Career Decision Ambiguity Tolerance and Its Relations With Adherence to the RIASEC Structure and Calling

Abstract

While career decision ambiguity tolerance (CDAT) has been repeatedly shown to play a salient role in career decision-making, a potential revision of its measurement model has been indicated in past research. More importantly, it remains less clear as to what individual characteristics are associated with high CDAT. We examined a four-factor model of CDAT (i.e., preference, tolerance, confidence, and aversion) and its associations with adherence to the realistic, investigative, artistic, social, enterprise, and conventional (RIASEC) structure and career calling in a sample of college students (N = 248). The results of the current study supported the four-factor model being superior to the original three-factor model in representing the structure of CDAT. It was also found that adherence to the RIASEC structure was positively associated with preference, presence of calling was positively associated with confidence, and search for calling was positively associated with aversion. The implications of the results were discussed with respect to the theory and research of CDAT. Limitations and suggestions for future research were also provided.

Keywords

ambiguity tolerance career decision measurement model RIASEC career calling

Career decision-making is an important yet difficult process. The extant primary career counseling models emphasize the importance of collecting and processing information regarding self and the world of work and then using it to make choices (e.g., Holland, 1997; Parsons, 1909; Sampson, Lenz, Reardon, & Peterson, 1999). However, there has been an emerging proposition emphasizing the importance of ambiguity tolerance (Xu & Tracey, 2014, 2015a, 2015b), given the fact that a key component of career decision-making is dealing with unfamiliar, complex, inconsistent, or unpredictable information (Gati, Krausz, & Osipow, 1996). Xu and Tracey (2015b) have proposed and demonstrated that ambiguity tolerance specific to career decision-making is an important construct with respect to career decision-making process and outcomes. While they initially revealed a three-factor structure of CDAT (i.e., preference, tolerance, and aversion) in the U.S. college students (Xu & Tracey, 2015b), subsequent research suggested a plausible four-factor structure (i.e., preference, tolerance, confidence, and aversion; Xu, Hou, Tracey, & Zhang, 2016). One purpose of the current study was thus to examine this four-factor structure with respect to its structural validity and criterion validity. While the previous research on CDAT has been focused on its role in career decision-making (Xu & Tracey, 2014, 2015a, 2015b), there has been little research exploring other key vocational characteristics associated with this construct. With a refined measurement model, the current study was thus intended to examine the relations of CDAT with adherence to the realistic, investigative, artistic, social, enterprise, and conventional (RIASEC) structure (Tracey, 2008; Tracey & Darcy, 2002; Tracey, Lent, Brown, Soresi, & Nota, 2006) and career calling (Dik, Eldridge, Steger, & Duffy, 2012; Duffy & Dik, 2013).

Structural Model of CDAT

Ambiguity tolerance defines how individuals evaluate and respond to ambiguous situations characterized by unfamiliar, complex, or inconsistent information (Budner, 1962; Furnham & Ribchester, 1995). Early research of ambiguity tolerance was focused on the relation of ambiguity tolerance with ethnic prejudice, and results found that people intolerant of ambiguity tend to hold more ethnic stereotypes (Frenkel-Brunswik, 1949; Furnham & Ribchester, 1995). Frenkel-Brunswik (1949) first argued that ambiguity tolerance plays an important role in the entire emotional and cognitive functioning of individuals. Over the years, the research of ambiguity tolerance has been diffused to many other areas, such as creativity (Zenasni, Besançon, & Lubart, 2008), entrepreneurship (Ng, 2013; Schere, 1982), and anxiety and worry (Buhr & Dugas, 2006; Dugas, Gosselin, & Ladouceur, 2001). It was revealed that individuals tolerant of ambiguity are likely to exhibit more creativity (Zenasni et al., 2008), more competence in complex decision-making (Ng, 2013), and less general anxiety (Buhr & Dugas, 2006). These results thus collectively support the theoretical proposition that individuals holding a positive and functional reaction to ambiguity tend to show better performance and more satisfaction in ambiguity-intense tasks.

Unfortunately, career decision-making is such a process full of ambiguity, where information is commonly unknown, complex, inconsistent, or unpredictable (Gati et al., 1996). Therefore, the ability to handle ambiguity is critical in terms of career decision-making outcomes. However, the salience of ambiguity tolerance in career decision-making has not been well recognized until recent years (e.g., Xu, Hou, & Tracey, 2014; Xu et al., 2016; Xu & Tracey, 2014, 2015a, 2015b). One plausible reason of scarce research on ambiguity tolerance is that the prevailing career decision models carry the assumption that people can make informed decisions with adequate information (e.g., Holland, 1997; Parsons, 1909; Sampson et al., 1999). While the utility of collecting and processing information (e.g., interest, value, and skills) is certainly indisputable, the limitation of information-based models becomes increasingly present in this more complex and fluid modern society (Savickas et al., 2009).

The importance of ambiguity tolerance in career decision-making has been examined in its associations with career indecision, career exploration, and career decision-making self-efficacy. Xu and Tracey (2014) found that general ambiguity tolerance positively predicted career exploration and negatively predicted career indecision. In addition, they found a moderation effect of general ambiguity tolerance on the link of career exploration with career indecision. Subsequently, Xu and Tracey (2015a) revealed that general ambiguity tolerance not only directly predicted career indecision but also indirectly predicted career indecision through the mediation of career decision-making self-efficacy. Given the positive results supporting the role of general ambiguity tolerance in career decision-making, Xu and Tracey (2015b) defined CDAT as people’s evaluations of and responses to ambiguity specifically encountered in career decision-making and they also developed a scale to measure this domain-specific construct. Through an exploratory and confirmatory factor analysis, they found a three-factor structure of preference, tolerance, and aversion underlying CDAT, and these three factors were found to additively predict career indecision, career decision-making self-efficacy, and career adaptability over and beyond general ambiguity tolerance (Xu & Tracey, 2015b).

