This study reports a statistical modification of a psychometric expatriate adjustment survey (e.g., the CernySmith Assessment or CSA) applied to expatriate missionary and humanitarian workers and their families. Earlier CSA survey items assisted in a need for rapid response clinical assessment originating from clinical observations, literature with a factor analysis resulting in 20 subscales and 5 qualitative content domains (Organizational, Cultural, Relational, Behavioral, and Personal). However, the assessment required more robust scale and factor analysis development to assure replicability. The present study analyzed 1,133 respondents working in 130 host countries. After various factor analytical iterations, a final CSA psychometric scale of 5 factors and 15 subscales (and a final 3 factor solution) emerged that compares with other known expatriate employee and spousal adjustment scales (Bhaskar-Shrinivas, Harrison, Shaffer, & Luk, 2005; Black, Mendenhall, & Oddou, 1991; Brown, 2008). This improved adjustment assessment answers Hippier, Caligiuri, and Johnson's (2014) call for robust expatriate adjustment assessments. The CSA moves adjustment measurement forward with new domains of resilience (Well-Being, Past Stresses, and Focus) and personal foundation (Spirituality, Health, and Habits). In light of 20-40% rate of preventable missionary attrition, the CSA should contribute to research, practice, and organizational support.

Introduction and Overview

Research in the cross-cultural adjustment tradition has identified factors and assessments that would improve work/task, interaction/relationships, and personal/family adjustment (Pinto, Cabral-Cardoso, & Werther, 2012; Ren, Shaffer, Harrison, Fu, & Fodchuk, 2014). Considering this direction, the goal has been to identify significant predictive variables, to determine what is meant by outcomes and effectiveness, and to develop adjustment measurements. Moreover, effectiveness traditionally implied work, relationship, and cultural functioning. Consequently, researchers expanded "adjustment" to include mental health adjustment, cultural adjustment, relationship adjustment, and family adjustment (Bhaskar-Shrinivas, Harrison, Shaffer, & Luk, 2005; Black, Mendenhall, & Odou, 1991).

With expatriate global mobility expected to increase by 47% by 2016 and attrition rates averaging 30% to 50% as of 2014 (Brookfield, 2014), the need for accurate measures and models is imperative. Surprisingly, the literature offers relatively little consideration to models that consider the dynamic flow of arrival, on-site support and monitoring, and return preparation (Moran, Harris, & Moran, 2016). Family and missionary adjustment receive even less attention (Van Meter, 2003). Consequently, recent recommendations for improved models abound (Haslberger, Brewster, & Hippier, 2013; Hechanova, Beehr, & Christiansen, 2003; Hemmasi, Downes, & Varner, 2010; Hippier, Caligiuri, & Johnson, 2014; Ren et al., 2014; Van Meter, 2003).

The purpose of this study is to present data supporting development of a comprehensive adjustment instrument designed for on-site self-discovery, mentoring, and coaching applications. The CernySmith Assessment (Cerny & Smith, 2005) original instrument served a purpose but requires reanalysis to address standards of a comprehensive adjustment measurement (Hippier, Caligiuri, & Johnson, 2014). This study presents supporting data to assure CSA replicability through factor analytic and reliability statistical techniques.

Review of Literature

Expatriate self-rating items typically survey psychological and sociological adjustment for everyday social situations and tasks in a new culture. The decades of this research identify a wide range of variables, models, and assessment with the goal to understand and predict expatriate performance and adjustment. The following review provides examples from expatriate research organized into orienting categories from vast and sometimes confusing investigations.

Expatriate Adjustment as Work and Outcome Performance

One early concern in the expatriate adjustment research concerned performance, perhaps driven partly by multicultural organization expansion and the rise of corporate expatriates. Black and Stephens' (1989) well-known 14-item cross-cultural adjustment measured three factors: work, interaction, and general culture. The three factor model unquestionably shaped theory development and measurement and also influenced the convergence of a unified cross-cultural adjustment theory (Black & Gregersen 1991; Black et al., 1991; Hechanova et al., 2003; Mendenhall, Kuhlmann, Stahl, & Osland, 2002).

But their model also launched additional cross-cultural adjustment research attempting to deepen understanding of the adjustment correlation with job dimensions. In a Swedish expatriate manager study, Anderzen and Arnetz (1997) added stress hormones, eleven psychosocial variables, aspects of mental well-being, and work adjustment to Black and Stephens' (1989) list of factors. Ward and Rana-Deuba's (1999) Acculturation Index measured host and co-national identification and four acculturation modes: integration, separation, marginalization, and assimilation. Caligiuri (2000) compared expatriate work performance with supervisor-rated interview measures to index employee performance and their desire to terminate assignments. Navara and James (2002) compared missionaries to non-missionary organizations, measuring their adjustment with a modified version (adding satisfaction) as one of Black and Stephens' (1989) adjustment ratings.

Additional examples illustrate new directions regarding how stress affects adjustment and work. The Bhaskar-Shrinivas et al. (2005) meta-analysis of over 50 determinants and outcomes of expatriate adjustment explored data from 8,474 expatriates in 66 studies. Their results supported the traditional adjustment model (Black et al., 1991) but highlighted the link between adjustment and cost benefits to companies and their expatriates. From the perspective of adjustment theory development, their analysis broadened the adjustment paradigm. They labeled this expansion the stressor-stress-strain model. Adjustment conceptualizations now included expatriate comfort or absence of stress to the traditional three factors: (a) Cultural (non-work factors like general living conditions, local food, transportation, entertainment, facilities, and health care services in the host country), (b) Interaction (interacting with host country nationals both inside and outside of work), (c) Work (job assignment), and now (d) Expatriate comfort/stress. Further studies included withdrawal intentions (from assignment, from organization, from occupation), task adjustment, career development, HQ-subsidiary coordination, assignment completion, professional/skill development, controlling the subsidiary, overall assignment effectiveness, and assignment satisfaction (Pinto et al., 2012; Bonache, 2005; Carmeli, 2005).

