religious questions over decisive answers when they have the tendencies for intellectual curiosity, divergent thinking, preference for variety, sensitivity to intrapersonal affect, and an active imagination (+O) (Beck and Jessup 2004; Howard and Howard 1995, 2010). For the facets and speaking specifically, religious undergraduates embrace doubt positivism emphasizing questions over conclusive answers when they have the proclivities to be easy-going or slow-to-anger (−n2), to be non-dogmatic about their values—a willingness to re-assess them (+o6), and to be intellectually curious and open to consider unconventional topics (+o5) (Beck and Jessup 2004; Howard and Howard 2010; McCrae and Costa 2010).

• Implications

There were a few implications regarding the models for the domain and facet traits. First, there was little variation between the percentages of explained variance (EV) of the two models (see Table 4). Emergent facets for TEN explained slightly more variance (9.9%) than the obtained domain trait (6.3%) as expected by Entringer et al. (2021). Second, as previously discussed, there are few studies to compare with this investigation’s results. Yet, Puffer (2017) did report that TEN was originally predicted by the empathy trait in the California Personality Inventory (Gough 1987). But in a subsequent analysis, a second standard multiple regression with only empathy, the p-value outcome was non-significant.

• Applications

There are a few real-world applications from these findings (see Appendix B for a summary); each seeks to strengthen religious college students’ management of Stage 4—doubt positivism and to equip their mentors with guidance ideas. First are the application suggestions from the domain trait, Openness (+O; Explorer) and its notable features. This personality characteristic elicits an intellectual curiosity with an active imagination and broad interests. Collectively, these can facilitate investigations of religious doctrines and practices; the drive to learn and probe beyond the surface can help expose perceived errors or misunderstandings. However, investigations fare better by examining multiple sides to any perceived error to achieve a balanced perspective. Students’ mentors can aid in the obtainment of helpful resources. Explorers are also divergent thinkers along with being relaxed with theory; they can aptly handle the unknown, intangible, immaterial, and abstract details of doctrines and practices. Yet, it is important not to get stuck in endless reflections which usually never offer meaningful insight or direction. A peer or a spiritual advisor can also encourage limits on excessive reflection.

In addition, explorers are strategic, able to project long-term goals and obtain the means to achieve them. This tendency can aid in the deconstruction and reconstruction of doctrines/church practices. Yet, some questions and doubts may never become resolved. Further, explorers tend to exercise independent judgment; this ability to make choices free of supervision can ensure that religious students do not succumb to social pressures and conform to the expectations of peers and church leaders. But this helpful determination will need some restraint to refrain from stubbornness and unteachability; a mentor can assist in encouraging honest self-reflection in these concerns (Howard and Howard 2010).

Second are the applicative recommendations from the three facets. Easy-going or slow to anger (−n2) elicits an even-tempered approach to life. People with this characteristic tend not to be anxious, moody, or easily irritated (McCrae and Costa 2010). The trait brings an internal calmness when college students are in a state of uncertainty. It can help reduce distracting “psychological noises” in the head (i.e., worry, obsessive thinking, shame) giving room for creative solutions to arise in the process of doubt. However, it is important not to overly romanticize or minimize religious doubt (Moldovan 2022). Spiritual advisors can remind undergraduates about doubt being a resultant from this fallen world and how doubt can quickly go awry turning into a painful and negative struggle and possibly into unbelief (Wallace 2021).

Values (+o6) evoke a non-dogmatic willingness to assess personal value systems (McCrae and Costa 2010). Such broad-mindedness and tolerance can limit rigid demandingness for certainty and leave room for authentic doubts without humiliation (Moldovan 2022). Yet, re-examinations need multiple sources of information that provide both support and dissent of a topic to allow for cross-checking of ideas and alternatives. Further, it is prudent to lean on peers and mentors in the community for feedback (Proverbs 15:22); communal dependence can help avoid narrow-mindedness.