In the original three-factor structural model of CDAT, preference defines an individual’s tendency to feel interested and excited for ambiguity (particularly new information) in career decision-making. Tolerance defines an individual’s tendency to accept ambiguity and feel competent in tackling ambiguity in career decision-making. Aversion defines an individual’s tendency to find ambiguity anxiety-provoking and avoid it in career decision-making (Xu & Tracey, 2015b). While the factors of preference and aversion have appeared to be psychometrically sound in the original and subsequent research (Xu et al., 2016; Xu & Tracey, 2015b), the factor of tolerance has received relatively less support for its structural coherence (Xu et al., 2016; Xu & Tracey, 2015b). It has been shown that the tolerance factor has a relatively lower internal consistency (Xu & Tracey, 2015b) and poor cross-cultural validity (Xu et al., 2016) compared to preference and aversion. Given the fact that acceptance and confidence are emphasized, respectively, in collectivistic and individualistic cultures (Bai, 2005; Triandis, 1995). Xu, Hou, Tracey, and Zhang (2016) suggested that blending tolerance/acceptance and confidence items into one single factor might potentially contribute to the structure ambivalence. Therefore, it is expected that a revised model wherein the original tolerance factor is divided into two separate factors of tolerance and confidence could account for item relations better and enrich the measurement of CDAT.

The revised four-factor model of CDAT also makes sense conceptually in that it captures four distinct types in terms of how people evaluate and react to ambiguity. While aversion taps a maladaptive style, preference, tolerance, and confidence tap a separate group of functional and adaptive styles. However, a general higher order factor of adaptive coping might be less appropriate than bivariate correlations in describing the relations of preference, tolerance, and confidence, because these three factors predominantly tap three specific and different styles. Among them, preference embraces a positive evaluation of ambiguity, while tolerance and confidence embrace a negative evaluation of ambiguity. Between tolerance and confidence, tolerance represents a passive coping style, while confidence represents a proactive and effortful coping style. Given the empirical results and theoretical differentiations of the four styles discussed above, it is reasonable to expect that a four-factor oblique model would be better than a three-factor oblique model and a four-factor hierarchical model in representing the structure of CDAT (Hypothesis a1).

The first purpose of the present study was thus to examine this four-factor model, particularly by comparing it with the original three-factor model and correlating it with career exploration, career decision-making self-efficacy, and career certainty. These constructs were chosen as the convergent criterion because they have been widely considered as important career decision variables and shown to relate to CDAT (Xu et al., 2016; Xu & Tracey, 2014, 2015a, 2015b). The previous research has portrayed preference and aversion being associated, respectively, with career exploration and indecision (Xu et al., 2016). It is thus anticipated that preference would positively predict all the criteria except certainty and aversion would negatively predict all the criteria except exploration (Hypothesis b1). We argued that tolerance mainly plays a role in decision outcomes and evaluations and thus hypothesized it being predictive of certainty only (Hypothesis b2). As confidence in coping with ambiguity is involved in any aspect of career decision-making, we hypothesized that confidence would positively predict all the criteria (Hypothesis b3).

Adherence to the RIASEC Structure With CDAT

While the previous research has well supported the significant role of CDAT in career decision-making (Xu et al., 2016; Xu & Tracey, 2014, 2015a, 2015b), there has been an increasing need to explore what contributes to a higher level of CDAT. We argue that adherence to the RIASEC structure (Tracey, 2008; Tracey & Darcy, 2002) and career calling (Dik et al., 2012; Duffy & Dik, 2013) are two plausible candidates.

Adherence to RIASEC structure defines the extent to which individuals perceive six interest types in the normative circular order prescribed by Holland’s (1997) model. Holland’s model proposes that the six interest types of realistic, investigative, artistic, social, enterprise, and conventional are arranged on a circular structure, where adjacent interest type (e.g., R and I) are closer than alternate interest types (e.g., R and A) and alternate interest types (e.g., R and A) are closer than opposite interest types (e.g., R and S). This normative structure of interest types has received extensive support for its structural validity (e.g., Darcy & Tracey, 2007; Gupta, Tracey, & Gore Jr., 2008; Tracey & Rounds, 1993) and has been widely used as a guiding model for career counseling and intervention.

While the RIASEC structure was derived based on correlational information among groups, individuals have been found to show variance in terms of their using this same circular structure in their understanding of interests and occupations (Tracey, 2008; Tracey & Darcy, 2002; Tracey et al., 2006). Tracey (2008) argued that a close resemblance of how individuals perceive interest types with the RIASEC model could mark an important progress in career decision-making, given the fact that the RIASEC model provides a simplistic and elegant structure organizing and mapping activities and careers. Based on this idiographic approach, it has been found that individual adherence to the RIASEC structure was negatively associated with career indecision and positively associated with career exploration, career certainty, interest–occupation congruence, and career decision-making self-efficacy (Tracey, 2008; Tracey & Darcy, 2002; Tracey et al., 2006).

Given the evidence portraying adherence to the RIASEC structure as a conducive construct in career decision-making, it is plausible to argue that adherence to the RIASEC structure could predict CDAT. Individuals using the RIASEC structure to organize information of the self and the vocational world are expected to show more interest in new information and less anxiety in handling ambiguity, as the RIASEC structure could serve as a tool accounting for complex and inconsistent information. We thus hypothesized in the current study that adherence to the RIASEC structure positively predicts preference and negatively predicts aversion (Hypothesis c1). As tolerance is conceived to be potentially related to cultural values (Xu et al., 2016) and confidence mainly results from social cognitive learning experiences (Bandura, 1977), the cognitive RIASEC structure is expected to be minimally correlated with these two factors (Hypothesis c2).

Career Calling With CDAT

In addition to adherence to the RIASEC structure, career calling is expected to relate to CDAT as well. There has been an increasing scholarly interest in career calling in the vocational psychology literature since 2007 (Duffy & Dik, 2013). The definition of calling is controversial and difficult particularly regarding the source of calling (Dik et al., 2012; Duffy & Dik, 2013). Traditionally, calling involves an external caller, who often comes in the form of a religious higher being. Dik and Duffy (2009) reviewed the literature and identified three components of calling, consisting of an external summons, meaning/purpose, and prosocial motivation. While the last two components are commonly agreed upon, the source of calling has been expanded to incorporate internal sources and the interplay of internal and external sources (Duffy & Dik, 2013). Currently, there have been three prominent perspectives posited regarding the source of calling: external summons, sense of destiny, and perfect fit (Duffy, Allan, Bott, & Dik, 2014). While perfect fit has been found to be the most common source of calling, it has been shown with limited research that the source of calling has little relation with life and work satisfaction (Duffy et al., 2014).