Family and Spousal Correlation with Expatriate Adjustment

Despite the hundreds of scholarly studies in adjustment, family and spousal adjustment researchers have only uncovered the centrality of spouse and family in more recent investigations. While earlier research minimally included spouse and family (Black & Gregersen 1991; Black & Stephens 1989; Shaffer & Harrison 1998), researchers like Brown (2008) introduced spousal and couple adjustment as a significant determinant of expatriate completion, adjustment, and performance. His list of 15 stressors sorted into four factors: (a) Reduced Self (feeling less valued, competent, or uncertain about the future), (b) Relationship Strains (decline in partner relationship, dealing with partner's disappointment, too many demands or expectations, insufficient time with partner), (c) Local Pressures (daily living challenges such as driving, shopping, engaging culture, concerns over health, safety and security), and (d) Isolation (no close friends, feeling isolated and cut off, disappointment in assignment benefits). Furthermore, Brown (2008) included write-in stressors on isolation (i.e., aspects of local culture, demands of work and working hours, stress from missing family and friends). Spousal adjustment finds support with other examples, such as the qualitative report by Gupta, Banerjee, and Gaur (2012) which classified expatriate failure with spousal adjustment. Also, new measurements in family adjustment dimensions have emerged (Rosenbusch & Cseh, 2012). As demonstrated subsequently, the CSA includes and refines family and spouse dimensions.

Mission Research on Expatriate Adjustment

Published missions-related research on expatriate adjustment and assessment appears limited. The need has been discussed for a long time. For instance The World Evangelical Fellowship (WEF) reported nearly two decades ago that 71% of attrition in missionaries was preventable (Taylor, 1997). Subsequent WEF-related research of 592 global missions agencies (Van Meter, 2003), including 555 U.S. agencies representing over 38,000 missionaries in 22 nations, was consistent with the 1997 research and reported a 20.5% average preventable loss, which increased to over 40% among non-denominational agencies. Identified causes include: family, marital discontent, relationship problems, organizational support (e.g., selection screening, finances, performance reviews, on-field management, danger calculation, clear objectives, member care, children's education, and home-church involvement), meeting psychological health criteria, personal satisfaction, and effective communication with sponsoring organization.

The beginning theoretical background conceptualized the problems associated with missionary stability and sticking ability in international humanitarian and missions work (Dye, 1974; Foyle, 1987; Gish, 1983; and Williams, 1988). Further conceptual work by Kelly and Michele O'Donnell on understanding and managing missionary stress should be especially noted (O'Donnell & O'Donnell, 1992). A pilot test of a stress assessment (CHOPS 100) identified 10 missionary stressors (O'Donnell, 1993). This identification encouraged further testing, item development, and operationalization resulting in a preliminary psychometric assessment (initially called CSAI; Cerny & Rosenbusch, 2010; Cerny & Smith, 2005; Cerny, Smith, Ritschard, & Dodd, 2007). However, these efforts were under reported academically and required additional rigorous methodological work along with recent theory developments.

Holistic and Multivariate Approaches to Expatriate Adjustment

However meaningful these foundational models were, their limits prompted more comprehensive multivariate and holistic models (Haslberger et al., 2013; Hippier et al., 2014). Some of these include: the person-environment needs (P-E); novelty of culture; family; organizational support; psychodynamic qualities in global adjustment; a stressor-stress-strain paradigm highlighting adaptation stressors; and the role of cognitive, affective, and behavioral strain implications. They also showed global demands, uncertainties, negative outcomes of withdrawal, poor performance, early return, and international assignment dissatisfaction. Further comprehensive expatriate adjustment modeling is illustrated by Lee and Kartika's (2014) regression model correlating individual factors (emotional intelligence, cultural intelligence, experience abroad), family factors (family support, family adaptability, parental demands, family to work conflict), social capital (social capital, mentoring behavior), organizational support, and a psychology contract to predict expatriate adjustment and performance ([R.sup.2] = .708, p = .000).

The inclusion of social relationships and mentoring also appeared in the adjustment outcomes from Zhuang, Wu, & Wen (2013). Haslberger et al. (2013) meta-analyzed multiple dimensional models of cognitive, affective, and behavioral outcomes typical of previous research, but expanded to build on P-E fit theory (person-vocation, person-job, person-organization, person-group, person-supervisor fit). They found a path analytic approach of culture novelty, work and non-work adjustments along with other cognitive/ affective/behavioral antecedents (i.e., adjustment tension between external demands and internal needs supply) to predict performance, satisfaction, and withdrawal. Similarly, Ren et al. (2014) tested various cross-cultural demands (cultural novelty and cultural value distance, host country language deficiency), proactive tactics (information seeking, relationship building, and positive framing), and a unique concept called embeddedness (breadth and depth of connectedness in one's foreign posting) to determine retention. Their findings from bivariate correlations and a holistic path analysis yielded a significant adjustment and retention prediction.

This comprehensive emphasis of adjustment and retention theory represents a shift forward. For instance, Firth, Chen, Kirkman, and Kim (2014) concluded that future research needs "to adopt, and rigorously test, a motivational process perspective in which expatriates engage in strategies that enable them to overcome contextual and personal challenges" (p. 296). These trends predict more robust psychometric dimensions and adaptation items which remain consistent with known instruments while comprehensively adding improved definitions and variables/ factors on person-environment (P-E), marital and family, organizational leadership, social networking, mentoring, differential factor weightings, and more exhaustive predictions of performance and personal adjustment (Haslberger et al., 2013; Hippier et al., 2014).