Ideas (+o5) prompt an active intellectual curiosity and imagination that is inventive and idealistic (McCrae and Costa 2010). Such characteristics are needed in grappling with questions, the unknown, uncertainty, and unconventional perspectives. For example, many Christians bristle at the notion of doubt positivism. This specific personality trait can incite undergraduates to search the biblical text for how God responded to the doubting proclivities of His people. Such an investigation can help them discover whether God was (is) doubt intolerant and condemning or more doubt tolerant (Yancey 2009, 2020).

Further, students’ or any religious persons’ findings can provide a workable model for them on how to perceive and respond to religious doubt. For example, several biblical stories offer insight into the intolerance versus tolerance debate. First, God patiently engaged prominent Old and New Testament characters (e.g., Abraham, Sarah, Gideon, Peter) when they doubted His promises (e.g., bearing a child) and clear directions (e.g., leading Israel into battle or walking on water). Second, God welcomed questions, often Christ’s comments aroused inquiries from His disciples. The apostle, Thomas, stated that the disciples didn’t know where Christ was going and asked how they can know the way (John 14:5). Third, God was not threatened by His followers’ doubts. In John 6:60; the disciples stated that eating Christ’s flesh and drinking His blood was a difficult statement to fathom and wondered out loud as to how anyone could believe the words. Fourth, God revealed a deep concern for the lives of His people which includes doubtfulness. Jesus shared in Matthew 11:28, “Come to Me, all who are weary and I will give you rest” (NLT). Fifth, God promised wisdom to His people to resolve issues in life such as doubts. The Proverb writer remarked in Proverbs 2:6, “The Lord grants wisdom, knowledge, and understanding” [NLT]. He also promised to faithfully remain present with His people, including doubters as they grapple with their unresolved questions and painful tragedies. In Hebrews 13:5, God declared, “I will never leave you, nor forsake you” (NLT; M. Long 2023).

4.4.2. Religious Angst/Negative Side of Doubt

Several of the items in religious angst, RA, reveal the gist of this measure of religious doubt. These include: “My religious development has often been filled with doubt and has been troubling at times”, “I would mostly describe my spiritual journey as a struggle”, “I have often felt lost and alone during my spiritual journey”, and “I have often felt abandoned by God during my spiritual journey” (Beck and Jessup 2004, p. 293). This MQOS subscale resembles many negativistic perspectives about doubt. For instance, religious doubt has been regarded as being uncomfortable (Piper 2020), scary (Patton 2010), unsafe (Moreland and Issler 2008), bad, frightening, and an enemy of the faith (P. Long 2023).

Further, there is a similarity between the negative side of doubt and dishonest/disbelieving doubt; both possess the element of instability or unsteadiness. The former eschews doubt; its alarming and dangerous nature is disquieting. The latter vacillates in trusting God and remaining attached to Him. Metaphorically speaking, religious people stand in two camps; one foot predominately leans on human wisdom and the other quasi-depends on God’s understanding (Osborne 2011; Piper 2020).

Two predictive models surfaced for RA. Collectively, six different predictors surfaced between them: in the domain model, there was Neuroticism (+N; Reactive) and in the facet model, there were depression (+n3), values (+o6), mistrust (−a1), reserved (−e1), and impulsive (−c6). For the former and speaking generally, religious college students feel isolated, anxious, doubtful, and other negative emotions in their spiritual pilgrimage when they are easily bothered by several environmental factors (+N) (Beck and Jessup 2004; Howard and Howard 1995). For the facets and speaking specifically, religious undergraduates experience a level of negative affect (e.g., loneliness, worry, abandonment, confusion) along with doubt in their spiritual pilgrimage when they have a tendency to be depressed (i.e., prone to sadness, guilt, etc.) (+n3), to be non-dogmatic and willing to review their value system (+o6), to be cynical and skeptical (−a1), to be more formal and reserved with people (−e1), and to be impulsive (−c6) (Beck and Jessup 2004; McCrae and Costa 2010).