There are two prominent aspects in the attainment of a calling, consisting of presence of calling and search for calling. Where presence of calling describes to which extent individuals perceive a career calling, search for calling describes to which extent individuals seek a career calling (Duffy & Sedlacek, 2007). As calling serves as a motivational function providing sense of meaning/purpose, it is reasonable to expect that individuals perceiving a calling would exhibit career development and academic/work/life satisfaction. There has been empirical research supporting the presence of a calling being associated with career decidedness, career decision-making self-efficacy, work and life satisfaction, and life meaning (Dik, Sargent, & Steger, 2008; Duffy, Allan, & Dik, 2011; Duffy & Sedlacek, 2007). In general, the research has suggested that individuals with a calling tend to feel more comfortable and confident with their career choice and find the career they are called to more meaningful.

Therefore, we argue that presence of calling could help individuals cope with ambiguity in career decision-making, because calling would direct individuals to a specific career and provide a cohesive structure of meaning/purpose. More specifically, we hypothesize that presence of calling would lead to more confidence in handling ambiguity (i.e., confidence) and less aversion with ambiguity (i.e., aversion; Hypothesis c3). We hypothesize that there are no links between presence of calling with preference and tolerance (Hypothesis c4), as calling is hardly related to interest in new information and acceptance of ambiguity.

While presence of calling has been shown to link to positive career outcomes across the board (Dik et al., 2008; Duffy et al., 2011; Duffy & Sedlacek, 2007), search of calling has received much less attention regarding its role in career development. The past research has found that search for calling positively correlates with indecisiveness and negatively correlates with career decidedness and self-clarity (Duffy & Sedlacek, 2007). These results speak to the phenomenon that search for a career calling could be driven by the barriers of career decision-making(e.g., indecision and lack of self-clarity). However, little is known about what individual characteristics are associated with the calling search behaviors. We proposed that CDAT relates to search for calling, given the reasoning that individuals showing a strong need for a coherent meaning and purpose structure are more likely to be people with difficulty in handling ambiguity. Among the four factors of preference, tolerance, confidence, and aversion, aversion is perceived as the most important motivational factor correlating with search of calling, as aversion specifically taps the difficulty in facing inconsistent and complex situations. We thus hypothesized a positive relation of search for calling with aversion (Hypothesis c5) and no links between search for calling with preference, tolerance, and confidence (Hypothesis c6).

Summary

Given the increasing speculation of a four-factor model of CDAT (i.e., preference, tolerance, confidence, and aversion; Xu et al., 2016; Xu & Tracey, 2015b), the current study examined the structural validity and criterion validity of this revised model in its superior model–data fit relative to the original three-factor model (Hypothesis a1) and its covariance with career exploration, career decision-making self-efficacy, and career certainty. It is anticipated that preference positively predict all the criteria except certainty and aversion negatively predicted all the criteria except exploration (Hypothesis b1). Tolerance was hypothesized to only positively predict career certainty (Hypothesis b2) and confidence was hypothesized to positively predict all the criteria (Hypothesis b3).

Based on the four-factor model, the relations of CDAT with adherence to the RIASEC structure, presence of calling, and search for calling were examined. We hypothesized that adherence to the RIASEC structure positively predicts preference and negatively predicts aversion (Hypothesis c1). The adherence to the RIASEC structure was not hypothesized to predict tolerance and confidence (Hypothesis c2). We hypothesized that presence of calling positively predicts confidence and negatively predicts aversion (Hypothesis c3), and presence of calling does not predict preference and tolerance (Hypothesis c4). Moreover, we hypothesized that search for calling positively predicts aversion (Hypothesis c5) and does not predict preference, tolerance, and confidence (Hypothesis c6).

Method

Sample

The current sample consisted of 248 undergraduate students recruited from a southwest state university. They ranged in age from 18 to 47 (M = 20.30, SD = 4.00). Of the sample, 31.9% were male (n = 79), 67.3% were female (n = 167), and 0.4% were self-identified as transgender (n = 1). In terms of race/ethnicity, 8.9% (n = 22) were African American/Black, 8.9% (n = 22) were Asian/Asian American, 19.0% (n = 47) were Latino(a)/Hispanic, 54.8% (n = 136) were Caucasian/White, 0.8% (n = 2) were Native American, 4.0% (n = 10) were multiracial, 3.2% (n = 8) were self-identified as others.

Measures

The CDAT Scale–Revised (CDAT-R)

The original 18-item CDAT (Xu & Tracey, 2015b) was developed to measure people’s evaluations of and responses to unfamiliar, complex, inconsistent, and unpredictable information in career decision-making. It contains three subscales of preference (6 items), tolerance (6 items), and aversion (6 items). Xu and Tracey (2015b) reported α coefficients of .83, .70, and .81 for the three subscales of preference, tolerance, and aversion, respectively. The validity of the CDAT has also been supported in its association with career indecision, career decision-making self-efficacy, and career adaptability (Xu & Tracey, 2015b). The current study revised this scale by adding 7 items to enhance the separate measurement of tolerance and confidence, 5 of which were specifically designed to measure confidence and 2 of which were designed to improve old tolerance items. The proposed four-factor model consists of preference, tolerance, confidence, and aversion. Preference measures individual tendency to feel interested and excited for ambiguity in career decision-making (e.g., “I am interested in exploring the many aspects of my personality and interests”). Tolerance measures individual tendency to experience acceptance of ambiguity in career decision-making (e.g., “I am tolerant of the unpredictability of a career”). Confidence measures individual tendency to feel competent in coping with ambiguity in career decision-making (e.g., “I am confident in tackling complex career decision-making tasks”). Aversion measures individual tendency to avoid and withdraw from ambiguity in career decision-making (e.g., “I try to avoid complicated career decision-making tasks”). Participants would be invited to rate each item on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Higher scores indicated higher endorsement of each factor. The current study found α coefficients of .90, .84, .86, and .82 for the four subscales of preference, tolerance, confidence, and aversion.

Career Exploration Survey (CES)

The 11-item CES was designed to measure individual exploratory behaviors in two domains, consisting of gathering information regarding self (self-exploration [SE]) and gathering information regarding the vocational world (environmental exploration [EE]; Stumpf, Colarelli, & Hartman, 1983). A 5-point Likert-type scale is used to measure participants’ career exploration, ranging from 1 (very little) to 5 (very much). Past research has supported adequate internal reliability of the CES in its internal consistency αs ranging from .60 to .88 (Bartley & Robitschek, 2000; Nauta, 2007; Stumpf et al., 1983). The current study found α coefficients of .87 and .84 for the EE and SE subscales, respectively.

