Date of Award

8-2007

Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)

Department

Resource Economics and Policy

Advisor

Mario F. Teisl

Second Committee Member

Kathleen P. Bell

Third Committee Member

Mary E. Davis

Abstract

Stated preference surveys have become an increasingly important tool for estimation of nonmarket values. In 1993, The National Oceanic and Atmospheric Administration (NOAA) convened a panel that set forth methodological guidelines for the most prevalent stated preference survey method, Contingent Valuation (CV). More recently, Conjoint (CJ) Analysis, an attribute-based stated preference method, has been applied to economic research. Although most stated preference surveys include multiple valuation questions, the assumption of stable preferences in economic theory implies that the order of presentation of valuation questions should not affect responses and corresponding willingness-to-pay estimates. A fair amount of methodological studies have found question order effects in willingness-to-pay (WTP) estimates produced by CV, but very few studies have tested for question order effects in WTP estimates produced by CJ. The presence of question order effects in stated preference surveys suggests either poor construct validity of the survey instrument or evidence the assumption of fixed preferences is untenable. This study tested for question order effects in a CJ choice survey intended to elicit preferences for attributes of GM food products. Using estimates of WTP for changes in GM content as the measure of preferences, no question order effects were found. This particular survey passed the validity test, providing support for the use of CJ choice surveys as an alternative to CV surveys for the valuation of nonmarket goods. However, more research is needed to test for test for question order effects in other CJ formats.

Share