Expatriate Adjustment Measurement

Research focused on instrumentation characterizes another part of the cross-cultural adjustment literature. This research emphasis is illustrated in analyses such as the correlation of 250 variables with competence highlighting spousal adjustment, expatriate failure (Spitzberg, 2015; Spitzberg & Changnon, 2009), state-of-the-art conceptualizations and assessment issues (Deardorff, 2015), dialectical interpersonal-intercultural cognitive and affective dimensions (Kealey, 2015; Martin, 2015; Spitzberg, 2015;), contextual analysis, power analysis, identity struggles, language, relationship synchrony, multi-faceted personality traits, values, motivations, and co-creation in social networks (Chi & Suthers, 2015; Kim, 2015).

Essays directed at critiquing measurements from the intercultural competency literature reported some 100 intercultural competence instruments. However, many of these measures often are limited, lacking a range of the most correlated adjustment variables, revealing few items that help diverse teams make holistic assessments, and offering inadequate relationship analysis. Moreover, the critique of limited scales highlights overly simplistic research designs (i.e., need enhanced design involving control groups, performance-based measures, little practical use related to improvements, and inaccessibility related to proprietary ownership (Deardorff, 2015). Kealey (2015) lamented the lack of predictive validity and the potential vulnerability of self-reports. Martin (2015) acknowledged limits of the popular affect, behaviors, cognition model and proposed future measurements to include holistic, relational-dialectic interaction, and spiritual views.

The adjustment literature overview in these examples outlines development in the expatriate adjustment research. First, the literature identified adjustment as an outcome performance issue and contributed a three factor model to define outcomes. Second, the research expanded to include socio-psychological variables, stress, and psychodynamic issues. Third, still another set of variables included spouse and family, organizational support, cultural demands and distance, and numerous dyadic and relationship dimensions. Fourth, the literature review pointed to the scarcity of published missions adjustment studies and the need for rigorous methodology. Fifth, researchers incorporated more complex multivariate adjustment models and identified concerns for intensifying measurements. In sum, this brief overview underscores the need enhanced research, better theory and modeling, and enriched measurements that establish more confidence and yet inspire practical usefulness. This final point explains the reason for the current research designed to evaluate and test a promising adjustment instrument.

Rationale

There are two reasons prompting this study. One is to refine instrumentation as noted above, particularly as Hippier et al. (2014) advocated more detailed scale development and stated that this was "the way forward." The present study analyzes the psychometric properties and dimensions of a renewed CSA and discusses its convergence with other known adjustment instruments. Second, the present study not only addresses overall trends for research advancements (Hemmasi et al., 2010; Haslberger et al., 2013; Hippier et al., 2014; Ren et al., 2014) but concentrates on assessing missionary transition, monitoring and attrition (Van Meter, 2003). Since the missions adjustment literature is incomplete (see the literature review above), this emphasis itself is compelling.

Considering the research and expatriate issues outlined, the present researchers decided that a rigorous psychometric analysis of the CSA would provide a basis for its use in further research and clinical application. The primary research question for the present study was: to what degree do the content derived scales of the CSA empirically identify factors consistent with prior measures and the framework of a stressor-stress-strain model of cross-cultural adjustment? Our approach followed the psychometric methodology employed by Hemmasi et al. (2010). A unidimensional set of items was identified for each scale. Then the items for all of the scales within each domain were analyzed to determine the degree of discriminant validity. Finally, all of the domain scales were analyzed to determine the relationship among domains. This re-analysis is expected to produce an efficient scale for measuring psychometric properties and dimensions of cross-culture adjustment, all aligned with accepted and known measures.

Methodology

Respondents and Data Collection

During a 19-month period, 1,133 expatriate employees, family members, or students in 130 host countries from 46 passport countries responded to 87 items in an online survey. Respondents were 47% male and 53% female, ages 20-77 (median 42). The current cross-cultural assignment time averaged 5 years with a total field experience mean of 13 years. Educational levels included 17% pre-Bachelor's degree, 38% Bachelor's degree, and 44% post-Bachelor's degree. The 35% single and 65% married respondents represented 87% non-profit, 7% business, 4% student, and 1% diplomatic. Participants were 86% from North America, 6% from Western Europe, and 8% from other countries.

Instrument Development and Testing Procedures

As an online assessment platform accessible to employees, family members, and coaches, the CSAI (as it was first called) developed from early missions-centered conceptual frameworks identified above. The initial items were subjected to Principle Components Factor Analysis resulting in 20 scales within five domains: Organizational (organizational leadership, ethos, management, and work demands), Cultural (adjustment to situational culture, transitions, and safety), Relational (relationship with nationals, expats, family, and friends), Behavioral (perceptual, cognitive, and affective internal regulation and buoyancy), and Personal (basic support of spirituality and faith, physical health, and habits). These pilot efforts stimulated subsequent confirmation research linking the 20 scales within five domains with predicted adjustment stressor impact (Rosenbusch, Cerny, & Earnest, 2015; Rosenbusch & Cseh, 2012).

In this original set of 87 items, an initial 8 items were judged as objectionable for a self-discovery and coaching application (e.g., unrealistic fears, having the same thought over and over, experiencing strange thoughts, worrying about a nervous breakdown, fear of hurting oneself, things feeling unreal, having thoughts of suicide, and sexual concerns.). The remaining 79 items were included in the present study. Based on their content, the developers divided the items into 15 scales which in turn were grouped into the 5 a priori domains. The Likert items, considered a "state" rather than "trait," assumed that an expatriate's level of functioning will vary across time depending on perception of personal factors (e.g., health) and/or situational factors (e.g., a change in administration).

Procedures and Methods of Data Analysis

The CSA is organized with selected items assigned to selected scales grouped within specified domains. That is, the CSA, as currently implemented, has a specified "measurement model." In turn, a 2-stage procedure tested the CSA measurement model. In stage one, an Exploratory Factor Analysis (EFA), with each of the 15 scales, determined if the scales were unidimensional. A scale was considered unidimensional if all items loaded above 0.40 and the set of items had one large eigenvalue with the remaining values below 1.0. Hemmasi et al. (2010) used the same approach analyzing sets of cross-cultural adjustment items. In the second stage, we examined the validity of the CSA organization of items into scales within each domain by comparing 3-factor solutions derived from different subsamples.