The Career Decision Self-Efficacy–Short Form (CDSE-SF)

Participants’ career decision-making self-efficacy was assessed by the 25-item CDSE-SF (Betz, Klein, & Taylor, 1996) based on a 5-point Likert-type scale ranging from 1 (no competence at all) to 5 (complete competence). Self-efficacy on five skill domains was evaluated based on career maturity model (Crites, 1978), consisting of accurate self-appraisal, gathering occupational information, goal selection, making plans for the future, and problem-solving. The internal reliability has been well supported in excellent internal consistency αs at the .90 range (Betz & Luzzo, 1996). There has also been extensive research supporting the validity of the CDSE-SF in the population being investigated in this study (e.g., Betz & Luzzo, 1996). The current data revealed an α coefficient of .95.

The Career Decision Scale (CDS)

The CDS is a 19-item measure with a 4-point Likert-type scale ranging from 1 (exactly like me) to 4 (not at all like me). The scale consists of the CDS-Certainty subscale (2 items) and the CDS-Indecision subscale (16 items). Only the CDS-Certainty subscale was used in the current study to measure to which extent individuals are certain about their career/major choice. The past research has revealed internal consistency coefficients of the CDS-Certainty in the .80 range (Hartman, Fuqua, & Hartman, 1983; Tracey, 2008). The validity of the CDS-Certainty has been supported in the findings of its association with a variety of important career decision variables (Osipow, 1987; Tracey, 2008). The current study found an internal consistency α of .79.

The Inventory of Children’s Activities–Paired (ICA-P)

The ICA-P (Tracey & Darcy, 2002) was developed to enable an assessment of individual structural perceptions of the RIASEC activities and its similarity to the normative RIASEC structure (i.e., adherence to the RIASEC structure). Based on the factor loadings, the best single item for each RIASEC scale of the ICA–Revised (ICA-R; (Tracey & Ward, 1998) was selected, and the total 6 items were combined into all possible pairs and arranged randomly to form the ICA-P. Although the ICA-R was originally designed for children, its items have been found to work well for college students as well (Tracey & Ward, 1998). Individuals are asked to rate the degree of similarity of each pair on a 6-point Likert-type scale ranging from 1 (totally different) to 6 (very much alike). A 6 × 6 matrix of similarity ratings among the RIASEC activities thus results for each individual. As the RIASEC circular model specifies 72 predictions regarding the orders of similarity ratings (Tracey & Rounds, 1993), the fit of each individual’s similarity matrix to the RIASEC circular structure (i.e., adherence to the RIASEC structure) is determined by the number of predictions met. While the previous research often calculated the adherence scores by comparing predictions met and unmet (Tracey, 2008; Tracey et al., 2006), the tie predictions were not taken into consideration. However, the current study was focused on the absolute numbers of predictions met, as both tie and unmet predictions could indicate a less differentiated interest structural profile, which is not ideal for organizing information. Because ratings are not aggregated through items, internal consistency reliability is not an appropriate psychometric index for the ICA-P. The validity of the ICA-P in assessing individual variation from the RIASEC circular structure has been supported in its associations with important career decision variables (Tracey, 2008; Tracey et al., 2006).

The Brief Calling Scale (BCS)

The BCS (Dik et al., 2012) was designed to measure people’s presence of and search for a calling. It contains two subscales, consisting of presence of calling (2 items) and search for calling (2 items). Presence of calling describes to which extent individuals perceive a calling (e.g., “I have a calling to a particular kind of work”). Search for calling describes to which extent individuals search for a calling (e.g., “I am searching for my calling as it applies to my career”). Participants were asked to rate on a 5-point Likert-type scale ranging from 1 (not at all true of me) to 5 (totally true of me). Higher scores indicate stronger presence of and search for calling. The 2 items of the BCS-presence and the 2 items of the BCS-search were reported to correlate at r = .81 and r = .75, respectively (Duffy & Sedlacek, 2007). The validity of the BCS has been supported by its correlations with criterion variables, such as career decision-making self-efficacy and life/work satisfaction (Dik et al., 2012; Dik et al., 2008; Duffy & Sedlacek, 2007). The current study found internal consistency coefficients of .86 and .82 for BCS-presence and BCS-search, respectively.

Procedure

College students attending career development, university orientation, social statistics, or introduction to psychology classes were invited to participate in this study as an extra credit opportunity. Institutional Review Board (IRB) approval was obtained for the study. Roughly 1,000 students received the public announcement regarding this research. Students consenting to this research filled a demographic questionnaire and the package of research instruments online. No identifying information was solicited on the online survey. All the individual responses were kept as anonymous and confidential through analysis. The missing rate in this study ranged from 4.8% to 10.5%. We followed Schlomer, Bauman, and Card’s (2010) suggestion and used the full information maximum likelihood estimation. This approach estimates model parameters based on all available information and has been shown to outperform mean substitution in estimations of coefficients and standard errors when the amount of missing data is small to moderate (Schlomer, Bauman, & Card, 2010).

Analysis

We used Mplus 7.0 to conduct structural equation modeling on three progressive analyses. The first confirmatory factor analysis precisely examined the four-factor model of CDAT. While the model–data fit of the four-factor model was examined, a particular focus was on the comparison of the nested three-factor model and four-factor model. The second analysis examined the criterion validity of the CDAT-R by looking at its covariance with career exploration, career decision-making self-efficacy, and career certainty. The third analysis focused on the key hypotheses of the current study and examined the hypothesized covariance of the four factors of the CDAT-R with adherence to the RIASEC structure, presence of calling, and search for calling.

The CDAT-R, EE, SE, CDSE-SF, and CDS-Certainty were assessed based on latent construct models with the items/subscales as the manifest indicators. The BCS-presence and BCS-search were assessed by the subscale scores, respectively. The fit of the models was evaluated using the criteria recommended by Hu and Bentler (1999): robust χ², Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). With the purpose of making the statistical tests robust to potential nonnormality, we adopted the robust maximum likelihood parameter estimation. Differences between nested models were examined with the Santorra–Bentler scaled χ² difference test (Muthén & Muthén, 2012). A nonsignificant result of the Santorra–Bentler scaled χ² difference test indicated that an invariance model is a better representation of the data (Muthén & Muthén, 2012).

Results

Table 1 summarizes the means, standard deviations, and the bivariate correlations of the CDAT-R, EE, SE, CDSE-SF, CDS-Certainty, BCS-presence, BCS-search, and adherence to the RIASEC structure. Table 2 summarizes the fit indices of all the models.

Table 1.
Means, SD, and Correlations of Variables.