• Implications

Concerning the models for the domain and facet traits of RA, there were a few implications. First, there was considerable variation between the percentages of explained variance (EV) (see Table 4). Obtained facets for RA explained 32.1% of the variance relative to the 22.3% of the emergent domain trait, an outcome predicted by Entringer et al. (2021) with a religious sample. Second, as mentioned earlier, there are few investigations in the religion-personality literature to compare with this study’s findings. However, research using RA provides pertinent information about the subscale—problematic associations with adaptive religious and psychological constructs. For instance, Beck (2006) indicated RA was positively correlated with disturbed attachment styles with God—anxiety about abandonment and avoidance of intimacy with God. Messay (2010) found RA negatively associated with forgiveness and positively related to psychological distress. Further, Puffer et al. (2023) reported RA negatively related to important career developmental variables—vocational identity and career commitment.

• Applications

Practically speaking, there are a few applications germane to these findings (see Appendix B for a summary); each attempt to assist religious college students in the management of this outcome in Stage 4—religious angst and to equip their mentors with guidance ideas. First are the applicative ideas from the domain trait, Neuroticism (−N; Reactive). This personality feature provokes a discriminating responsiveness to environmental factors leading to the experience of several negative affect (e.g., fear, anger, guilt, etc.) (McCrae and Costa 2010). Reactives come across to peers as tense, restless, moody, discouraged along with taking challenges personally. Yet, their tendencies to be alert, sensitive, and concerned can be marshalled toward understanding and managing their powerful negative feelings and habits. Alertness and sensitivity can be the needed indicators (i.e., much like an amber traffic light) to re-direct reactives’ emotional functioning. For instance, this attentiveness can prompt them to pause, which can afford time to reflect on emotionally savvy questions—what are they afraid of (e.g., what is the threat to them?) and how can they constructively respond to the fearfulness? What is an alternative response to the threat, something other than avoiding it or being moody? Can they practice deep breathing and then consult with a supervisor (e.g., mentor) to create a solution?

Second are the application ideas from the five facets. Collectively, the traits aptly describe the intrapersonal and interpersonal world of religious undergraduates experiencing religious angst/the negative side of doubt. For instance, while these religious persons are open minded and non-dogmatic (+o6) in their doubting experience, they are also depressed, discouraged (+n3), and impulsive (−c6). With others, they are cynical, skeptical (−a1), closed off, and aloof (−e1).

Examining the fivesome more thoroughly, values (+o6), on the personal level, elicit a willingness to re-examine personal values including the tendencies for tolerance and broad-mindedness which are needed when questioning religious doctrine and practices. Depression (+n3) evokes a proneness to grapple with guilt, sadness, and hopelessness along with discouragement, discontentment, and non-confidence. Impulsive (−c6) prompts the tendency to not being deliberate—reacting with a sense of urgency and becoming reckless. In the social realm, mistrust (−a1) triggers a suspiciousness toward others fearing them to be deceitful and unsafe. Reserved (−e1) sparks a social aloofness that appears unfriendly, frosty, and detached (McCrae and Costa 2010).

The cause or causes for the religious undergraduates’ angst is (are) unknown. Possibilities include multiple trials (i.e., James 2:1–4) (e.g., cancer diagnosis, death of a sibling, etc.) for unknown reasons that have arrived in their life, one right after another. Another potential cause is a season referred to as a dark night of the soul (Psalm 23:4) where God leads Christians through a character transformation (see Appendix A). Often, this spiritual experience is accompanied with unanswered prayers, repeated disappointments, and self-perceptions of being a failure, defeated, and empty (Scazzero 2017).