Our approach used EFA, rather than Confirmatory Factor Analysis (CFA), following the recommendations of Asparouhova and Muthen (2009), Browne (2001), and Lorenzo-Seva and ten Berge (2006). Asparouhova and Muthen (2009) observed that using CFA has disadvantages when there is question of replicating a final model. Since we were testing a modified version of the CSA, we accepted the Asparouhova and Muthen (2009) recommendation that a well-fitting measurement model is better carried out by EFA. Similarly, Browne (2001) advocated EFA rather than CFA approaches in situations of model uncertainty stating, "The discovery of mis-specified loadings ... is more direct through rotation of the factor matrix than through the examination of model modification indices" (p. 1020). Finally, Lorenzo-Seva and ten Berge (2006) recommended testing the replicability of EFA-derived factors across subsamples as a way of confirming the stability of the EFA derived factors. They recommended using Tucker's Coefficient of Congruence to empirically test the similarity of factors derived from different subsamples.

In each domain, we extracted the number of factors equal to the number of scales in that domain. In each case, three factors were extracted. We did separate EFAs for male and female respondents and for two random 50% subsamples. We calculated the Tucker's Congruence Coefficient to determine the degree of replicability between the factor pattern loadings across the samples. For solutions that had a high degree of correspondence (>.90), we then performed an EFA on the total sample.

Finally, we created scale scores for each scale within each domain and performed an EFA of these 15 scales. In each analyses, the method of factor extraction for the EFA was Principle Components Analysis (PCA).

Results

The first inquiry in the process asked, "Do the items on a given scale form a single factor?" We calculated the first principle component of each set of items for each of the 15 scales. The 15 PCA analyses of the item clusters confirmed that each cluster formed a single scale. The scree plots of the eigenvalues indicated one main cluster for each scale. We dropped a total of 11 items because they did not have loadings above 0.40. The 11 omitted items were: inadequate technical support, inadequate pay, lacking education or training, male/female role expectations, government red tape, not practicing religious disciplines, tensions related to belief system, loss of needed support, communication problems, social opposition in one's host country, and social instability. This process reduced the total items to 68.

Using these 68 items, each scale was subject to a second PCA. The loadings of the items on the first princi pal component were above 0.50 in all cases. Table 1 reports the number of items and Cronbach's Coefficient Alpha for each of the 15 scales within the five domains (Alpha ranges = 0.61 to 0.85). The Alpha coefficients for each of the domain scales are also given.

We conducted a PCA for the sets of items within each of the five domains. Since each domain in the CSA had three subscales, we extracted the first three principle components and subjected them to oblique rotation. To test the stability of the factor solutions we compared factors derived from different subsamples. First, this was done for the male and female samples separately, which allowed us to examine the replicability of factors across gender. Then the sample was randomly split in half using the SPSS sampling procedure, and PCAs were conducted within each half sample to test the stability of the factors. The Tucker Congruence Coefficients measuring the correspondence between the three factor solutions derived from the male and female samples and the 50% split-samples were:

           Male & Female

Domain    F1     F2     F3

1        0.92   0.98   0.90
2        0.94   0.95   0.92
3        0.95   0.94   0.85
4        0.59   0.84   0.06
5        0.96   0.98   0.87

          Half 1 & Half 2

Domain    F1     F2     F3

1        0.97   0.97   0.94
2        0.98   0.96   0.93
3        0.99   0.98   0.96
4        0.50   0.41   0.86
5        0.99   0.95   0.95

Regarding the desired value of the Congruence Coefficient, Lorenzo-Seva and ten Berge (2006) provide the following guidelines: "Our results suggest that a value in the range .85-.94 corresponds to a fair similarity, while a value higher than .95 implies that the two factors or components compared can be considered equal" (p. 57). For four of the five domains, the 3-factor solutions had very high correspondence across the subsamples, such that they can be considered equivalent. In the Resilience Domain (4), only one of the three factors had acceptable correspondence. We examined the loadings of items in the 3-factor solution and found that three items were unstable across the subsamples. These items were: Troubling Dreams, Making Decisions, and Fear of Failure. When these three items were omitted, the correspondence coefficients for the three factors derived from the remaining 10 items had very high correspondences coefficients:

                Male & Female        Half 1& Half 2

Domain        F1     F2      F3     F1     F2      F3

Modified 4   0.99   -0.98   0.98   0.99   0.98    0.96

Given the high degree of replicability across sub-sample analyses, we performed PCAs within each domain using the entire sample. We extracted three factors and subjected them to oblique rotation. The results of these five analyses are shown in Tables 2-6. Item loadings in three of the domains (Organizational, Cultural, and Foundational) showed 100% correspondence between their pattern of coefficients and the a priori assignment of items to domain scales. The pattern of coefficients was above 0.50 for all but three items, which were above 0.40. For scale items in the Relational domain, all but one of the 15 items fit the pre-designated 3-scale clustering. For the above four domains, the scree test of the eigenvalues from the EFA supported the 3-scale structure.

The Resilience domain had three items that did not consistently load on the same component cluster across subsamples. The other nine items did form a very stable 3-factor solution across subsamples. The three unstable items loaded either on the Well-being factor or the Focus factor. Table 5 shows the loadings for all 12 items derived from the PC A with the entire sample. The inclusion of the three items does not affect the loading pattern of the other nine items. The Resilience domain could use further development to establish a stable set of items.

Tables 2-6 report the PC A results of item loadings for the three scales in each of the five a priori domains.

