Commodity Modeling and Pricing: Methods for Analyzing Resource Market Behavior

CHAPTER 15

Public Input in Rural Land Preservation

Modeling Preference Asymmetries in Stated Preference Data

Robert J. Johnston and Kelly Giraud Cullen

INTRODUCTION

Public preferences for land use or other public policies often are elicited using variants of the common Likert scale, such as a scale from 1 to 5, where 1 = strongly oppose and 5 = strongly support (Bateman et al. 2002; Danielson et al. 1995; Kline and Wichelns 1998; Lynne, Shonkwiler, and Rola 1988; Variyam, Jorday, and Epperson 1990). Such scales provide information regarding a respondent’s strength of preference beyond a simple referendum, that is, a binary yes-or-no response. However, in return for the ability to model the increased information provided by Likert scales (LS), researchers often accept implicit assumptions not required when modeling binary responses. These include the assumption that respondents choose a cardinal, rather than a more basic ordinal, response on the continuum of the provided scale by reference to a single underlying preference function.

Despite the common use of simplifying assumptions when working with LS data, the literature addressing other choice contexts suggests that responses to such preference scales may be complex. For example, since Likert scales usually allow respondents to express varying degrees of support or opposition, there is the possibility that responses will manifest preference or response asymmetries. Response asymmetries formally imply that different factor ...