Variables M SD Preference Tolerance Confidence Aversion EE SE CDSE-SF Certainty Presence Search

Preference 6.03 0.97 1.00

Tolerance 5.12 1.12 .54 1.00

Confidence 4.89 1.11 .40 .45 1.00

Aversion 4.10 1.17 .08 .02 –.27 1.00

EE 2.87 0.96 .06 .05 .34 –.22 1.00

SE 3.30 0.95 .20 .13 .18 –.03 .51 1.00

CDSE-SF 3.58 0.68 .35 .31 .59 –.32 .45 .37 1.00

Certainty 2.93 0.75 –.01 –.09 .36 –.40 .38 .14* .47 1.00

Presence 6.43 2.20 .09 –.01 .19 –.18 .32 .28 .45 .41 1.00

Search 6.19 2.37 .01 .02 –.11 .27 .06 .19** –.17* –.31 –.21 1.00

Adherence 27.09 10.06 .20** .06 –.05 –.01 –.10 .04 .08 –.09 .03 .06

Note. N = 248. EE = career exploration survey-environmental exploration; SE = career exploration survey-self-exploration; CDSE-SF = Career Decision Self-Efficacy–Short Form; certainty = Career Decision Scale-certainty; presence = Brief Calling Scale-presence of calling; search = Brief Calling Scale-search of calling; adherence = adherence to the realistic, investigative, artistic, social, enterprise, and conventional structure.

*p < .05. **p < .01.

Table 2.
Summary of Model Fit Indices.

Variables χ² df AIC BIC CFI RMSEA SRMR

Estimate 90% CI

Model a1. Three-factor model with original items 394.06 132 13,655.22 13,852.66 .83 .092 [.081, .102] .08

Model a2. Four-factor model with original items 313.28 129 13,580.27 13,788.10 .88 .078 [.067, .089] .07

Model a3. Three-factor model with new items 637.64 167 14,376.37 14,594.59 .77 .109 [.100, .118] .10

Model a4. Four-factor model with new items 346.07 164 14,070.88 14,299.49 .91 .069 [.058, .079] .06

Model a5. Four-factor hierarchical model 377.61 166 14,100.25 14,321.94 .90 .073 [.064, .083] .08

Model b1. Criterion validity model 1,123.39 643 23,327.11 23,798.19 .90 .056 [.051, .062] .07

Model c1. Relation with calling and adherence 422.70 219 13,231.20 13,472.79 .90 .065 [.055, .074] .06

Note. N = 248. AIC = Akaike information criterion; BIC = Bayesian information criterion; CI = confidence interval.

In the first analysis, we first examined Model a1 with the original CDAT items fitting the three-factor model. As can been seen by the values of RMSEA (.092), SRMR (.08), and CFI (.83) in Table 2, the three-factor model did not fit the data adequately. We then examined Model a2, where the tolerance and confidence items of the original CDAT were treated as two separate factors (i.e., the four-factor model). As can been seen from the fit values presented in Table 2, (i.e., RMSEA = .078, SRMR = .07, and CFI = .88), the four-factor model fit the data adequately. The corrected χ² difference test indicated that Model a2 was a better representation of the data compared to Model a1, scaled Δχ²(3, N = 248) = 58.45, p < .05. Therefore, the four-factor model was supported particularly in its structural superiority to the three-factor model based on the original item set of the CDAT.

We then replicated the analysis using the new, expanded item set of the CDAT-R. We first specified Model a3, where tolerance and confidence of the CDAT-R were combined as a single factor (i.e., the three-factor model). As can been seen by the value of RMSEA (.109), SRMR (.10), and CFI (.77) in Table 2, the three-factor model did not fit the data adequately. We then examined Model a4, where tolerance and confidence of the CDAT-R were treated as two separate factors (i.e., the four-factor model). As can been seen from the fit values presented in Table 2 (i.e., RMSEA = .069, SRMR = .06, and CFI = .91), the four-factor model fit the data adequately. The corrected χ² difference test indicated that Model a4 fit the data better than Model a3, scaled Δχ²(3, N = 248) = 1,194.62, p < .05. Therefore, the structural superiority of the four-factor model was supported again based on the CDAT-R. The fit values cannot be statistically compared between the CDAT and the CDAT-R, as they employ different indicators. However, the CDAT-R provides a better conceptual breadth in measuring the multidimensional model of CDAT. We additionally examined the four-factor hierarchical model (Model a5) using the CDAT-R. While the fit indices were adequate (RMSEA = .073, SRMR = .08, and CFI = .90), the Akaike information criterion and Bayesian information criterion of this model were worse than those of the four-factor oblique model. An examination of the factor loadings also revealed the higher order factor being predictive of the four first-order factors unevenly, indicating that a general factor is inappropriate in accounting for the correlations of the four factors. Hypothesis a1 was thus well supported.

In the second analysis, we examined Model b1, where the four factors of the CDAT-R were associated with the criterion constructs in the hypothesized way. As can been seen by the values of RMSEA (.056), SRMR (.07), and CFI (.90), this model fit the data adequately. An examination of the modification indices did not reveal any major area of misfit that could be corrected by adding regression paths. We then examined individual parameter estimates and found that preference was positively predictive of CDSE-SF and SE (but not EE); tolerance was negatively predictive of career certainty; confidence was positively predictive of CDSE-SF, career certainty, and EE (but not SE); and aversion was negatively predictive of CDSE-SF and career certainty. Therefore, Hypotheses b1, b2, and b3 were largely supported in this study. Figure 1 depicts all the standardized parameter estimates. In general, the four factors showed differential associations with EE, SE, career decision-making self-efficacy, and career certainty in expected directions, which further supported the divergent validity of the four-factor CDAT-R.

Figure 1.
Standardized parameter estimates for Model b1. CDATP = Career Decision Ambiguity Tolerance Scale–Preference; CDATT = Career Decision Ambiguity Tolerance Scale–Tolerance; CDATC = Career Decision Ambiguity Tolerance Scale–Confidence; CDATA= Career Decision Ambiguity Tolerance Scale–Aversion; EE = environmental exploration; SE = self-exploration; CDSE = career decision self-efficacy. Bold regression paths are significant, p < .05.