Finally, in Phase three, the 15 scale scores were subjected to a secondary principal components analysis. The scale scores were calculated by averaging the ratings on the items within each scale. The PCA result of the 15 cluster scales is reported in Table 7. The scree test suggested that three components were appropriate. Three components were extracted and subjected to oblique rotation. The three components mostly corresponded to the Organization domain, a combination of the Cultural and Relational domains, and a combination of the Resilience and Foundational domains, respectively. Therefore, the resulting three factors could be broadly interpreted to identify work management, relationship management, and self-management aspects of cross-cultural adjustment.

Discussion and Conclusions

The results revealed that the Principle Components Analyses (PCA) of 68 items yielded 15 unidimensional constructs comprising scales of expatriate adjustment. The PCAs within domains demonstrated substantial empirical convergence with the a priori categorizing of items on scales within each of the functioning domains: Organizational, Cultural, Relational, Resilience and Foundational. The 15 adjustment scales may be used independently or in combination in future research. Taken together, the scale and domain structure of the CSA provide a rationally derived and empirically supported map in a single comprehensive instrument. The PCA analysis in Phase 3, clustering the 15 scale measures, was essentially a second-order PCA resulting in three factors broadly interpreted as work management, relationship management, and self-management aspects of cross-cultural adjustment.

One focus for this study was comparison with previous models and a research need for improved measurement. Figures 1 and 2 visually and conceptually align the CSA with other known models and instruments. Future research will need to add appropriate statistical validity to such comparisons. The Bhaskar-Shrinivas et al. (2005) comprehensive model (Figure 1) and instrument in their stressor-stress-strain formulation is an underlying paradigm for the current CSA assumptions, which produced such outcomes as wellness, resilience, and retention. Figure 2 depicts CSA domains and items conceptually aligned with two other standard models (Black et ah, 1991; Brown 2008). Two of the three clusters identified in this analysis, work management and relationship management, appear to be inclusive of the three primary factors identified by the original Black model. This involves alignment with work management, corresponding to Black's Work factor with the CSA Cultural domain and parallel to Black's General (culture) factor. The scales within the CSA Relational domain are inclusive but broader than the Black model's Interaction factor, which is limited to interaction with host nationals. The broader relationship focus for the CSA includes domestic and national colleagues, family, and friends.

The CSA self-management cluster has no parallel with Black's three primary factors, and thus we consider it an additional construct for ongoing consideration and research. The CSA's Resilience and Foundational domains include scales measuring emotional, cognitive, and physical stresses along with psychological vulnerability, habit patterns, and spirituality. An exception is that some questions from the Health scale of the CSA Foundational domain parallel health resource questions in the Black model's General (cultural) factor. The CSA self-management cluster, however, includes positive parallels to Brown's (2008) Reduced Self scale, while the CSA relationship management cluster offers meaningful parallels to Brown's (2008) Local Pressures, Isolation, and Relationship Strains scales. The only overlap between Brown (2008) and Black and Stephens' (1989) expatriate scales is the parallel between Brown's Local Conditions and Black's General Conditions (cultural) scale as well as a smaller overlap between Brown's Isolation and Black's Interaction scale.

The Black and Brown models utilize survey questions developed by university-based business consultants in international management. The CSA link to those items (Figure 2) and the Bhaskar-Shrinivas et al. (2005) model jointly suggest a comparison validity framework. Differences noted are that the CSA survey questions were developed by clinical psychologists who were providing coaching primarily for international non-profit groups and missionaries in the field with the goals of increasing well-being and reducing failed assignments. This perspective explains the self-management content focus of CSA questions in addition to the work management and relationship management aspects. Moreover, with the current study offering a more robust assessment, CSA future reports will rely on these findings while continuing to use easy-to-understand, nonjudgmental language, balanced strength and challenge formats, multivariate avenues to self-discovery, mentor-friendly feedback and discussion, and online accessibility. The assessment has meaningful factor and reliability structures, which also meet recent theory trends that are advancing adjustment theory, measurement, and practice (Hippier et al., 2014; Ren et al., 2014; Zhuang et al., 2013; Firth et al., 2014). The CSA could be a valuable tool for organizations to recognize complex personal adjustment in the field and thus circumvent problems before they spiral out of control.

As with other research studies of this type, several limitations occur. First, the single use at a point in time results in a snapshot of perceived experience without accounting for self-report validity issues or individual response styles. A second limitation is that the CSA had been field tested before these comprehensive analyses were conducted, and so researchers were working with a fixed data set. The results modified the domains and scales, which future research should pursue as we recommend more items in future scales. The redevelopment presented here of the 15 adjustment scales anticipates further research for more validity testing. Also, further research needs to test differences in CSA adjustment patterns among employees, spouses, and children and to uncover influences of pre-field training, time on-assignment, total cross-cultural experience, and difficulty of setting. Another need is that future research should provide confirmatory factor analysis in order to substantiate the factor structures presented in this study. We suggest that a program that implements assessment within a coaching/consultation model could leverage expatriate adjustment strengths (positive feedback) and identify and reduce adjustment challenges (focused feedback and coaching/scaffolding to coping efficacy) before such challenges become distress that erodes well-being, job satisfaction, and job performance.

In sum, moving to work in a foreign country often involves simultaneous changes in many aspects of an expatriate's life, including unusual sources of stress that impact functioning and retention. When adjustment fails, resulting psychological and economic costs are high. Our hope is for a comprehensive assessment of adjustment with useful feedback to guide expatriate functioning, early intervention, and support. This assessment would be most useful during the initial cross-cultural engagement period when the risks are profound. Finally, periodic assessments could provide "early warning" of changes in the expatriate's functioning.

Keith J. Edwards

Biola University Rosemead School of Psychology

Carley H. Dodd

Abilene Christian University

Katherine H. Rosenbusch

George Mason University

Leonard J. Cerny II

Missionary Care International and CernySmith, LLC

References

Anderzen, I., & Arnetz, B. B. (1997). Psychological reactions during the first year of a foreign assignment: Results of a controlled longitudinal study. Work and Stress, 11(4), 304-318.

