Bayesian Hierarchical Models for Response Rates in Hunter Attitude Surveys

• Activity Number: 17
• Type: Topic Contributed
• Date/Time: Sunday, August 7, 2005 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Xiaoming Gao*+ and Zhuoqiong He and Dongchu Sun
• Companies: Missouri Department of Conservation and University of Missouri, Columbia and University of Missouri, Columbia
• Keywords: attitude survey ; nonresponse ; response rate ; Bayesian hierarchical model ; spatial correlation ; logistic linear mixed model

Abstract

Hunters and farm operators are two key elements in deer management in Missouri. Deer hunter attitude surveys have been developed to assess the opinions of firearms deer hunters about deer populations and hunting opportunities. Survey results are used to set deer population goals and hunting regulations. Despite the best efforts by survey practitioners to reduce nonresponse through careful design of questionnaires and implementation---including multiple waves---nonresponse is a persistent problem for survey researchers. The response rates for deer hunter attitude surveys went from 72% in 1979, to 73% in 1991, to 55% in 2001. The impact of nonresponse bias upon management decisions can produce unexpected outcomes. This is especially critical when small adjustments in management regulation stimulate major changes in the outcome. In this paper, we propose Bayesian hierarchical and spatial models for characterizing respondents and non-espondents, including the information of various explanatory variables such as hunter's age, date of permit purchase, and residence. We explore the effects of such covariates on the response rates.