<論文>メタ分析によるセクハラ大学教授の平均年齢と身分の推定

Abstract

Meta-analysis is a method used to synthesize several different studies measuring an effect size, e.g., an odds ratio, difference of means, or the slope of a regression line. Although meta-analysis is often used to unify the results of different professional academic papers estimating the same effect size, it can be applied to unify essays written by undergraduate students as well. In this paper, I meta-analyzed essays written by students taking my class on survey methods. I assigned the students the task of collecting a probabilistic sample of articles on sexual harassment cases caused by full professors or associate professors, based on reports published in Asahi Newspapers from 2000 to 2020. The students were then tasked with estimating the 95% confidence intervals of the mean age and rate of harassment by full professors reported in their samples. I then meta-analyzed those results in order to obtain better estimates. The results show no significant difference between the rates of harassment by professors and those by associate professors. However, the mean age of the harassers was lower than the other professors by 1.3 years, which contradicts the hypothesis that older and more powerful people would be more likely to commit sexual harassment.