Asparouhova, T., & Muthen, B. (2009). Exploratory structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal, 16(3), 397-438.

Bhaskar-Shrinivas, P., Harrison, D. A., Shaffer, M. A., & Luk, D. M. (2005). Input-based and time-based models of international adjustment: Meta-analytic evidence and theoretical extensions. Academy of ManagementJournal, 48, 257-281.

Black, J. S., & Gregersen, H. B. (1991). The other half of the picture: Antecedents to spouse cross-cultural adjustment. Journal of International Business Studies, 22(3), 461-477.

Black, J. S., Mendenhall, M., & Oddou, G. (1991). Toward a comprehensive model of international adjustment: An integration of multiple theoretical perspectives. Academy of Management Review, 16, 291-317.

Black, J. S., & Stephens, G. K. (1989). The influence of the spouse on American expatriate adjustment and intent to stay in Pacific Rim overseas assignments. Journal of Management, 15, 529-544.

Bonache, J. (2005). Job satisfaction among expatriates, repatriates and domestic employees: The perceived impact on international assignments on work related variables. Personnel Review, 34, 110-124.

Brookfield Global Relocation Services. (2014). Global mobility trends survey 2014. New York: Brookfield GRS.

Brown, R. J. (2008). Dominant stressors on expatriate couples during international assignments. The International Journal of Human Resource Management, 19, 1018-1034.

Browne, M. W. (2001). An overview of analytic rotation in exploratory factor analysis. Multivariate Behavioral Research, 36, 111-150.

Caligiuri, P. M. (2000). The big five personality characteristics as predictors of expatriate's desire to terminate the assignment and supervisor-rated performance. Personnel Psychology, 53, 67-88.

Carmeli, A. (2005). The relationship between organizational culture and withdrawal intentions and behavior. International Journal of Manpower, 26, 177-195.

Cerny, L., & Rosenbusch, K. (2010, November). Impact of family system variables on intercultural adjustment. Paper presented at Mental Health and Missions Conference, Angola, IN.

Cerny, L. J., & Smith, D. S. (2005, March). CernySmith adjustment index: Research and development. Paper presented at Research Colloquium, Rosemead Graduate School of Psychology, Biola University, La Mirada, CA.

Cerny, L. J., Smith, D. S., Ritschard, H., & Dodd, C. H. (2007, March). The CSAI: An expatriate on-field adjustment index to measure intercultural intelligence. Paper presented to the Families in Global Transition Conference, Houston, TX.

Chi, R., & Suthers, D. (2015). Assessing intercultural communication competence as a relational construct using social network analysis. International Journal of Intercultural Relations, 48, 108-119.

Deardorff, D. K. (2015). Intercultural competence: Mapping the future research agenda. International Journal of Intercultural Relations, 48, 3-5.

Dye, W. (1974). Stress producing factors in cultural adjustment. Missiology, 2, 67-77.

Firth, B. M., Chen, G., Kirkman, B. L., & Kim, K. (2014). Newcomers abroad: Expatriate adaptation during early phases of international assignments. Academy of Management Journal, 57(1), 280-300. doi: 10.5465/amj.2011.0574

Foyle, M. (1987). Overcoming missionary stress. Wheaton, IL: Evangelical Missions Information Service.

Gish, D. (1983). Sources of missionary stress. Journal of Psychology and Theology, 11, 238-242.

Gupta, R., Banerjee, P., & Gaur, J. (2012). Exploring the role of the spouse in expatriate failure: A grounded theory-based investigation of expatriate's spouse adjustment issues from India. International Journal of Human Resource Management, 23(17), 3559-3577. doi: 10.1080/09585192.2011.645052

Haslberger, A., Brewster, C., & Hippier, T. (2013). The dimensions of expatriate adjustment. Human Resource Management, 52(3), 333-351. doi:10.1002/hrm.21531

Hechanova, R., Beehr, T. A., & Christiansen, N. D. (2003). Antecedents and consequences of employees' adjustment to overseas assignment: A meta-analytic review. Applied Psychology: An International Review, 52,213-236.

Hemmasi, M., Downes, M., & Varner, I. (2010). An empirically driven multidimensional measure of expatriate success: Reconciling the discord. The International Journal of Human Resource Management, 21, 982-998.

Hippier, T., Caligiuri, P., & Johnson, J. (2014). Revisiting the construct of expatriate adjustment. International Studies of Management & Organization, 44(3), 8-24. doi:10.2753/IM000208825440301

Kealey, D. J. (2015). Some strengths and weakness of 23 years of research on intercultural communication competence: Personal reflections. International Journal of Intercultural Relations, 48, 14-16.

Kim, Y. Y. (2015). Achieving synchrony: A foundational dimension of intercultural communication competence. International Journal of Intercultural Relations, 48, 21-31.

Lee, L., & Kartika, N. (2014). The influence of individual, family, and social capital factors on expatriate adjustment and performance: The moderating effect of psychology contract and organizational support. Expert Systems with Applications, 41, 5483-5494. doi: 10.1016/j.eswa.2014.02.030

Lorenzo-Seva, U., & ten Berge, J. M. F. (2006). Tucker's Congruence Coefficient as a meaningful index of factor similarity. Methodology, 2(2), 57-64.

Martin, J. N. (2015). Revisiting intercultural communication competence: Where to go from here. International Journal of Intercultural Relations, 48, 6-8.

Mendenhall, M., Kuhlmann, T., Stahl, G., & Osland, J. (2002). Employee development and expatriate assignments: A review of the expatriate adjustment theory literature. In M. Gannon, & K. Newman (Eds.), Handbook of cross-cultural management (pp. 155-183). Thousand Oaks, CA: Sage.

Moran, R. T., Harris, P. R., & Moran, S. (2016). Managing cultural differences (9th ed.), New York, NY: Elsevier.

