Exploring Preferences for Poplar Biotechnology and Integrated Choice Latent Variable Models

  • Author / Creator
    Rollins, Curtis
  • This thesis is divided in two papers, each focusing on a distinct issue. In the first paper (Chapter 2), public preferences for poplar biotechnology are examined in western Canada. Using a sample of the public from British Columbia, Alberta, Saskatchewan, and Manitoba, respondents were asked to vote in a series of hypothetical referenda comparing proposed forest policies to the current policy. Proposed policies varied based on poplar breeding method (traditional, genomics, or genetic modification) and whether poplars may be used as a biofuel feedstock. British Columbians were least accepting of new policies promoting poplar biotechnology or genomics on public land, while Albertans were most accepting. Little variation in policy acceptance was noted between different breeding methods, though genetic modification is least preferred. Policies involving poplar-derived biofuel production were preferred to policies involving no biofuel production. Respondents who were more certain of their voting choice were less likely to prefer a proposed policy versus the current policy. The second paper (Chapter 3), examines an integrated choice latent variable (ICLV) model in comparison to other choice models. The ICLV model allows for simultaneous estimation of models of individuals’ latent attitudes and choices. While attitudes are often specified as observed covariates in the social sciences to gain explanatory power, this approach may result in biased estimates, which the ICLV model addresses. Using the Albertan data subset, the ICLV model is compared to a model specifying attitudes as observed variables (No ME model). No significant differences in effects on the choice outcome were detected between the No ME and ICLV models. However, the No ME model is prone to Type I and II errors when estimating the effects of demographic variables on the choice outcome. Thus, the largest benefit of the ICLV model seems to be its ability to estimate relationships between latent and observed variables.

  • Subjects / Keywords
  • Graduation date
  • Type of Item
  • Degree
    Master of Science
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.