ANOTHER LOOK AT THE STRATIFICATION OF EDUCATIONAL TRANSITIONS: THE LOGISTIC RESPONSE MODEL WITH PARTIAL PROPORTIONALITY CONSTRAINTS

In this paper we reanalyze Robert D. Mare's highly influential work on educational transitions among American men born in the first half of the 20th century. Contrary to previous belief, Mare found that the effects of socioeconomic background variables decline regularly across educational transitions in conditional logistic regression analyses. We have reconfirmed Mare's findings and tested them by introducing a modified logistic response model that constrains selected social background effects to vary proportionally across educational transitions. We refer to our preferred model as the logistic response model with partial proportionality constraints (LRPPC). The model can easily be estimated in Stata or using other standard statistical software. Partial proportionality constraints may also prove useful in interpopulation comparisons based on other linear models.     

Assumptions of social mobility research in the U.S.: The case of occupational status

This paper examines the theoretical and empirical basis for common rank-orderings of occupational roles by raters throughout the world. From an illustrative comparison of occupational stratification in Australia and the United States we conclude tentatively that (1) commonalities in the socioeconomic characteristics of occupational roles provide the basis for interplace consistencies in “prestige” scores, (2) the structure of occupational mobility in the U.S. and Australia is largely similar, (3) this similar structure of occupational stratification manifests a common socioeconomic process which defines a (the?) major component of occupational mobility in capitalist, industrial (only?) societies. We infer that “prestige” scores for occupations are fallible estimates of the socioeconomic statuses of occupation, within the context of mobility processes in (at least) the U.S. and Australia.     

Bad news indeed for Ryff’s six-factor model of well-being

Springer and Hauser (An Assessment of the Construct Validity of Ryff’s Scales of Psychological Well-Being: Method, Mode, and Measurement Effects. 2006. Social Science Research 35) tested one key aspect of the validity of Ryff’s six-factor model of psychological well-being (RPWB), namely, whether there is substantial independent variation among the six factors. In several large and heterogeneous samples, under a variety of model specifications, and using various sets of RPWB items, we found very high factor correlations among the dimensions of well-being, especially personal growth, purpose in life, self-acceptance, and environmental mastery. That is, the six-factor model makes theoretical claims that do not yield large or consistent empirical distinctions when standard measures and instrumentation are used. Where Ryff and Singer’s comment (Best News Yet on the Six-Factor Model of Well-Being. 2006. Social Science Research 35) refers directly to that analysis, their methodological discussion is most often irrelevant or incorrect. Their text largely ignores and fails to challenge our strong empirical findings about the factorial structure of well-being. In this response, we reinforce these findings and their implications for the (in)validity of the six-factor well-being model as implemented by Ryff. We also explain why Ryff and Singer’s lengthy review of studies that show differential relationships of RPWB factors with other variables should be interpreted with far greater caution than Ryff and Singer recognize. We offer recommendations for analyzing RPWB items in surveys that have already been conducted, but we also emphasize the need for a thorough rethinking of the measurement and dimensionality of psychological well-being.     

Religion,socialization, and fertility

Longitudinal data from a large .sample of Wisconsin men and women are used to examine the effects on fertility of religious and secular socialization, including farm upbringing. Analyses of children ever born (CEB) and of parity progression show that current religious choice is more important in explaining fertility than is religion of orientation or denomination of secondary school. The effects of current and background religion are additive, and the effect of current religion is the same for men as for women at each parity progression. Catholic religious background affects fertility primarily by increasing the likelihood of having a third or fourth child; its indirect effects on fertility operate through religious schooling and current religious affiliation. Unlike religious background, the positive influence of farm background on fertility persists among men and women, even when current farm employment is controlled.     

Explaining the Gender Wealth Gap

To assess and explain the United States’ gender wealth gap, we use the Wisconsin Longitudinal Study to examine wealth accumulated by a single cohort over 50 years by gender, by marital status, and limited to the respondents who are their family’s best financial reporters. We find large gender wealth gaps between currently married men and women, and between never-married men and women. The never-married accumulate less wealth than the currently married, and there is a marital disruption cost to wealth accumulation. The status-attainment model shows the most power in explaining gender wealth gaps between these groups explaining about one-third to one-half of the gap, followed by the human-capital explanation. In other words, a lifetime of lower earnings for women translates into greatly reduced wealth accumulation. After controlling for the full model, we find that a gender wealth gap remains between married men and women that we speculate may be related to gender differences in investment strategies and selection effects.     

Exchange rate regimes and the cross-country distribution of the 1997 financial crisis

We study variations in the severity of the 1997 financial crisis in a sample of 25 developing countries. We use both currency depreciation and stock market returns as crisis measures. Our key findings are that countries that started 1997 with an exchange rate peg experienced significantly greater currency depreciation and significantly lower stock returns than would be predicted from the levels of various macroeconomic indicators.     

Model averaging procedure for partially linear single-index models

This paper is concerned with model selection and model averaging procedures for partially linear single-index models. The profile least squares procedure is employed to estimate regression coefficients for the full model and submodels. We show that the estimators for submodels are asymptotically normal. Based on the asymptotic distribution of the estimators, we derive the focused information criterion (FIC), formulate the frequentist model average (FMA) estimators and construct proper confidence intervals for FMA estimators and FIC estimator, a special case of FMA estimators. Monte Carlo studies are performed to demonstrate the superiority of the proposed method over the full model, and over models chosen by AIC or BIC in terms of coverage probability and mean squared error. Our approach is further applied to real data from a male fertility study to explore potential factors related to sperm concentration and estimate the relationship between sperm concentration and monobutyl phthalate.