Whatever the cause(s) for religious angst, there is hope. Several responses will aid in the constructive management of this outcome in Stage 4. Religious undergraduates can: (1) share doubtful thoughts with God and ask for wisdom in handling this season (M. Long 2023; Poor 2024). The psalm writers modeled candidness. In Psalm 22, David cried out questions—why am I forsaken and why is God so far away? In Chronicles 1:10, Solomon requested wisdom to lead Israel, a task unfamiliar to him; (2) choose not to suppress thoughts and feelings. Mentors can function as a sounding board. Krause and Ellison (2009) reported that suppression of doubts was considered a poor coping response and related to health problems; (3) find a ‘religious angst companion’ (Yancey 2020). Whatever questions or doubts that emerge in this season, college students need to make a conscious effort to confide in a trusted friend (Wallace 2021) or mentor. Loneliness can fuel unrealistic thinking. Some leave the Christian faith because they do not have ‘safe’ people as confidants (Moldovan 2022); (4) refrain from making major changes in lifestyle that are sudden and overly reactive. Instead, they need to focus on working through their season of religious angst and not impulsively altering their religious lifestyle (e.g., stop attending church, ghosting close friends); (5) recognize depression impacts many Christians—young and old. Martin Luther delineated in his journal the common features of his depression—loneliness, self-reliance, and divine thinking errors (e.g., God is not good) (Griffith 2017); and (6) decide to attend to depressive symptoms and impulsive tendencies. There are no quick fixes for either one. Yet, seeking help from a therapist can aid the journey. Counselors can assist in the discovery of irrational beliefs contributing to depression and impulsiveness; these mental health professionals can also help with understanding and managing the emotions (i.e., fear, guilt, sadness) related to both. Further, a medical doctor can potentially prescribe medicine that can raise mood levels and slow down impulsivity. A spiritual advisor could accompany students during the appointments to help them remember details.

4.5. Predictors for the Unidimensional Measure of Quest Religiosity (Batson’s Quest Scale)

Batson’s (1976) quest religious orientation and subsequent Quest Scale (QS) were a protest orientation and measurement. He argued, in the late 1970′s, that the operationalization of Allport’s intrinsic and extrinsic orientations was missing three features of the religious sentiment. The threesome included a responsiveness to existential angst and life circumstances (e.g., trials, tragedies), doubt positivism, and openness to new religious information and possible belief modification (Batson et al. 1993). Eventually, the trio was operationalized into a 12-item assessment instrument which over the years has been popular among researchers and frequently utilized in their research. Yet, the QS has received criticism for lacking adequate discriminative and predictive validity and for the ambiguity of the composite score (Beck et al. 2001; Leak 2011). Interestingly, most correlational data involving quest religiosity within the research literature includes a composite score (Crosby 2013). In this study, the total score of the QS was utilized for comparison reasons with the subscales in the multidimensional model of QRO (i.e., MQOS).

Two predictive models surfaced for QS. Collectively, seven different predictors emerged between them: in the domain model, there was Openness (+O; Explorer), Neuroticism (+N; Reactive), and Agreeableness (−A; Challenger) and in the facet model, there were values (+o6), mistrust (−a1), ideas (+o5), and actions (+o4) (Howard and Howard 1995, 2010). For the former and speaking generally, religious undergraduates have an inquisitive religious motive that embodies three important traits (i.e., responsiveness, doubt positivism, and openness) of mature and authentic religiousness when they have the propensity for an openness to change, divergent thinking, an active imagination, and an intrapersonal sensitivity (+O), when they have the tendency to react to environmental factors prompting negative affect (e.g., sadness, anger) (+N), and when they have the proclivity to be suspicious of authorities—being disagreeable and argumentative (−A) (Batson et al. 1993). For the facets and speaking specifically, religious college students have a quest religious orientation when they are not dogmatic—expressing a readiness to re-examine their value system—religious, social and political views (+o6), when they tend to be mistrusting, known for being cynical and skeptical (−a1), when they possess an active intellectual curiosity including progressive, non-conventional subjects (+o5), and when they are behaviorally inclined to seek out and experience novel kinds of foods, locations, and activities (+o4) (Batson et al. 1993; McCrae and Costa 2010).

4.5.1. Implications

There are a few implications concerning the outcomes from the domain and facet models. First, there was some variation between the explained variance (EV) percentages between the two models (see Table 4). The obtained facets for QS (19.5%) explained slightly more variance than the emergent domain traits (16.7%); this was anticipated by Entringer et al. (2021) for religious samples.

