Extent and impact of response biases in cross-national survey research

摘要

Survey research is fraught with serious response biases. This study examines the extent and impact of three important response biases in cross-national research: socially desirable responding, yea-saying, and nay-saying. From a survey of 5569 respondents across 15 countries, the study finds evidence of substantial differences in response biases across countries. Socially desirable responding is highest in Singapore and Italy, yea-saying is highest in Brazil and India, and nay-saying is highest in the Netherlands and Japan. These response biases lead to erroneous conclusions about innovativeness based on surveys as compared to that based on the market penetration of new products. In particular, the biases lead to the over-reporting or underreporting of innovative traits and the over-reporting of the adoption of new products. Overall, negatively valenced items that show the least susceptibility to these response biases can help predict both actual penetration at the aggregate level as well as individual probabilities of adoption and should be included in cross-national surveys.