Abstract

Alcohol consumption is likely underreported in the National Health and Nutrition Examination Survey. The problem, common among most self-reported data, stems from social desirability bias. As a result of this bias, the true level of alcohol consumption, specifically heavy episodic drinking, or binge drinking, is undercounted. By applying zero inefficiency stochastic frontier analysis, we adjust for the inefficiency caused by survey participants underreporting their alcohol consumption. This paper serves to partially correct the estimates of heavy episodic drinking and to serve as an example that stochastic frontier analysis can be used outside of its standard application to correct underreported and self-reported data. The study concludes that among the sample of 2,901 NHANES participants, 27.51% underreported their alcohol consumption. It further concludes that among those who did originally admit to binge drinking, 69.51% underreported the true extent of their binge drinking episodes. However, of the people who did not report binge drinking, none of them underreported. While more research should be done to determine why the model stated that none of the nondrinkers underreported, this paper demonstrates a possible use of the zero inefficiency stochastic frontier model in regards to adjusting self-reported data.

Thesis Completion

2020

Semester

Spring

Thesis Chair/Advisor

Hofler, Richard

Degree

Bachelor of Science (B.S.)

College

College of Business Administration

Department

Economics

Degree Program

Economics

Language

English

Access Status

Open Access

Release Date

5-1-2020

Included in

Economics Commons

Share

COinS