Further, comparing the EVs (i.e., facet and domains) for QS relative to the five Multidimensional Quest Orientation Scale (MQOS) subscales (i.e., existential motives [EXM], complexity [COM], tentativeness [TEN], exploration [EXP], and change [CHAN]) that directly measure QRO (i.e., excluding RA), QS had the highest EV percentage with domain traits relative to the EVs of the MQOS subscales with domain traits (e.g., 16.7% [QS] versus 15.6% for CHAN and 14.6% for COM, etc.) and was second among the facet traits (i.e., 19.5% [QS] vs. 21.7% for complexity [COM]). In addition, QS had the highest number of facet predictors (4) and tied with COM and CHAN for the highest number of domain predictors (3). Possibly, the composite score of QS is not as ambiguous as noted by Leak (2011) when it is studied with personality traits.

Second, the correlational results in present study support and contrast previous outcomes found in the literature (see Table 2). In this study, QS was positively related to Neuroticism and Openness (i.e., small coefficients, 0.10–0.30) while being unrelated to Extraversion. These findings corroborate the outcomes reported by Henningsgaard and Arnau (2008) and by Barrett and Roesch (2009). Yet, in this study, QS was negatively related to Agreeableness while Barrett and Roesch along with Kosek (1999) found QS positively associated. Further, in this study, QS was negatively related to Conscientiousness which Henningsgaard and Arnau (2008) also reported but Kosek (1999) indicated QS was positively correlated.

Third, for regression outcomes, Henningsgaard and Arnau (2008) discovered QS was positively predicted by Neuroticism and Openness and negatively projected by Conscientiousness explaining 8% of the variance. However, in the present study, QS was also positively predicted Neuroticism and Openness but negatively anticipated by Agreeableness explaining 16.7% of variance, twice the amount of Henningsgaard and Arnau (2008). Possibly, different instruments measuring QRO account for the different emergent predictors and explained variance percentages. The present study utilized Batson’s (1976) Quest Scale (i.e., 12 items) while the latter employed the Revised Religious Life Inventory (RLI; Hills et al. 2005) that had only 8 items for QRO.

Fourth, Puffer (2013) reported that QS was projected by three social personality traits from the California Personality Inventory (CPI)—empathy (+EM, an adept social understanding), intolerance (−GI, disregard for making a good impression), and feminine/masculinity (+F/M, an androgenous social thoughtfulness) (Gough 1987; McAllister 1988). Yet, the explained variance was about one-quarter of this study’s explained variance (5% vs. 19.6%). Possibly, different personality instruments account for the differences; the CPI has mainly interpersonal personality traits, while the FFM involves both inter- and intrapersonal features.

Last, when analyzing which facet and domain traits emerged for QS relative to the subscales of the MQOS, two observations are notable. The emergent domain features for QS were the same ones that surfaced for change, CHAN. Both had Openness, Neuroticism, and Agreeableness. The percentages of EV were similar as well; 16.7% for QS and 15.6% for CHAN.

Concerning the obtained facet traits in the QS, there seems to be a combination of key traits in exploration (EXP) and religious angst (RA). EXP had actions (+o4) and idea (+o5); the facets for RA included values (+o6) and mistrusting (−a1). While +o5 and +o6 were repeated predictors for several of the MQOS subscales, only +o4 appeared with EXP and only −a1 manifested with RA. Further, three openness facets for the prediction of QS comports with quest religiosity; QRO entails a religious lifestyle that resembles a growth mindset—an openness to behaviorally try new things (+o4) and a willingness to cognitively consider different and out of the box religious thoughts and values (+o6 and +o5) (McCrae and Costa 2010).

4.5.2. Applications

From these findings, there are a few practical applications. First is the responsiveness component in the quest religious orientation (QRO). As mentioned earlier, it entails an inclination to address personal existential struggles (e.g., identity, death, freedom) and complex questions relative to ordinary living and to trials/tragedies in life (Batson et al. 1993). Several of the obtained predictors for QS facilitate this feature. The Openness (+O; Explorer) trait elicits an intellectual curiosity that is imaginative and reflective along with divergent thinking enabling religious students to engage with any concerns or struggles (e.g., lack of finances, career selection, health of a relative) related to their pursuit of life goals and dreams. Each of the Openness facets offers a specific nuance to this curiosity. For instance, actions (+o4) stir an adventurous and behavioral bent, one that actively pursues new ideas and experiences. Ideas (+o5) prompt a willingness to consider unconventional topics—meaning no subject is off limits (e.g., students can believe differently than their parents or caregivers). Values (+o6) provokes a non-dogmatic approach, a flexibility to tackle complex features in topics avoiding being superficial.

