長谷部 拓也

In this article, I describe the commands that implement the estimation of three endogenous models of binary choice outcome. The command esbinary fits the endogenously switching model, where a potential outcome differs across two treatment states. The command edbinary fits the endogenous dummy model, which includes a dummy variable indicating the treatment state as one of the explanatory variables. After one estimates the parameters of these models, various treatment effects can be estimated as postestimation statistics. The command ssbinary fits the sample-selection model, where an outcome is observed in only one of the states. The commands fit these models using copula-based maximumlikelihood estimation.

In this article, I describe the escount command, which implements the estimation of an endogenous switching model with count-data outcomes, where a potential outcome differs across two alternate treatment statuses. escount allows for either a Poisson or a negative binomial regression model with lognormal latent heterogeneity. After estimating the parameters of the switching regression model, one can estimate various treatment effects with the command teescount. I also describe the command lncount, which fits the Poisson or negative binomial regression model with lognormal latent heterogeneity.

Reversing the obesity epidemic has been a persistent global public health challenge, particularly among low socioeconomic status populations and racial/ethnic minorities. We developed a novel concept of community-based incentives to approach this problem in such communities. Applying this concept, we proposed a school intervention to promote obesity prevention in the U.S. We conducted a pilot survey to explore attitudes towards this future intervention. The survey was collected as a nonprobability sample (N = 137 school-aged children (5–12 years)) in northern California in July 2013. We implemented multivariable logistic regression analyses where the dependent variable indicated the intention to participate in the future intervention. The covariates included the body mass index (BMI) based weight categories, demographics, and others. We found that the future intervention is expected to motivate generally-high-risk populations (such as children and parents who have never joined a past health-improvement program compared to those who have completed a past health-improvement program (the odds-ratio (OR) = 5.84, p &lt  0.05) and children with an obese/overweight parent (OR = 2.72, p &lt  0.05 compared to those without one)) to participate in future obesity-prevention activities. Our analyses also showed that some subgroups of high-risk populations, such as Hispanic children (OR = 0.27, p &lt  0.05) and children eligible for a free or reduced price meal program (OR = 0.37, p &lt  0.06), remain difficult to reach and need an intensive outreach activity for the future intervention. The survey indicated high interest in the future school intervention among high-risk parents who have never joined a past health-improvement program or are obese/overweight. These findings will help design and implement a future intervention.

We derive the asymptotic variance of the Blinder-Oaxaca decomposition effects. We show that the delta method approach that builds on the assumption of fixed regressors understates true variability of the decomposition effects when regressors are stochastic. Our proposed variance estimator takes randomness of regressors into consideration. Our approach is applicable to both the linear and nonlinear decompositions. Previously, only a bootstrap method has been a valid option for nonlinear decompositions. As our derivation follows the general framework of m-estimation, it is straightforward to extend our variance estimator to a cluster-robust variance estimator. We demonstrate the finite-sample performance of our variance estimator with a Monte Carlo study and present a real-data application.

While persistent racial/ethnic disparities in influenza vaccination have been reported among the elderly, characteristics contributing to disparities are poorly understood. This study aimed to assess characteristics associated with racial/ethnic disparities in influenza vaccination using a nonlinear Oaxaca-Blinder decomposition method. We performed cross-sectional multivariable logistic regression analyses for which the dependent variable was self-reported receipt of influenza vaccine during the 2010-2011 season among community dwelling non-Hispanic African American (AA), non-Hispanic White (W), English-speaking Hispanic (EH) and Spanish-speaking Hispanic (SH) elderly, enrolled in the 2011 Medicare Current Beneficiary Survey (MCBS) (un-weighted/weighted N=6,095/19.2million). Using the nonlinear Oaxaca-Blinder decomposition method, we assessed the relative contribution of seventeen covariates including socio-demographic characteristics, health status, insurance, access, preference regarding healthcare, and geographic regions - to disparities in influenza vaccination. Unadjusted racial/ethnic disparities in influenza vaccination were 14.1 percentage points (pp) (W AA disparity, p<0.001), 25.7 pp (W SH disparity, p<0.001) and 0.6 pp (W EH disparity, p>.8). The Oaxaca-Blinder decomposition method estimated that the unadjusted W AA and W SH disparities in vaccination could be reduced by only 45% even if AA and SH groups become equivalent to Whites in all covariates in multivariable regression models. The remaining 55% of disparities were attributed to (a) racial/ethnic differences in the estimated coefficients (e.g., odds ratios) in the regression models and (b) characteristics not included in the regression models. Our analysis found that only about 45% of racial/ethnic disparities in influenza vaccination among the elderly could be reduced by equalizing recognized characteristics among racial/ethnic groups. Future studies are needed to identify additional modifiable characteristics causing disparities in influenza vaccination. (C) 2015 Elsevier Ltd. All rights reserved.

This paper discusses the copula-based approach of a bivariate binary choice model. We derive the marginal effects of explanatory variables on an outcome of interest (both direct and indirect) in the model. We also show that the signs of the marginal effects are determined by the signs of the coefficient parameters. A real-data application is provided. (C) 2013 Elsevier B.V. All rights reserved.