Navara, G. S., & James, S. (2002). Sojourner adjustment: Does missionary status affect acculturation? International Journal of Intercultural Relations, 26, 695-709.

O'Donnell, K. (1993) Personal communication to Leonard Cerny listing 10 areas of missionary stress under consideration.

O'Donnell, K., & O'Donnell, M. (1992). Understanding and managing stress. In K. O'Donnell, K. (Ed), Missionary care: Counting the cost of world evangelization (pp. 110-122). Pasadena, CA: William Carey Library.

Pinto, L. H., Cabral-Cardoso, C., & Werther, W. B. (2012). Adjustment elusiveness: An empirical investigation of the effects of cross-cultural adjustment on general assignment satisfaction and withdrawal intentions. International Journal of Intercultural Relations, 36(2), 188-199.

Ren, H., Shaffer, M. A., Harrison, D. A., Fu, C., & Fodchuk, K. M. (2014). Reactive adjustment or proactive embedding? Multistudy, multiwave evidence for dual pathways to expatriate retention. Personnel Psychology, 67( 1), 203-239. doi: 10.1111/peps. 12034

Rosenbusch, K., Cerny, L. J, & Earnest, D. (2015). The impact of stressors during international assignments. Cross Cultural Management, 22(3), 405-430. doi: 10.1108/CCM-09-2013-0134

Rosenbusch, K., & Cseh, M. (2012). The cross-cultural adjustment process of expatriate families in a multinational organization: A family system theory perspective. Human Resource Development International, 75(1), 61-77. doi: 10.1080/13678868.2011.646895

Shaffer, M., & Harrison, D. (1998). Expatriates' psychological withdrawal from international assignments: Work, non-work and family influences. Personnel Psychology, 51, 87-118.

Spitzberg, B. H. (2015). Is past prologue, or just passed and lacking presence? International Journal of Intercultural Relations, 48,24-26.

Spitzberg, B. H., & Chagnon, G. (2009). Conceptualizing intercultural communication competence. In D. K. Deardorff (Ed.), The SAGE handbook of intercultural competence (pp. 2-52). Thousand Oaks, CA: Sage.

Taylor, W. (1997). Too Valuable to Lose: Exploring the Causes and Cures of Missionary Attrition Pasadena, CA: William Carey Library.

Van Meter, J. (2003). US report of findings on missionary retention. Paper developed through steering committee of the High Leigh Conference. London, England. Retrieved from http://www.worldevangelicals.org/resources/rfiles/res3_95_link_1292358708.pdf.

Ward, C., & Rana-Deuba, A. (1999). Acculturation and adaptation revisited .Journal of Cross- Cultural Psychology, 30, 422-442.

Williams, K. (1988). Worksheet for balanced living. In K. O'Donnell, K. & M. O'Donnell (Eds.), Helping missionaries grow: Readings in mental health and missions (pp. 390-398). Pasadena, CA: William Carey Library.

Zhuang, W., Wu, M., & Wen, S. (2013). Relationship of mentoring functions to expatriate adjustments: Comparing home country mentorship and host country mentorship. International Journal of Human Resource Management, 24(1), 35-49. doi:10.1080/095851 92.2012.669784

Author Note: Disclosure Statement: CernySmith LLC has annual contracts with approximately 60 international humanitarian, mission, and business organizations providing unlimited use of this assessment for coaching employees and family members. This assessment and component scales may be freely published and used for research purposes.

Correspondence concerning this article should be addressed to Carley H. Dodd, Department of Intercultural Communication, Abilene Christian University, Box 28156, Abilene, TX 79699, USA. Phone: T325-674-2293. Email: [email protected]

Author Information

EDWARDS, KEITH J. PhD. Address;: 13800 Biola Avenue, La Mirada, CA 90639. Email: [email protected]. Title: Professor of Psychology. Degrees: B.Ed. (Mathematics), University of Wisconsin, Whitewater; MA (Research Methods), New Mexico State University; PhD (Research methods), New Mexico State University; PhD (Clinical and Social Psychology), University of Southern California. Specializations: Marital Therapy, Emotion-Focused Therapy, Interpersonal Neurobiology of Experience, Spiritual Functioning, Research Methods, Cross-Cultural Adaptation.

DODD, CARLEY H. PhD. Address: ACU Box 28156 Abilene, TX 79699. Email: [email protected]. Title: Professor of Intercultural Communication. Degrees: BA (Bible, Speech Communication), Abilene Christian University; MA (Communication), Abilene Christian University; PhD (Communication), University of Oklahoma. Specializations: Intercultural Communication and Adaptation, Social Change, Interpersonal Communication.

ROSENBUSCH, KATHERINE H. EdD. Address: 4400 University Drive, Fairfax, VA 22030. Email: [email protected]. Title: Assistant Professor of Management. Degrees: BS (Agricultural Development), Texas A&M University; MS (Agricultural Leadership Education), Texas A&M University; EdD (Human Resource Development), The George Washington University. Specializations: International Human Resource Development, Global Talent Development and Leadership, Expatriate Preparation and Adjustment.

CERNY II, LEONARD J. PhD. Address: 746 E. Chapman Avenue, Orange, CA 92866. Email: [email protected]. Title: Psychologist. Degrees: BS (Psychology), United States Air Force Academy; MDiv (Theology), Melodyland School of Theology; MA (Marriage, Family & Child Counseling), Chapman College; MA (Counseling Psychology), Rosemead Graduate School of Psychology; PhD (Clinical Psychology) Rosemead Graduate School of Psychology. Specialization: Stress and Resilience Management.

