Introduction

The symposium brings together a number of studies that deal with issues in cross-cultural assessment. Comparing cultures in terms of their psychological characteristics remains challenging despite the tremendous advancements of the last decades. Procedures that have been proposed in the past to adapt instruments for multilingual studies and to analyze data obtained in such studies seem to work well in small-scale studies. However, in more complex studies such procedures are not straightforward at all. Examples of complex studies are large-scale and longitudinal studies. The contributions of the symposium describe design and analysis of complex cross-cultural studies.

Contributions

Bond and Jing study moral tolerance for uncivil practices (e.g., cheating on taxes) and for non-traditional lifestyle practices (e.g., homosexuality) across 44 nations of the World Values Survey (2005-2007). Zeinoun, Daouk-Oyry, Choueiri, and Van De Vijver describe a personality study in the Arab-Levant, using qualitative and quantitative procedures. Nel, Meiring, Fetvadjiev, Hill and Van de Vijver describe a related study in South Africa. Finally, Benítez, Adams, and He address various types of bias a longitudinal survey in the United Kingdom, using the General Health Questionnaire (GHQ-12).

Conclusions

Assessment issues are pivotal in all the presentations. The presenters give examples of challenges in complex cross-cultural studies and ways in which these challenges can be resolved. The papers do not present a one-size-fits-all type of approach, but propose dedicated solutions for context-specific issues.

 

*****

Paper 1: Linking a Citizen’s Trust of Regulatory Institutions and of Out-Groups to Tolerance for Morally Questionable Practices in 44 Nations
Michael Harris Bond & Yiming Jing

Introduction: Ethical constructs differ from country to country as suggested by the voluminous cross-cultural research on values, but little comparative research has examined the justifiability of specific moral behaviors. The World Values Survey offers the opportunity to examine the predictors of the justifiability for morally questionable behaviors around the world. Further, the range of nations involved in the WVS enables an assessment of how national variation in culture may impact upon the relative power of these predictors from nation to nation.

Objectives: This study examines whether a citizen’s trust in regulatory institutions and trust of out-groups in the nation predict his or her tolerance for morally questionable behaviors. Further, it explores whether these individual-level linkages differ between countries with different priority goals for socializing children, viz., Self-directedness (e.g., independence) vs. Other-directedness, e.g., obedience, and Civility, e.g., tolerance vs. Practicality, e.g., instrumentality.

Design/Methodology: We tested this model across 44 nations using representative data from the World Values Survey (2005-2007). Age, gender, and education level were included as control variables in the HLM analysis.

Results: The justifiability of a variety of morally questionable behaviors was simplified into a 2-factor solution which tapped non-traditional and uncivil practices. We found that non-traditional practices, e.g., homosexuality, are more justifiable in both Self-Directed countries and in Civil countries. Additionally, socialization for Self-directedness promotes the contribution of a citizen’s institutional trust to his or her rated unjustifiability of uncivil practices, e.g., not paying taxes, whereas socialization for Civility promotes the contribution of a citizen’s out-group trust to his or her rated justifiability of non-traditional lifestyle practices.

Conclusions: The justifiability of two dimensions of morally questionable behaviors varies across nations as a function of how children in those nations are socialized. These different socialization practices influence the strength with which a person’s trust in society’s regulatory institutions and trust of out-groups links to the justifiability of non-traditional and uncivil practices, respectively.

 

Paper 2: Personality in the Arab-Levant: Quantitative and Qualitative Findings
Pia Zeinoun, Lina Daouk-Oyry, Lina Choueiri, Fons Van De Vijver

Introduction

Established personality tests are originally developed in English (e.g., Big Five Inventory), have been translated or adapted to other languages (e.g., Arabic) with varying psychometric properties. Moreover, the adoption/ adaptation of tests to another language/culture may preclude the identification of culture-specific factors.

Objectives

This project aimed at constructing an Arabic personality test using multiple methods.

Design/Methodology

First we culled personality descriptors of written Arabic, obtained ratings, and analyzed their factor structure (n = 803). Second, we gathered and qualitatively analyzed free descriptors of personality from Lebanon, Syria, Jordan, and the West Bank (n = 504). Finally, we analyzed the factor structure of an Arabic version of the IPIP.

Results

Results indicate that the Arab-Levant personality shares similarities with “Western†models of personality, but also has culture-specific dimensions that are not measured by popular personality tests. Conclusion: Implications for test and item construction in Arabic and for Arabic-speaking populations are discussed.

 

Paper 3: South African Personality Inventory: Where Are We Now?
Alewyn Nel, Deon Meiring, Velichko Fetvadjiev, Carin Hill, & Fons van de Vijver

Introduction

The South African Personality Inventory (SAPI) project was initiated in 2005 to overcome personality assessment issues in a multicultural and multilingual context. South Africa consists of 4 ethnocultural groups and 11 official languages which make it challenging for psychometric instrument constructors.

Objectives

This project aims to develop a fair, valid and reliable personality instrument for the diverse South African context which shows equivalence across all linguistic and cultural groups.

Design/Methodology

A modified lexical approach was utilized in the first phase of the project (N = 1216) where more than 55,000 personality descriptions were collected and content analyzed. Ensuing phases included various pilot studies with the developed items (total: 2573) generated from the responses of phase 1 (N = 6735), and workshops with linguistic and cultural experts to reduce total of items. Further samples were taken to evaluate the retained items in various validation studies.

Results

Phase 1 resulted in a proposed nine-cluster model which included Conscientiousness, Emotional Stability, Extraversion, Facilitating, Integrity, Intellect, Openness, Relationship Harmony, and Soft-Heartedness. Some aspects in these clusters resembled the universal Big Five model, while indigenous aspects were also identified. These clusters were further refined in various pilot and validation studies and yielded a six-cluster model which included Conscientiousness, Extraversion, Intellect/Openness, Negative Social-Relational, Neuroticism, and Positive Social-Relational.

Conclusions

Personality assessment is important for recruiting, selecting and developing employees. This study contributes to the processes researchers followed when they develop a personality instrument which aims to measure personality fairly across various linguistic and cultural groups.

 

Paper 4: An Integrated Approach to Bias in a Longitudinal Survey in United Kingdom: Assessing Construct, Method and Item Bias in the General Health Questionnaire (GHQ-12)
Isabel Benítez, Byron G. Adams, & Jia He

Introduction

Construct, method, and item bias are three levels of bias evaluated for valid group comparisons. Previous studies usually focus on one level of bias with integrated views still missing, especially in longitudinal designs.

Objectives

The aim of this study is to address bias in an integrated manner, using four waves of data in the UK Longitudinal Household Panel Survey.

Design/Methodology

Responses on the General Health Questionnaire (GHQ-12) from English natives and two generations of immigrants (first and second), and from different ethnocultural groups (White, Asian and Black) were used to analyze construct equivalence (construct bias), response styles (a core source of method bias), and Differential Item Functioning (DIF; item bias).

Results

While the basic structure of the GHQ-12 was stable across groups and time, levels of item bias and extreme response style decreased with repeated administrations. At the method level, the reduction of extreme response style was related to changes in the mode of administration, whereas at the item level, the systematic presence of DIF was inherent to the use of negations in the formulation of items.

Conclusions

The integrated results allowed the distinction between temporal sources of bias that became smaller over time and sources affecting comparisons consistently. We discuss the utility of the longitudinal and integrated approach for learning about bias.