Further, Neuroticism (+N; Reactive) educes a sensitivity and a familiarity with common negative affect (e.g., fear, sadness) that may be evoked as students engage in real-life problems and struggles (e.g., same-sex and opposite sex relationships). In addition, Agreeableness (−A; Challenger) with its hard-headed and persistent proclivities provides a means to ensure that solutions and answers to religious questions remain complex and not defaulted or compromised toward simplistic, naïve, and reductionistic answers (Howard and Howard 2010; McCrae and Costa 2010).

During responsiveness in QRO, religious undergraduates, as previously stated, need to embrace intellectual accountability. They can: (1) tackle complex existential issues in life with a trusted friend or mentor (Proverbs 15:22); they should not “go it alone” instead they need to collaborate; (2) research existential issues by pursuing multiple sources that offer the pro and cons on the subject; this helps them avoid being one-sided, narrow-minded, and overly simplistic; (3) prepare for possible condemnation from close associates and church leaders with a mentor. Others may not appreciate challenges to the status quo; hence, students should not be surprised or caught off guard; and (4) make responses to criticism respectful, even more so than the messages that come from critics. This can be accomplished by effectively listening to the critics’ whole message, asking clarifying questions, and then responding empathically to the critics’ messages.

Second is the doubt positivism component of QRO. As shared earlier, it refers to a comfortableness with uncertainty and with unanswered religious questions (Batson et al. 1993). Comfortability in this context refers to being relaxed with and at ease with the unknown, hesitation, and questions. Many of the emergent traits for QS advance this part of quest religiosity. Openness (+O; Explorer) prompts an openness to life experiences—not a closed mindset, contemplative habits, an aesthetic sensitivity, and an emotional attentiveness. Collectively, these proclivities provide experience handling the abstract and uncertainty by repeated learning opportunities that teach religious persons to release their false sense of control on certainty. For instance, emotional savviness educates and reminds people what they can and can’t control with their emergent feelings.

Values (+o6) elicit a non-dogmatic perspective, this invites broad-mindedness, non-rigidity, and tolerance. Each tendency makes room for multiple religious ideas and opinions and grants permission for differences; thus uncertainty, hesitation, and unanswered questions are not problematic when they enter college students’ belief system. They are familiar with opposing ideas and are experienced in mentally juggling them (Howard and Howard 2010).

Mistrust (−a1), an awkward social presentation, can assist religious undergraduates. The propensities of hard-headedness, persistence, and reluctance to accept information from others can create the social space between self and others (i.e., peers or church leaders) who make it their business to persuade students to follow their ideas (Howard and Howard 2010; McCrae and Costa 2010).

Yet, amid doubt positivism, religious college students need to remember two cautions: (1) avoid endless reflection; incessant and boundless considerations tend to be profitless and without benefit somewhat like a dog chasing its own tail; and (2) elude stubbornness and unteachability which commonly promote a fixed mindset and narrowmindedness.

Third is the openness feature in QRO. As indicated previously, it involves a willingness to search for new religious knowledge (exploration) and to potentially modify beliefs (change) (Batson et al. 1993). Some obtained traits for QS facilitate this part of quest religion. Although Neuroticism (+N; Reactive) is associated with an awkward social presentation (e.g., worried, tense, restless), it can provoke sensitivity, concern, and attentiveness that can empower college students to detect what might be wrong in their belief system and to have courage to begin the modification process. Put another way, Reactive can be the conscience of the exploration and change process.