In the third analysis, we examined Model c1, which specified the hypothesized associations of the four factors of the CDAT-R with the BCS-presence, BCS-search, and adherence to the RIASEC structure. As can been seen by the values of RMSEA (.065), SRMR (.06), and CFI (.90), this model fit the data adequately. An examination of the modification indices did not reveal any major association ignored in the hypothesized model. A subsequent examination of the individual parameter estimates (see Figure 2) indicated that adherence to the RIASEC structure only positively predicted preference, the BCS-presence only positively predicted confidence, and the BCS-search only positively predicted aversion. Therefore, Hypotheses c2, c4, c5, and c6 were fully supported and Hypotheses c1 and c3 were partially supported. The results suggested that individuals perceiving activities/careers in a more similar way to the RIASEC structure tend to find new information interesting and exciting. It was also suggested that individuals perceiving a calling tend to have more confidence in handling ambiguity in career decision-making. Last, it was suggested that individuals searching for a calling tend to be individuals perceiving ambiguity as an undesirable challenge.

Figure 2.
Standardized parameter estimates for Model c1. CDATP = Career Decision Ambiguity Tolerance Scale–Preference; CDATT = Career Decision Ambiguity Tolerance Scale–Tolerance; CDATC = Career Decision Ambiguity Tolerance Scale–Confidence; CDATA = Career Decision Ambiguity Tolerance Scale–Aversion; presence = presence of calling; search = search for calling. Bold regression paths are significant, p < .05.

Discussion

The present study first examined the four-factor model of CDAT. The results of confirmatory factor analyses were repeatedly supportive of the four-factor model (i.e., preference, tolerance, confidence, and aversion) being superior to the three-factor model (i.e., preference, tolerance, and aversion) in terms of model–data fit. Conceptually, the factor of confidence captures a unique aspect of how individuals evaluate and respond to ambiguity in career decision-making, which has not been fully assessed in the original CDAT model (Xu & Tracey, 2015b). Based on the three-factor model, individuals might choose approaching, acceptance, and avoidance of ambiguity as a reaction to their evaluations of ambiguity. However, among preference, tolerance, and aversion, none of them embraces a tendency to fight with and potentially change ambiguity. Therefore, it appears meaningful to incorporate such a proactive and effortful style (i.e., confidence) in the measurement of CDAT, particularly given the wide interest in self-regulation and self-efficacy across domains of psychology (Bandura, 1977).

The empirical results of the current study also supported the meaning of confidence in CDAT as can be seen by its associations with EE, career decision-making self-efficacy, and career certainty. It was found that confidence strongly predicted EE, career decision-making self-efficacy, and career certainty. Such a relational pattern portrays confidence as an important aspect of CDAT in the U.S. context. As confidence is highly valued in individualistic cultures than collectivistic cultures (Triandis, 1995), it would be interesting to see future research investigating how confidence plays out in career decision-making in collectivistic contexts.

Related to confidence, it would be important to examine how tolerance plays out in career decision-making in collectivistic contexts, as acceptance of unpredictability is often emphasized in those cultures (Bai, 2005). It can be seen from the current study that tolerance negatively predicted career certainty. This result supported the separate assessment of tolerance by revealing that individuals with higher tolerance and acceptance of ambiguity tend to have lower certainty as to their career choice. However, the direction of this association is opposite of the hypothesis. It challenges our assumption of the dynamic role of tolerance in career decision-making. Does tolerance with ambiguity help individuals feel more confident with their career choice or allow individuals to wonder about their choice more? The current results are more than far away from conclusive and future research using a longitudinal approach to examine this question would be helpful.

It is also noteworthy that preference was not associated with aversion. This result is not unique to the current study and replicates the previous research revealing a weak link between these two seemingly opposite factors (Xu et al., 2016; Xu & Tracey, 2015b). Theoretically, these two factors deviate from each other as they belong to adaptive and maladaptive styles, respectively. An examination of the items of preference and aversion additionally suggested that the negligible link is attributable to the fact that preference is more targeted at new and unknown information, while aversion is more targeted at complex and unpredictable information. Therefore, not only tolerance and confidence but also preference and aversion are distinct factors.

Based on the four-factor model, the current study examined the relations of CDAT with adherence to the normative RIASEC structure (Tracey, 2008; Tracey & Darcy, 2002). While we anticipated that a close resemblance of how individuals perceive activities/careers with the RIASEC structure would help with preference and aversion, the present results only found support for its link with preference. It was thus suggested that people using the RIASEC structure to organize career-related information are more likely to find new and unknown information interesting and exciting in career decision-making, although they might still perceive complex and unpredictable information undesirable. This finding is consistent with Tracey, Lent, Brown, Soresi, and Nota’s (2006) research revealing the positive link of adherence to the RIASEC structure with career exploration, as preference has been shown to associate with career exploration in Xu et al.’s (2016) and in this study. Together, the research supports the utility of the RIASEC structure in helping individuals organize and find fun in new and complex information, which might subsequently help individuals engage in career exploration.

While the positive function of career calling has been revealed in the previous research (Dik et al., 2008; Duffy & Dik, 2013; Duffy & Sedlacek, 2007), the present study further found support for the role of presence of calling in CDAT. It was suggested that individuals perceiving a calling tend to have more confidence in coping with ambiguity in career decision-making. This result is not surprising, as one common understanding of calling rests on its motivational function in providing a coherent structure of meaning and purpose (Dik & Duffy, 2009; Duffy & Dik, 2013). It can be reasoned that a single structure of meaning and purpose could reduce people’s sense of ambiguity as to how to interpret and use the information about the self the world of work. Interestingly, presence of calling was not found to predict aversion, although we anticipated that individuals perceiving a calling tend to feel more comfortable with ambiguity. This finding might make more sense when taking into consideration the revealed link of search for calling with aversion. It was suggested in the current study that individuals searching for a calling are likely to be the individuals feeling uncomfortable with ambiguity. Therefore, a plausible mechanism of calling interacting with ambiguity implied by this study is that individuals anxious about ambiguity tend to search for a career calling, but such a coping strategy (i.e., presence of calling) does not necessarily change their perception about ambiguity. As the current results are preliminary, more research examining the dynamic interplay of calling with CDAT would be needed.

There are several limitations regarding the conclusions drawn from the current results. First, this study investigated a sample of college students. Therefore, the results might not generalize to other populations, such as older adults facing career transitions. It would thus be interesting to see future research examining the individual characteristics of people with high CDAT. Second, the current study was focused on presence of and search for calling in investigating calling. However, living a calling has been proposed as a crucial mediator in the link of calling with life/job satisfaction (Duffy, Allan, Autin, & Bott, 2013; Duffy, Bott, Allan, Torrey, & Dik, 2012). Hence, it would potentially provide a more complete picture of how calling interacts with CDAT, if living a calling, presence of calling, and search for calling could be modeled together in future research.

