Three essays in demography: method, gendered generational change, assortative mating, and fertility.

This thesis explores the differential effect of socioeconomic status (SES) on a series of family and fertility attitudes and behaviors from different methodological lights. The first essay studies response pattern from factorial survey experiment and its relationship with traditional survey question and respondent’s perceived SES. Across 8 survey countries, I find that participants who tend to see vignettes as successful (i.e., whose mean vignette ratings in the 75% quantile with mean rating = 7.20) also tend to see their own lives more satisfied, less impacted by COVID-19, and report higher subjective wealth. In other words, optimistic participants tend to give more positive ratings overall. Meanwhile, they also rate different vignettes with lower variance, and show milder preference in all factors. For relational factors (i.e., communication, contact with grandparents, gendered work-life balance), optimistic respondents see relational factors at twice less impactful than others. While objective SES (e.g., educational level, household income) is positively associated with perceived SES and well-being, I find few differential effects of objective SES on family ideals, and its relationship with response style differ across countries. In China, Japan, Korea and Spain, SES is negatively associated with the optimistic response style, while in the US, Singapore and Italy, SES is positively associated with the associated with the optimistic response style. In Norway, SES is unrelated. The complicated and endogenous relationship between survey response style, perceived social status, and family related attitudes contributes to the null finding of SES on family ideal resonating with the Hainmueller and his colleagues’ finding that binary choice (0 vs. 1) experiment could better validate real-world behavior. The second essay explores the strategic choices of multi-generational caregiving. Grandparents pass down resources not only through the middle generation but also through direct channels such as caring for the grandchildren, and in some countries, which grandparent to provide grandchildren care seem to be strategic. In China, Spain and Italy, the more highly educated grandparents (out of mother’s and father’s parents) are more likely to regularly take care the grandchildren. However, this pattern is absent in Korea and Japan where the family institution has long been categorized to the East Asian family with China. The third paper studies the relationship between assortative mating and differential fertility with large-scale representative sample. I find that educational hypergamy is negatively related to women’s decision making in childrearing decisions, but not related to in decisions related to leisure activities. In addition, Hypergamous parents are less likely to have only one daughter, and they are more likely to have a boy as the last born child, especially when the first born is a girl.