Consumer choice models applied on a biobased product

Abstract

Biobased products are non-food/non-fuel goods derived from plants and other renewable agricultural and forestry materials. Everyday household products made from biobased materials could replace products made with petroleum-based inputs. This research estimates consumer willingness to pay (WTP) for extrinsic attributes (Source, Biodegradable, Biobased label, and Origin) in single-used eating-ware (SUEW) made from a biobased product using data from nationally representative surveys.

Study I applies the generalized multinomial logit (GMNL) model based on the data collected by the conjoint surveys. A split sample survey is used to address availability bias and attention bias. Study II extends a rank ordered logit (ROL) regression to a generalized rank ordered logit model (GROL). Best-worst scaling data collected from a survey is used in this study. Study III extends the hybrid mixed choice model (HMC) to a hybrid generalized multinomial logit (HGML) model based on the data collected from a conjoint survey. The choice models of studies 2 and 3 are estimated using Bayesian procedures.

Results suggest that the rapidity of product degradability and using non-plastic materials to make SUEW plates were valued most by consumers. Availability bias (e.g., the order of information provided to participants before choice set completion) did not affect WTP estimates in study I while did affect WTP estimates in study II. Inattention bias decreases consumer WTP for SUEW plate study I while did not affect WTP estimates in study II. This difference may be due to differences in the sample, on differences in the methods used to elicit WTP.