Comparative Studies on Survey Sampling Bias in Cross-cultural Social Research

Social statistical studies have demonstrated that different sampling techniques can cause diverse biases toward the response results of cross-cultural survey. However, the unification to sampling methods in all countries is almost impossible because the sampling frame which includes all individuals of the target population does not necessarily exist in each country. This paper will discuss the intensity of impacts from the representative sampling approaches such as multi-stage sampling, stratified sampling, area sampling, quota sampling etc. which have been widely used in most cross-cultural surveys. Main statistical analyses have been carried out using the survey record data collected from two large-scale cross-cultural surveys. The important findings have shown that the non-probability sampling is a convenient approach in most situations, but it often causes a significant bias toward the sample unit drawing and an arbitrariness of individual selection for the face-to-face interview with the cross-cultural context. Sampling operations on decreasing the non-sampling error will also be demonstrated based on the real survey results