We examine the effect of dental insurance coverage on the probability of having a dental care visit in light of selection bias.
We use data from the 2003 Medical Expenditure Panel Survey and use 3 different approaches to control for selection bias. First, we use a probit specification and include a rich set of independent variables that we posit control for unobserved attitudes toward risk and health care. Second, we use an instrumental variable model with family employment status as our instrument. Finally, we use a nonparametric approach to identify the upper and lower bounds of a dental insurance effect. We also ran a base probit model that did not include controls for attitudes toward risk and health care.
The base probit, the probit including measure of attitudes, and the instrumental variable models provided similar estimates of the effect of dental insurance on the probability to seek dental care. This may indicate that selection bias may not be a concern. All estimates were within the bounds obtained through the nonparametric approach.
Despite concerns of the potential endogeneity of dental insurance in models that estimate dental care use, we find evidence that these concerns may be unfounded.
*Center for Financing, Access and Cost Trends, Agency for Healthcare Research and Quality, Rockville, MD
†Dental Public Health, University of Maryland School of Dentistry, College Park, MD
‡Department of Economics, University of Virginia, Charlottesville, VI
Supported by the Agency for Healthcare Research and Quality, Rockville, MD.
The views expressed in this paper are those of the authors, and no official endorsement by AHRQ or the Department of Health and Human Services is intended or should be inferred.
The authors declare no conflict of interest.
Reprints: Philip F. Cooper, PhD, Center for Financing, Access and Cost Trends, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville, MD 20850. E-mail: firstname.lastname@example.org.