Cross-cultural differences in response styles

Abstract

In the present study, I compared cross-cultural differences in response bias in college students in the United States, Mexico, Australia, the Philippines, Malaysia, and Japan. Specifically, I examined the convergence of indices of the same response bias for a given instrument; the stability of response bias indexes across instruments in the same data set; cross-cultural differences in acquiescence response style (ARS) in relation to individualism and collectivism; differences in extreme response style (ERS) based on the number of rating scale response points; cross-cultural differences in ERS; cross-cultural differences in midpoint responding (MRS); and the effects of controlling for response biases.

I examined several data sets that included a variety of instruments with a variety of constructs measured, response anchors, and numbers of rating points. I found at least moderate convergence of indices of the same response bias for a given instrument. Individual differences in acquiescence and ERS response biases were found to have some stability, but primarily across instruments with similar rating scales and constructs. I found mixed support for the hypothesis that ARS would be higher in collectivist cultures than in individualist cultures but ARS tended to be higher in Mexico than in the other cultures. ERS was found to be higher on scales with fewer response points compared to scales with higher numbers of response points. Hofstede's Power Distance dimension did not account for cross-cultural differences in ERS. There was some support for the hypothesis that MRS would be higher in the East Asian country of Japan but results were not statistically significant. There was not definitive support for the hypothesis that MRS would be higher in low power distance cultures than in high power distance cultures. Controlling for response bias significantly reduced cross-cultural mean differences in personality scores, although some response biases confound response bias and substantive variance. Theoretical and practical implications of this study are discussed.