TABLE 1

The Number of Items and Cronbach's Alpha for each
of the 15 Scales in the 5 Domains

                               Number
                              of Items   Alpha

ORGANIZATIONAL DOMAIN            15       .85

Organizational Relationship       6       .83
Organizational Support            4       .76
Workload                          5       .73

CULTURAL DOMAIN                  13       .83

Adaptation                        6       .76
Transitions                       4       .67
Situational Crisis                3       .74

RELATIONAL DOMAIN                15       .80

Effective Relationships           8       .78
Family Adjustment                 3       .78
Extended Family/Friends           4       .65

RESILIENCE DOMAIN                12       .83

Well-being                        5       .75
Past Stresses                     3       .69
Focus                             4       .66

FOUNDATIONAL DOMAIN              13       .83

Spirituality                      5       .85
Health                            4       .61
Habits                            4       .74

For Total CSA Score computed from 15 scale averages, Cronbach's
alpha = .89

TABLE 2

PCA of Items on Scales in the Organizational Domain

                                                    Loadings

Scale                Items                        1     2     3

Organizational Relationship
  Supervisor micromanagement (a)                 .83
  Disagreeing with your leader's decisions       .80
  Insensitivity of leaders                       .73
  Conflicts on work team                         .67
  Leadership concerns                            .66
  Difficulty fitting the organizational style    .42
Organizational Support
  Lack of organizational support                       .87
  Too little organizational direction                  .86
  Organizational expectations                          .59
  Unhelpful organizational policies                    .52
Workload
  Email traffic load                                         .79
  Work interfering with your spiritual life                  .74
  Pressure to keep in touch with others                      .69
  Working too many hours                                     .65
  The need for a vacation                                    .52

Note: Blank entries < 0.30.

(a) Items are preceded by, "How much have you been stressed by ..."

TABLE 3

PCA of Items on Scales in the Cultural Domain

                                                Loadings

Scale               Items                     1     2     3

Adaptation
  Language learning struggles1               .86
  Not fitting in or feeling uncomfortable    .71
  Unfamiliar surroundings &                  .68
  customs
  Lack of familiar foods                     .63
  Hospitality expectations ....              .51
  The inability to fulfil your needs ....    .47
Transitions
  Severe disruption of living conditions           .78
  Inadequate housing                               .74
  Not enough privacy                               .69
  Stressful transitions                            .64
Situational Crisis
  Danger to yourself or family                           .83
  A violent experience                                   .81
  A potentially traumatic event                          .81

Note: Blank entries < 0.30.

(a) Items are preceded by, "How much have you been stressed by ..."

TABLE 4

PCA of Items on Scales in the Relational Domain

                                            Loadings

Scale             Items                   1     2     3

Effective Relationships
  Not feeling understood (a)             .75
  Relationship problems                  .72
  Loneliness                             .62
  Being suspicious of others             .60
  Being disappointed by                  .53
    friends or family
  Arguments with a
    significant other
  Difficulty controlling anger           .50
  Friends not honoring                   .49
    commitments
Family Adjustment
  Concerns for children                        .79
  Poor adjustment by a                         .74
    family member
  Disagreements about                          .69
    childrearing
Extended Family/Friends
  A friend or family member being ill                .76
  Concerns about parents                             .73
  A friend or family member dying                    .63
  Missing friends and family                         .50

Note: Blank entries < 0.30.

(a) Items are preceded by, "How much have you been stressed by ..."

TABLE 5

PCA of Items on Scales in the Resilience Domain

                                          Loadings

Scale             Items                1     2      3

Well-being
  Not enough emotional support (a)    .81
  Getting too emotional               .72
  Feeling depressed or sad            .69
  Finding it hard to relax            .65
  Troubling dreams                    .50
Past Stresses
  Early childhood trauma                    -.84
  Painful memories                          -.74
  Trouble concentrating                     -.69
Focus
  Difficulty focusing on tasks                     .80
  Difficulty remembering things                    .70
  Making decisions                                 .60
  Fear of failure                                  .50

Note: Blank entries < 0.30.

(a) Items are preceded by, "How much have you been stressed by ..."

TABLE 6

PCA of Items on Scales in the Foundational Domain

                                       Loadings

Scale           Items               1     2     3

Spirituality
  Lacking spiritual help and       .83
    direction (a)
  A lack of vital spiritual        .82
    relationships
  Feeling spiritually empty        .81
  Discouraged about your           .74
    spiritual life
  Being ineffective spiritually    .72
Health
  Health concerns about self or          .79
    family
  Lack of needed medical                 .70
    resources
  Sleep problems                         .57
  Low energy                             .53
Habits
  Feeling guilty about unwanted                -.91
    habits
  Unwanted desires or personal                 -.86
    habits
  Feeling guilty                               -.53
  Weight loss or gain                          -.43

Note: Blank entries < 0.30.

(a) Items are preceded by, "How much have you been stressed by ..."

TABLE 7

Components Analysis of 15 Scales Demonstrating 3
Broad Factors

                                     Pattern
                                  Coefficients

Factor         Scale              1     2     3

Work Management
  Organizational Relationship                .86
  Organizational Support                     .86
  Workload                                    *
Relationship Management
  Adaptation                           .39
  Transitions                          .59
  Situational Crisis                   .78
  Effective Relationships        .56
  Family Adjustment                    .62
  Extended Family/Friends              .68
Self-Management
  Well-being                     .70
  Past Stresses                  .72
  Focus                          .75
  Spirituality                   .69
  Habits                         .83
  Health                               .57

Note: * Blank entries were less than 0.30.

FIGURE 1

Stressor--Stress--Strain Model (Bhaskar-Shrinivas et al. 2005).

Environmental   Environmental Demand vs.    Expatriate Stress
demand          Expatriate Resource         Reaction
                Response

Stressor                 Stress                       Strain

Foreign         Stress emerges when the     Strains are reactions to
environmental   expatriate fails to cope    stress experiences
demands and     adaptively to these         including affective,
uncertainties   stressors. Stress           cognitive, and behavioral
impinge on      results from a response     outcomes such as job
expatriate's    pattern expressed in        dissatisfaction, early
resources.      maladjustment.              return, psychological
                                            withdrawal cognitions,
                                            and poor performance.