Jonathan Marc Bearak
Summary of Dissertation

My dissertation analyzes marriage, sexual risk behavior, and earnings inequality. These reflect separate strands of my research, yet each also reflects the application of common methodological principles to move forward the areas they address. Social science tends to study the average outcome, or the mythical average subject, rather than how treatments affect the whole distribution, and often oversimplifies the temporal structure of processes, in ways which can obfuscate the true nature of phenomena.

More than thirty years of research finds that married men earn more. Scholars have long suggested that gender-normative specialization increases male earnings, but recent research shows that the marriage differential is inconsistent with this, as both sexes earn more when they marry and some of this differential arises prior to the marriage. This leaves the literature with a sizable yet unexplained phenomenon.

The marriage differential is often discussed in ethnography, economics and demography. Consider that the relationship between marriage and male earnings may systematically vary across the distribution. Ethnography focuses on the low end of the distribution. The economic and demographic arguments focus on the middle. A popular narrative also concerns itself with those at the top. The ethnographic literature invites inferences which contrast lower and upper strata, but it is inappropriate to make inferences beyond the population studied. If different processes occur at the bottom and at the top of the distribution, then looking only at the middle obfuscates the true relationship.

Marriage is an event, but it is also a process. The marriage differential reflects not simply the state of marriage, but competition within marriage markets and normative expectations of the economic requisites for marriage. My research applies quantile regression methods to analyze the gradient in the marriage differential. It is much larger among relatively low earners. This gradient emerges just prior to marrying, and peaks around the time of marriage. The results are consistent with, and highly suggestive of, economic thresholds for socially acceptable marriage.

Studies of sexual risk behavior often ignore how human beings make decisions, implicitly assuming condom use equally risk-reducing regardless of relationship context. When studying populations like American four-year college students, it is illogical to treat all kinds of sexual behavior as equally risky. My research analyzes changes in the odds of coitus, and condom use when coitus occurs, in the casual sexual events students typically call “hookups” between years 1-4 of college. The odds of coitus increase gradually through senior year, whereas the odds of condom use decrease abruptly between freshman and sophomore year. This decline occurs as students from lower-class backgrounds adapt to the behavior of upper-class peers. In contrast to typical health outcomes, such as smoking or diet, the lower-class students, initially, exhibit the safer behavior.

Underlying each of these three outcomes are several decisions which may be sequenced in different ways: whether to buy a condom in advance of a potential liaison, whether to have sex in the liaison, whether to use a condom conditional on having one, and whether to interrupt a liaison to acquire a condom. Each decision must be modeled in sequence. Simulations reveal the distribution of the three outcomes in different scenarios, whereas the models by themselves do not.

The third part of my dissertation addresses changes in family earnings inequality since the mid-twentieth century. Family earnings reflects those of one or both partners, and a number of determinants, including marriage, fertility, and education. Their relationship with earnings inequality, and each other, has changed over time. Marriage and education at one point in the life course are related to fertility in the future, and fertility at one point in the life course is related to marriage and education in the future. Prior analyses of changes in inequality over time have not accounted for the temporal structure of these processes, and how these differentially affect inequality between the poor and the middle class, and between the middle and the most well-to-do.

My research addresses inequality with respect to gender, relationship and family formation, fertility and education. These are essential factors to the human condition in modern society. They affect inequality in earnings, health, and well-being broadly. Their relative importance changes over time, between cohorts, and over the life course, within cohorts. Social science often looks at average differences between groups, and compares outcomes before and after an event. I focus on how paying closer attention to the distributional and temporal components of social processes, using new methods and in new ways, improves our understanding of the mechanics of human society.