To sum up, the current study contributes to the measurement of CDAT by establishing a four-factor model: preference, tolerance, confidence, and aversion, along with a validated refined measure (i.e., CDAT-R). More importantly, the current study first revealed certain individual characteristics being associated with CDAT. It was suggested that individuals perceiving activities/careers based on the RIASEC structure tend to feel more interested in finding new and unknown information in career decision-making. It was suggested that individuals perceiving a career calling tend to feel more confident in handling ambiguity in career decision-making. Last but not least, it was suggested that individuals searching for a calling tend to be people perceiving ambiguity as an undesirable challenge in career decision-making.

Variables	M	SD	Preference	Tolerance	Confidence	Aversion	EE	SE	CDSE-SF	Certainty	Presence	Search
Preference	6.03	0.97	1.00
Tolerance	5.12	1.12	.54**	1.00
Confidence	4.89	1.11	.40**	.45**	1.00
Aversion	4.10	1.17	.08	.02	–.27**	1.00
EE	2.87	0.96	.06	.05	.34**	–.22**	1.00
SE	3.30	0.95	.20**	.13	.18**	–.03	.51**	1.00
CDSE-SF	3.58	0.68	.35**	.31**	.59**	–.32**	.45**	.37**	1.00
Certainty	2.93	0.75	–.01	–.09	.36**	–.40**	.38**	.14*	.47**	1.00
Presence	6.43	2.20	.09	–.01	.19**	–.18**	.32**	.28**	.45**	.41**	1.00
Search	6.19	2.37	.01	.02	–.11	.27**	.06	.19**	–.17*	–.31**	–.21**	1.00
Adherence	27.09	10.06	.20**	.06	–.05	–.01	–.10	.04	.08	–.09	.03	.06

Variables	χ²	df	AIC	BIC	CFI	RMSEA	SRMR
Model a1. Three-factor model with original items	394.06	132	13,655.22	13,852.66	.83	.092	[.081, .102]	.08
Model a2. Four-factor model with original items	313.28	129	13,580.27	13,788.10	.88	.078	[.067, .089]	.07
Model a3. Three-factor model with new items	637.64	167	14,376.37	14,594.59	.77	.109	[.100, .118]	.10
Model a4. Four-factor model with new items	346.07	164	14,070.88	14,299.49	.91	.069	[.058, .079]	.06
Model a5. Four-factor hierarchical model	377.61	166	14,100.25	14,321.94	.90	.073	[.064, .083]	.08
Model b1. Criterion validity model	1,123.39	643	23,327.11	23,798.19	.90	.056	[.051, .062]	.07
Model c1. Relation with calling and adherence	422.70	219	13,231.20	13,472.79	.90	.065	[.055, .074]	.06

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Bai

(2005). An outline history of China. Beijing, China: Tongxin Press.

Bandura

(1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215. doi:10.1037/0033-295X.84.2.191

Bartley

D. F.

Robitschek

(2000). Career exploration: A multivariate analysis of predictors. Journal of Vocational Behavior, 56, 63–81.

Betz

N. E.

Klein

K. L.

Taylor

K. M.

(1996). Evaluation of a short form of the Career Decision-Making Self-Efficacy Scale. Journal of Career Assessment, 4, 47–57. doi:10.1177/106907279600400103

Betz

N. E.

Luzzo

D. A.

(1996). Career assessment and the Career Decision-Making Self-Efficacy Scale. Journal of Career Assessment, 4, 413–428. doi:10.1177/106907279600400405

Budner

(1962). Intolerance of ambiguity as a personality variable. Journal of personality, 30, 29–50. doi:10.1111/j.1467-6494.1962.tb02303.x

Buhr

Dugas

M. J.

(2006). Investigating the construct validity of intolerance of uncertainty and its unique relationship with worry. Journal of Anxiety Disorders, 20, 222–236.

Crites

(1978). Manual for the career maturity inventory for adults. Monterey, CA: CTB/McGraw-Hill.

Darcy

M. U.

Tracey

T. J.

(2007). Circumplex structure of Holland’s RIASEC interests across gender and time. Journal of Counseling Psychology, 54, 17–31.

10.

Dik

B. J.

Duffy

R. D.

(2009). Calling and vocation at work definitions and prospects for research and practice. The Counseling Psychologist, 37, 424–450.

11.

Dik

B. J.

Eldridge

B. M.

Steger

M. F.

Duffy

R. D.

(2012). Development and validation of the Calling and Vocation Questionnaire (CVQ) and Brief Calling Scale (BCS). Journal of Career Assessment, 20, 242–263.

12.

Dik

B. J.

Sargent

A. M.

Steger

M. F.

(2008). Career development strivings assessing goals and motivation in career decision-making and planning. Journal of Career Development, 35, 23–41.

13.

Duffy

R. D.

Allan

B. A.

Autin

K. L.

Bott

E. M.

(2013). Calling and life satisfaction: It’s not about having it, it’s about living it. Journal of Counseling Psychology, 60, 42–52. doi:10.1037/a0030635

14.

Duffy

R. D.

Allan

B. A.

Bott

E. M.

Dik

B. J.

(2014). Does the source of a calling matter? External summons, destiny, and perfect fit. Journal of Career Assessment, 22, 562–574.

15.

Duffy

R. D.

Allan

B. A.

Dik

B. J.

(2011). The presence of a calling and academic satisfaction: Examining potential mediators. Journal of Vocational Behavior, 79, 74–80.

16.

Duffy

R. D.

Bott

E. M.

Allan

B. A.

Torrey

C. L.

Dik

B. J.

(2012). Perceiving a calling, living a calling, and job satisfaction: Testing a moderated, multiple mediator model. Journal of Counseling Psychology, 59, 50–59. doi:10.1037/a0026129

17.

Duffy

R. D.

Dik

B. J.

(2013). Research on calling: What have we learned and where are we going? Journal of Vocational Behavior, 83, 428–436.

18.

Duffy

R. D.

Sedlacek

W. E.

(2007). The presence of and search for a calling: Connections to career development. Journal of Vocational Behavior, 70, 590–601.

19.

Dugas

M. J.

Gosselin

Ladouceur

(2001). Intolerance of uncertainty and worry: Investigating specificity in a nonclinical sample. Cognitive Therapy and Research, 25, 551–558.

