Testing for Causal Effects in a Generalized Regression Model With Endogenous Regressors

A unifying framework to test for causal effects in nonlinear models is proposed. We consider a generalized linear‐index regression model with endogenous regressors and no parametric assumptions on the error disturbances. To test the significance of the effect of an endogenous regressor, we propose a statistic that is a kernel‐weighted version of the rank correlation statistic (tau) of Kendall (1938). The semiparametric model encompasses previous cases considered in the literature (continuous endogenous regressors (Blundell and Powell (2003)) and a single binary endogenous regressor (Vytlacil and Yildiz (2007))), but the testing approach is the first to allow for (i) multiple discrete endogenous regressors, (ii) endogenous regressors that are neither discrete nor continuous (e.g., a censored variable), and (iii) an arbitrary “mix” of endogenous regressors (e.g., one binary regressor and one continuous regressor).     

Block clustering with collapsed latent block models

We introduce a Bayesian extension of the latent block model for model-based block clustering of data matrices. Our approach considers a block model where block parameters may be integrated out. The result is a posterior defined over the number of clusters in rows and columns and cluster memberships. The number of row and column clusters need not be known in advance as these are sampled along with cluster memberhips using Markov chain Monte Carlo. This differs from existing work on latent block models, where the number of clusters is assumed known or is chosen using some information criteria. We analyze both simulated and real data to validate the technique.     

Education and the changing shape of the income gradient in health

Research on the social determinants of health has increasingly sought to understand the relative importance of different features of socioeconomic status. Much of the ensuing debate has wavered between education and income, with recent research leaning increasingly toward income. This research has not, however, consistently explored interactions between different features of socioeconomic status and, in trying to understand the independent effects of different components of socioeconomic status, may have missed important features of socioeconomic position. With an eye toward examining how features of socioeconomic status combine and coalesce, this paper examines variation in the income-health association by level of education. Theories derived both from medical sociology and health economics suggest synergistic interactions between income and education, but they are unclear as to the direction and magnitude of these interactions. Results from two large and nationally representative data sets (the 1996-1997 Community Tracking Study and the 1972-2000 General Social Survey) indicate that the positive relationship between income and health varies substantially in both its strength and shape by level of education. Education improves health, and its effects are larger at lower levels of income. Moreover, education reduces the strength and curvature of the income-health relationship. Consequently, those with more education have better health for all levels of income, and fewer income-based disparities exist among the well educated than among the less well educated. The linear "gradient" relationship between income and health is, thus, more characteristic of groups with higher levels of education. Additional analyses indicate that these interactions existed in the United States in each of the last three decades. The results are discussed in light of theory regarding the perpetuation of health disparities, as well as current debates regarding the apparent incompatibility of distributive versus aggregative goals in health policy.     

Factors influencing middle and high schools' active parental consent return rates

The authors examined factors influencing the return rates for attempting to collect active parental consent forms from 21,123 students in the 7th through 10th grades in 41 middle and high schools. Overall return rates from middle schools were higher than from high schools. Schools that offered high levels of staff support for collecting consent forms had higher return rates. Procedures where the consent form was attached to a school form that parents had to complete and return to the school yielded the highest return rate. Implications for how researchers can obtain a high parent consent form return rate are discussed.     

Misgivings of medicine?: African Americans' skepticism of psychiatric medication

Research has consistently documented black-white differences in rates of voluntary psychiatric treatment that cannot be reduced entirely to differences in either need or access. A variety of explanations have been offered for the gap that remains, but the empirical testing of alternative propositions has thus far been minimal. Using the 1998 General Social Survey's Pressing Issues in Health and Medical Care module (n = 1,387), I find consistent and substantial black-white differences in a variety of beliefs about psychiatric medications, one of the most common treatments for mental illness, and the predisposition to use them. Specifically, blacks express less willingness to use psychiatric medications themselves or to administer them to a child for whom they are responsible. Neither socioeconomic status, knowledge, religious involvement, nor trust in medicine appears to explain this reluctance. Rather, it stems almost entirely from blacks' beliefs about psychiatric medications' efficacy and side-effects. The results indicate, first, that researchers should not assume that African Americans will use psychiatric medications at rates similar to whites if offered equal access. Second, the results indicate that blacks' skepticism of psychiatric medications may be rooted in specific beliefs about psychiatric medications, rather than general ideologies about medical practice. Health beliefs about psychiatric treatment, therefore, will continue to play an important role in understanding race differences in the use of psychiatric medications.     

Preliminary evaluation of contingent meals and telephone use to improve evidence-supported family therapy session attendance in mothers referred by Child Protective Services for substance abuse

Persons referred by Child Protective Services (CPS) for substance abuse evidence high rates of nonattendance to therapy sessions, taxing systems of care and exacerbating outcomes. This study examined the influence of two telephone-based incentive programs on therapy session attendance in mothers who were referred by CPS for substance abuse. After baseline therapy session attendance was established in an evidence-based clinic that incorporated a telephone engagement intervention (Phase I), participants were provided free cellular telephones with limited minutes and permitted to order free meals delivered by the therapist during upcoming sessions (Phase II). The third phase was similar to Phase II, but participants were provided unlimited minutes. Results indicated that participants’ attendance was significantly improved when meals and cellular telephone minutes were contingent on attendance. Although the percentage of sessions attended by participants during Phase III was higher than Phase II, unlimited minutes and meals did not significantly enhance attendance relative to limited minutes and meals. Session attendance for significant others of these participants was significantly higher during Phase III as compared with Phase I. Session attendance of significant others was statistically similar between Phase I and II and between Phase II and Phase III. Study implications and recommendations for future research and practice are discussed in light of the findings.     

FLIP‐FLOPPING: IDEOLOGICAL ADJUSTMENT COSTS IN THE UNITED STATES SENATE

Using a long panel of roll call voting data, I find that “flip‐flopping” senators face significant electoral costs when changing positions. In models of electoral competition, as the costs to candidates changing position approach zero, the equilibrium prediction is the convergence of platforms. Such convergence is at odds with empirical observation. Using a dynamic, structural model of candidate positioning, I identify the nature of the costs associated with changing position that may result in such non‐convergence. (JEL D72, C61, H11)