Scott M. Lynch

Cohort and life-course patterns in the relationship between education and health: A hierarchical approach

Recent medical sociological research has examined whether the relationship between education and health is dynamic across age, whereas recent demographic research has examined whether the relationship varies across cohorts. In this study, I examine how cohort structures the influence of education on life-course health trajectories. At the cohort level, changes in education and in the distribution of health and mortality make cohort differences in education’s effect probable. At the life-course level, the effect of education may vary across age because the mediators of the education-health relationship may vary in their relevance to health across the life course. Using basic regression analyses and random-effects models of two national data sets, I find that the effect of education strengthens across age, that this pattern is becoming stronger across cohorts, and that these patterns are suppressed when either effect is ignored.

Dreams Fulfilled, Dreams Shattered: Determinants of Segmented Assimilation in the Second Generation

We summarize prior theories on the adaptation process of the contemporary immigrant second generation as a prelude to presenting additive and interactive models showing the impact of family variables, school contexts and academic outcomes on the process. For this purpose, we regress indicators of educational and occupational achievement in early adulthood on predictors measured three and six years earlier. The Children of Immigrants Longitudinal Study, used for the analysis, allows us to establish a clear temporal order among exogenous predictors and the two dependent variables. We also construct a Downward Assimilation Index, based on six indicators and regress it on the same set of predictors. Results confirm a pattern of segmented assimilation in the second generation, with a significant proportion of the sample experiencing downward assimilation. Predictors of the latter are the obverse of those of educational and occupational achievement. Significant interaction effects emerge between these predictors and early school contexts, defined by different class and racial compositions. Implications of these results for theory and policy are examined.

Race differences in depressive symptoms: a dynamic perspective on stress exposure and vulnerability.

The existence, nature, and strength of race differences in mental health remain unclear after several decades of research. In this research, we examine black-white differences in the relationship between acute stressors and depressive symptoms. We reframe the stress exposure and differential vulnerability hypotheses in the context of long-term trajectories of stress and depression, and we hypothesize that trajectories of stress growth will be associated with trajectories of depressive symptom growth. Using latent growth curve analysis of a sample of 1,972 older persons interviewed three times at three-year intervals, we test the hypotheses that (1) growth in exposure to loss-related events will predict growth in depressive symptoms, and (2) African Americans will experience greater stress growth than whites. Results support the hypotheses. Stress growth exhibited a linear increase for blacks but not for whites, and predicted depression growth for both races, but explained more variance for blacks than for whites.

Explaining life course and cohort variation in the relationship between education and health: the role of income.

Researchers have attempted to explain the relationship between education and health by incorporating mediators—such as income—through which schooling affects health. Research has also shown that the education-health relationship varies across age and cohort. This article integrates these themes in an investigation of the role income plays in the changing relationship between education and health. Using data from the 1972–2001 National Health Interview Surveys, I find that the direct relationship between education and health is weakening across cohorts, while the indirect effect of education through income is strengthening across cohorts. Cohort change in this indirect effect is equally attributable to change in the relationships between education and income and between income and health. Life course differences in direct and indirect relationships vary in shape, suggesting that a substantive explanation—one that does not rely solely on selective mortality—is warranted. The results suggest that future research should continue to employ a life course framework.

Who Goes Who Stays: An Assessment of the Effect of a Freshman Mentoring and Unit Registration Program on College Persistence.

We present an evaluation of a freshman block registration and mentoring program at a major state university. In an effort to improve retention, the University initiated a freshman block registration and mentoring program in the fall of 1994. The program was developed to strengthen social support, which we hypothesize leads to higher rates of persistence. We present the effects of the program on retention and academic performance for successive cohorts from 1994 through 1998. Academic performance and retention rates are used as outcome measures and are related to student and university characteristics. We describe the nature of the program and the impact that it has on persistence (survival times), grades, and graduation. Our analysis uses event history models (follow back life tables and Discrete-Time Logit Models).

Measurement and Prediction of Aging Anxiety

Anxiety about aging has been virtually neglected in extant research, but such anxiety should become increasingly important to understand as the population ages. This research addresses three questions pertaining to aging anxiety. First, is aging anxiety a meaningful unidimensional construct? Second, what is the age pattern of aging anxiety? Third, what factors influence aging anxiety? Confirmatory factor analyses are conducted to examine the factorial validity of an aging anxiety scale across age, sex, and race aggregates. Structural equation models are used to determine the influence of several covariates. There were several important findings, including that (1) aging anxiety can be measured as a single construct, although the scale taps a somewhat different dimension for younger persons; (2) fear of social losses does not appear to be a concern about aging; (3) the age pattern of anxiety is declining, except across midlife; and (4) race, gender, socioeconomic status, current health, income, education, and knowledge about aging each affect aging anxiety.

Reconsidering Mortality Compression and Deceleration: An Alternative Model of Mortality Rates

In this research we develop a model of mortality rates that parameterizes mortality deceleration and compression, permits hypothesis tests for change in these parameters over time, and allows for formal gender comparisons. Our model fits mortality data well across all adult ages 20–105 for 1968–1992 U.S. white data, and the results offer some confirmation of findings of mortality research using conventional methods. We find that the age at which mortality deceleration begins is increasing over time, that decompression of mortality is occurring, and that these trends vary substantially across genders, although male and female mortality patterns appear to be converging to some extent.

Refining the Measurement of Women's Autonomy: An International Application of a Multi-dimensional Construct

Womens autonomy has long been a central concern for researchers examining the social position of women in developing countries. However, little emphasis has been placed on the measurement of autonomy, despite its importance for assessing the validity of comparative research. In this research, we use confirmatory factor analyses to determine (1) whether items thought to measure autonomy in fact form a reliable measure of autonomy, (2) whether the relationship between multiple dimensions of autonomy are strong enough to justify a discussion of autonomy as a single underlying construct, and (3) whether comparative research on autonomy is possible between two countries (India and Pakistan). We find that our indicators capture four distinct dimensions of autonomy that are moderately related, and that, while the model structures replicate fairly well across the two countries we study, there are measurement differences that make comparative research challenging.

The Pursuit of Happiness in China: Individualism, Collectivism, and Subjective Well-Being During China’s Economic and Social Transformation

This paper examines the consequences of China’s dramatic socioeconomic and political transformations for individual subjective well-being (SWB) from 1990 to 2007. Although many still consider China to be a collectivist country, and some scholars have argued that collectivist factors would be important predictors of individual well-being in such a context, our analysis demonstrates that the Chinese are increasingly prioritizing individualist factors in assessments of their own happiness and life satisfaction thus substantiating descriptions of their society as increasingly individualistic. While the vast majority of quality of life studies have focused on Westerners, this study contributes findings from the unique cultural context of China. Moreover, concentration on this particular period in Chinese history offers insight into the relationship between SWB and rapid socioeconomic and political change.

Bayesian Posterior Predictive Checks for Complex Models

In sociological research, it is often difficult to compare nonnested models and to evaluate the fit of models in which outcome variables are not normally distributed. In this article, the authors demonstrate the utility of Bayesian posterior predictive distributions specifically, as well as a Bayesian approach to modeling more generally, in tackling these issues. First, they review the Bayesian approach to statistics and computation. Second, they discuss the evaluation of model fit in a bivariate probit model. Third, they discuss comparing fixed- and random-effects hierarchical linear models. Both examples highlight the use of Bayesian posterior predictive distributions beyond these particular cases.