20.

Frenkel-Brunswik

(1949). Intolerance of ambiguity as an emotional and perceptual personality variable. Journal of Personality, 18, 108–143.

21.

Furnham

Ribchester

(1995). Tolerance of ambiguity: A review of the concept, its measurement and applications. Current Psychology, 14, 179–199. doi:10.1007/BF02686907

22.

Gati

Krausz

Osipow

S. H.

(1996). A taxonomy of difficulties in career decision making. Journal of Counseling Psychology, 43, 510–526. doi:10.1037/0022-0167.43.4.510

23.

Gupta

Tracey

T. J.

Gore

Jr., P. A.

(2008). Structural examination of RIASEC scales in high school students: Variation across ethnicity and method. Journal of Vocational Behavior, 72, 1–13.

24.

Hartman

B. W.

Fuqua

D. R.

Hartman

P. T.

(1983). The construct validity of the career decision scale administered to high school students. Vocational Guidance Quarterly, 31, 250–258.

25.

Holland

J. L.

(1997). Making vocational choices: A theory of vocational personalities and work environments (3rd ed.). Odessa, FL: Psychological Assessment Resources.

26.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55. doi:10.1080/10705519909540118

27.

Muthén

L. K.

Muthén

B. O.

(2012). Chi-square difference testing using the Satorra–Bentler scaled chi-square. Retrieved from http://www.statmodel.com/chidiff.shtml

28.

Nauta

M. M.

(2007). Career interests, self-efficacy, and personality as antecedents of career exploration. Journal of Career Assessment, 15, 162–180. doi:10.1177/1069072706298018

29.

(2013). Seeing through the fog of ambiguity: Entrepreneurial judgments under decision settings of complexity. Human Systems Management, 32, 57–66.

30.

Osipow

S. H.

(1987). Career Decision Scale: Manual. Odessa, FL: Psychological Assessment Resources.

31.

Parsons

(1909). Choosing a vocation. Boston, MA: Houghton Mifflin Company.

32.

Sampson

Lenz

Reardon

Peterson

(1999). A cognitive information processing approach to employment problem solving and decision making. Career Development Quarterly, 48, 3–18. doi:10.1002/j.2161-0045.1999.tb00271.x

33.

Savickas

M. L.

Nota

Rossier

Dauwalder

J. P.

Duarte

M. E.

Guichard

… van Vianen

A. E. M.

(2009). Life designing: A paradigm for career construction in the 21st century. Journal of Vocational Behavior, 75, 239–250.

34.

Schere

J. L.

(1982). Tolerance of ambiguity as a discriminating variable between entrepreneurs and managers. Paper presented at the Academy of Management Proceedings. Briarcliff Manor, NY: Academy of Management.

35.

Schlomer

G. L.

Bauman

Card

N. A.

(2010). Best practices for missing data management in counseling psychology. Journal of Counseling Psychology, 57, 1–10. doi:10.1037/a0018082

36.

Stumpf

S. A.

Colarelli

S. M.

Hartman

(1983). Development of the career exploration survey (CES). Journal of Vocational Behavior, 22, 191–226.

37.

Tracey

T. J.

(2008). Adherence to RIASEC structure as a key career decision construct. Journal of Counseling Psychology, 55, 146–157.

38.

Tracey

T. J.

Darcy

(2002). An idiothetic examination of vocational interests and their relation to career decidedness. Journal of Counseling Psychology, 49, 420–427.

39.

Tracey

T. J.

Lent

R. W.

Brown

S. D.

Soresi

Nota

(2006). Adherence to RIASEC structure in relation to career exploration and parenting style: Longitudinal and idiothetic considerations. Journal of Vocational Behavior, 69, 248–261.

40.

Tracey

T. J.

Rounds

J. B.

(1993). Evaluating Holland’s and Gati’s vocational-interest models: A structural meta-analysis. Psychological Bulletin, 113, 229–246.

41.

Tracey

T. J.

Ward

C. C.

(1998). The structure of children’s interests and competence perceptions. Journal of Counseling Psychology, 45, 290–303.

42.

Triandis

H. C.

(1995). Individualism & collectivism. Boulder, CO: Westview Press.

43.

Hou

Tracey

T. J.

(2014). Relation of environmental and self-career exploration with career decision-making difficulties in Chinese students. Journal of Career Assessment, 22, 654–665. doi:10.1177/1069072713515628

44.

Hou

Tracey

T. J. G.

Zhang

(2016). Variations of career decision ambiguity tolerance between China and the United States and between high school and college. Journal of Vocational Behavior, 93, 120–128. doi:10.1016/j.jvb.2016.01.007

45.

Tracey

T. J.

(2014). The role of ambiguity tolerance in career decision making. Journal of Vocational Behavior, 85, 18–26. doi:10.1016/j.jvb.2014.04.001

46.

Tracey

T. J.

(2015a). Ambiguity tolerance with career indecision: An examination of the mediation effect of career decision-making self-efficacy. Journal of Career Assessment, 23, 519–532. doi:10.1177/1069072713515628

47.

Tracey

T. J.

(2015b). Career decision ambiguity tolerance scale: Construction and initial validations. Journal of Vocational Behavior, 88, 1–9.

48.

Zenasni

Besançon

Lubart

(2008). Creativity and tolerance of ambiguity: An empirical study. The Journal of Creative Behavior, 42, 61–73.

Variables	χ²	df	AIC	BIC	CFI	RMSEA		SRMR
	χ²	df	AIC	BIC	CFI	Estimate	90% CI	SRMR
Model a1. Three-factor model with original items	394.06	132	13,655.22	13,852.66	.83	.092	[.081, .102]	.08
Model a2. Four-factor model with original items	313.28	129	13,580.27	13,788.10	.88	.078	[.067, .089]	.07
Model a3. Three-factor model with new items	637.64	167	14,376.37	14,594.59	.77	.109	[.100, .118]	.10
Model a4. Four-factor model with new items	346.07	164	14,070.88	14,299.49	.91	.069	[.058, .079]	.06
Model a5. Four-factor hierarchical model	377.61	166	14,100.25	14,321.94	.90	.073	[.064, .083]	.08
Model b1. Criterion validity model	1,123.39	643	23,327.11	23,798.19	.90	.056	[.051, .062]	.07
Model c1. Relation with calling and adherence	422.70	219	13,231.20	13,472.79	.90	.065	[.055, .074]	.06