Towards an Integrated Analysis of Bias in Cross-Cultural Assessment

Abstract

A central methodological aspect of cross-cultural assessment is the interpretability of intergroup differences: Do scores obtained by subjects from different cultural groups have the same psychological meaning? Equivalence (or the absence of bias) is required in making valid cross-cultural comparisons. As cross-cultural comparisons are becoming increasingly popular and important, the problem of bias and its detection is receiving increased attention from researchers. Three kinds of bias are discussed and illustrated, namely construct bias, method bias, and item bias (or differential item functioning). Methods to identify bias are reviewed. An overview is given of common sources of each kind of bias. It is argued that an integrated treatment of all forms of bias is needed to enhance the validity of cross-cultural comparisons. The predominant focus on item bias techniques has the unfortunate implication that construct and method bias are examined insufficiently.

Validating a Measure of Productive Organisational Energy in the South African Context

“…This could indicate that there are differences between how a native English speaking individual understood and interpreted the questions as opposed to a native Afrikaans speaker. As mentioned by Van de Vijver and Poortinga (1997) the cultural backgrounds of individuals can have a significant impact on how tests are interpreted. It is therefore possible that the cultural background of the different Indigenous groups, the English-speaking White group, and the Afrikaans-speaking White group had an impact on their interpretation of productive energy in the workplace.…”

“…Construct indicates the extent to which the same construct is measured by a measurement instrument for differing cultural groups. According to Van de Vijver and Poortinga (1997), meaningful comparison of responses from multiple cultural or language backgrounds is dependent on the equivalence of the constructs being measured.…”

Suicidal Behavior and Psychological Distress in University Students: A 12-nation Study

“…Inorder to interpret the results from cross-national comparisons one should be able to make the assumption that mesaures across the groups are equal. For He, & van de Vijver, (2012) and Van de Vijver, & Poortinga, (1997) there are three kinds of bias that may endanger the validity and the generalizibility of results in cross-cultural comparisons. They are: 1) construct bias, 2) method bias and 3) item bias.…”

Are there gender differences in cognitive reflection? Invariance and differences related to mathematics

“…If the test does not measure the same construct across different groups, results are not comparable and inferences about group differences are misleading. The general term used to describe the lack of correspondence between measures applied to different groups is bias (Van de Vijver & Poortinga, 1997). Measurement equivalence might be threatened by different forms of bias.…”