DescriptionThis dissertation concerns the formation and the consequences of people’s beliefs and preferences. Empirically and experimentally, I focus on individuals’ beliefs and ideologies through mechanisms such as group dynamics, education, and educational content. In Chapter 2, I conduct a laboratory experiment in which subjects play a variant of a public goods game without free-riding incentives. With no private information, the prospect of investment in the experiment involves self-evaluation of uncertain personal and group types, which depend on performance on a pretest. I compare subjects’ individual investment decisions with their decisions when in groups. Timing and structure of communication are the two dimensions of controlled treatments. The results suggest that individuals tend to invest more often and that they increase their subjective beliefs of being a “high” type when in groups, especially when group decisions are made prior to individual decisions. Chapter 3 estimates the causal effect of education on religiosity in the United States using NLSY97. Fixed effects and instrumental variable method are used as identification strategies. Although cross-sectional ordinary least squares estimation shows a positive correlation between religious outcomes and educational attainment, both fixed effect models and IV estimation show statistically significant negative effects of education, even when cognitive test score is controlled. This suggests that conventional OLS omits factors that push both education and religiosity. In Chapter 4, I employ a regression discontinuity design and make use of a quasi-experimental approach by looking at the 1997 junior high school curriculum reform, “Knowing Taiwan,” to examine the impacts of the increased focus on Taiwan in the textbooks on a variety of civic engagement and national identity in favor of Taiwan. Data is from the Taiwan Social Change Survey. I find a significant positive effect of the new curriculum on national identity. However, most measures of civic participation of the treatment group do not exhibit jumps because of the textbook reform.