This symposium brings together modern developments in the area of cross-cultural assessments. The emphasis will go beyond traditional psychometric invariance testing. Papers will be presented on response styles, the use of ipsatization to address response styles, qualitative methods to assess bias, and the structure of emotions.

Within-subject standardization (ipsatization) has been advocated as a possible means to control for culture-specific responding (e.g., Fisher, 2004). However, the consequences of different kinds of ipsatization procedures for the interpretation of mean differences remain unclear. The current study compared several ipsatization procedures with ANCOVA-style procedures using response style indicators for the construct of family orientation with data from 14 cultures and two generations from the Value-of-Children-(VOC)-Study (4135 dyads). Results showed that within-subject centering/standardizing across all Likert-scale items of the comprehensive VOC-questionnaire removed most of the original cross-cultural variation in family orientation and lead to a non-interpretable pattern of means in both generations. Within-subject centering/standardizing using a subset of 19 unrelated items lead to a decrease to about half of the original effect size and produced a theoretically meaningful pattern of means. A similar effect size and similar mean differences were obtained when using a measure of acquiescent responding based on the same set of items in an ANCOVA-style analysis. Additional models controlling for extremity and modesty performed worse, and combinations did not differ from the acquiescence-only model. The usefulness of different approaches to control for uniform response styles (scalar equivalence not given) in cross-cultural comparisons is discussed.

Defining and assessing bias have been two of the main methodological topics in the cross-cultural field. Most of the attention has been paid to the development of statistical procedures to detect several kinds of biases and the interpretation of results in quantitative terms. However, qualitative procedures can be also useful for understanding the presence of bias when comparing different cultural or linguistic groups. The aim of this study is to illustrate potential contributions of Cognitive Interviews (CI) when investigating bias. On one hand, conclusions of integrating CI findings with quantitative data from analysing item bias will be presented by enhancing the advantages for understanding bias sources. On the other hand, utility of CI for extracting information of different levels of bias (item, method, and construct) will be described. The approach will be illustrated by studying responses and response processes of Dutch and Spanish participants to “Quality of Life” items from five international studies. The qualitative perspective of bias will be discussed as well as the potentiality of qualitative procedures for investigating bias, as single or as part of mixed methods studies.

Cross-cultural as well as in Western scientific literature is plagued with inconsistent theory development on the nature and the role of guilt and shame.  We present a large cross-cultural study that assesses these emotions on the basis of the componential emotion approach using two different methods, namely an episode and a frequency method. In total 3684 participants from 20 countries across the world rated appraisals, action tendencies, bodily reactions, expressions, and feelings in the three last episodes where they experienced a self-conscious emotion (episode method) and they also rated the frequencies of these emotional reactions in general (frequency method). Cultural stability of the internal structure was investigated by comparing classical principal component analysis with simultaneous principal component analysis. Both for the episodes and the frequencies a five-componential structure emerged stably across cultural groups. Four of the five components had the same meaning between the two methods, namely guilt, embarrassment, negative esteem of the self, and anger. The fifth component in the episode structure referred to the seriousness of the situation and the fifth factor in frequency structure could be interpreted as general distress. These cross-culturally stable internal structures allow for more consistent theorizing both in Western and in cross-cultural research.

This paper investigated the cross-cultural similarities and differences of extreme response style (ERS) extracted from self-report data of attitudinal and behavioral questions. Data of a subsample of 3,255 young adults with different immigration and racial backgrounds in the third wave of the UK household survey were analyzed. For each participant, responses to items concerning general mental health and identity were used to extract one ERS index for attitudinal questions and responses of items concerning family and school activities were used to extract the ERS index for behavioral questions. The two indexes were positively correlated, indicating similar response style preference in different types of questions. The pattern of cross-cultural mean differences were similar in both indexes, with minority groups (immigrants, Asian, African, Black, Caribbean, and mixed-race in UK) showing higher extreme response style compared with the majority group (nonimmigrants and whites in UK). However, there were more cross-cultural variations in attitudinal ERS than behavioral ERS. Implications are discussed.