DISSERTATION RESEARCH: Complex Family Backgrounds: A Latent Class Approach

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

This dissertation reanimates an unresolved debate among intergenerational social mobility researchers about how to measure social class background at the family level. The debate concerns whether or not the social status of both parents, as opposed to that of the father or the head of household only, define the status of the family unit as a whole. The dissertation takes a new approach to resolving this debate and in doing so makes a methodological contribution to social mobility research. At the same time, the dissertation extends beyond concerns with this debate per se to focus on ways in which superior measures of overall family class can advance knowledge about how other aspects of families shape children's outcomes. Specifically, the dissertation illustrates that prior measures of overall family social class include significant measurement error; it also shows that findings on issues ranging from divorce to changes in mobility patterns over time have suffered from omitted variable bias as a result. The dissertation utilizes publicly available national survey data including the General Social Survey and the National Longitudinal Survey of Youth. Strategies of analysis primarily include log-linear and log-multiplicative models and latent class models of social mobility and attainment. The central research hypotheses of the dissertation are the following: First, the joint influence of spouses on overall family class background is of a cumulative nature, meaning that the status of both spouses must be included for accuracy. Second, information about the exact nature of the joint parent influence can be used to create a parsimonious measure of family background that includes the effects of both spouses. Third, use of the comprehensive measures(s) in new investigations of change over time in social mobility patterns and effects of non-intact family types and of race on children's outcomes will improve the accuracy of knowledge about these important topics. The broader impacts of this research include the following: This research will be beneficial to both the scientific and policy-making communities. Other social scientists will benefit from the ability to use the joint measure of family status designed in this project rather than one defined by the fathers' status only. The research will inform theoretical debates in sociology around the mechanisms behind negative effects of family disruption as well as behind social mobility processes more generally. Finally, I will work to disseminate the findings to a policy audience as they will potentially inform discussion on pro-marriage anti-poverty policies, or policies to promote non-residential parent involvement in non-intact families.