社会心理学データに対する分位点回帰分析の適用 : ネットワーク・サイズを例として [in Japanese] Extraversion changes the shape of the distribution of personal network size : An application of quantile regression to social psychological data [in Japanese]
Access this Article
Search this Article
Traditionally, research in social psychology has focused on the mean values of target variables. In most cases, this is simply because the mean value is the target of most statistical methods and often does not reflect the theoretical basis. Consequently, this has narrowed the perspective of researchers and possibly caused misunderstandings of social phenomena. In this study, we introduce quantile regression to solve this problem, which predicts the pth percentile of a target variable for any value of p. As an example, we theoretically predicted that the effect of extraversion on personal network size is different among the right (upper) and left (lower) parts of the distribution and tested this prediction using quantile regression. The result showed that extraversion positively correlates with the 70th-90th percentile of personal network size to a greater extent than that of the 30th-10th percentile. This result indicates that the distribution of personal network size not only moves toward the right but also becomes right-skewed as extraversion increases.
- Japanese Journal of Social Psychology
Japanese Journal of Social Psychology 29(1), 11-20, 2013
The Japanese Society of Social Psychology