Actions (+o4) prompt an adventurous spirit and preference for novelty; both can be sources of energy for religious students to explore new religious informational sources. Ideas (+o5) stir an intellectual curiosity; this includes a wide variety of interests and courage to pursue the unconventional which can translate into the liberty of having no topic being off limits in an explorative journey. In addition, this creative ingenuity assists religious undergraduates in tweaking an aspect of faith; it offers new ideas as to what and how the change can appear. Values (+o6) elicit a non-dogmatic broadmindedness which offers freedom and space to pursue religious knowledge from a wide variety of sources. This benefits exploration with a growth mindset—pursue information for improving spiritual well-being. It also aids the change process with the same freedom to judge one’s belief system and alter it as needed to enhance spiritual health.

However, amid openness, there are a few cautions. Religious undergraduates need to: (1) remember exploration and change inherently challenge the status quo and can be interpretated as something is wrong by church peers and acquaintances. Hence, they can prepare with a mentor for being misunderstood. They can anticipate how religious associates might respond to the explorative and change agendas and craft respectful, candid retorts; (2) collaborate with others and not explore or change alone. They can employ trusted friends and mentors to bounce ideas off them (Proverbs 15:22). Further, they can seek out multiple sources of information for exploration and change and practice the habit of cross-checking information; (3) propose changes to a belief system that are meaningful—ones that offer an improvement to the belief system; and (4) recall that change can be scary; they should not be surprised by emergent anxious moments; these can be managed.

Finally, the proclivities associated with each of the previously mentioned obtained traits with QS also underscore that a personality assessment (e.g., the Five-Factor model) can inform therapeutic work with undergraduates struggling with doubt. Assessment results can provide counselors with detailed information about their clients’ personality structure and help them develop tailored treatment plans. For instance, the negative affects (e.g., fear, anger) connected to Neuroticism (i.e., +N; Reactive) may be misunderstood and mismanaged by the religious undergraduate. Counselors can help students become more emotionally savvy about the adaptiveness of their negative emotions, especially in a doubting season.

4.6. Limitations and Future Research

The present study was not without limitations. First, the nature of the research questions and hierarchical regression procedures prevent cause and effect conclusions from being made regarding the variables of interest. Second, generalizations from the findings are limited to populations with similar features of this study’s sample (i.e., European American, Protestant).

These limitations, notwithstanding, the importance of studying the complexity of the doubt process, its stages and possible personality traits as predictors needs to remain a research priority. Future studies can include a replication of the present study to check the stability of its findings. Future investigations must define and operationalize religious doubt in a wider and realistic sense. There is a need to move beyond just one stage of the doubt process and include the other related stages as espoused by Krause and Ellison (2009); researchers also need to recognize the limits of imbalanced dichotomous discussions of religious doubt. Both positive and negative outcomes co-occur in real life; it is imperative that research studies investigate and discuss them together, not separately.

Relative to this sample, the experiences of other ethnic groups (e.g., Hispanic American, African American) and Christian sects (e.g., Catholic) with the stages of doubt and personality predictors along with the development of other doubt management strategies need investigation for comparisons with this study’s findings. In addition, the application of longitudinal methods would provide insight into possible changes in obtained personality traits over time per doubt stage. Further, the discovery of how religious people’s doubt relates to their belief and disbelief warrants attention. Smith (2014) argued Christians believe while they doubt, and one of the authors of this study expands it to—Christians believe while they doubt and disbelieve. Ecological Momentary Assessment (EMA) using smartphone app technology can be an appropriate methodology for such a research topic (Runyan et al. 2013). EMA records real-time data allowing for a reduction in recall bias and the collection of dynamic changes in doubt related thoughts, emotions, and behaviors over time.

5. Conclusions

Ortberg (2008) wrote, “One of the paradoxes of faith and doubt is that it is the ultimate intellectual challenge” (p. 25). The present study engaged in that challenge by investigating the predictive relationship between religious undergraduates’ stages of doubt and personality traits. From hierarchical regression analyses that attended to gender and suppression effects, eleven facet and five domain traits were discovered predictors for doubt criterion variables. Suggestions for doubt management per the four stages of doubt were also submitted. Collectively, these contributions expand the Protestant Christian church’s understanding of the complexity within religious doubt—its multiple stages, co-existing positive and negative sides along with doubters’ need for skillful management responses.