Importance of Age-Period-Cohort Analysis: Time Series Change of Authoritarianism in Japan

Abstract

This paper discusses three effects of age, period, and cohort that are the most fundamental factors of time series change. We confirm the mathematical aspects of identification problem and its overcoming, and that it affects the sociological interpretation of data. Then, the effectiveness of APC analysis is demonstrated by comparing the results of this study with previous studies. In short, APC analysis is comprehensively shown to be important by bridging mathematical aspects and sociological interpretation through the phenomenon of the mixture of linear components, which is the source of the identification problem. As an example of the authoritarianism we verify the influence of each effect and the long-term trend by APC analysis assuming a random walk. The combined data from the SSM surveys from 1995 to 2015 show that there is a cohort effect rather than an age effect, and the change between cohorts is a U-shaped curve with the people born from 1945 to 1955 as the turning points. It has been suggested that the generations who led the student movement are the most anti-authoritarian, and that later generations are gradually becoming authoritarian.