Factors Explaining Hypothetical Bias: How to Improve Models for Meta-Analyses
By: Atozou, Baoubadi; Tamini, Lota D.; Bergeronm, Stephane; Doyon, Maurice
Atozou, Baoubadi; Tamini, Lota D.; Bergeronm, Stephane; Doyon, Maurice, Factors Explaining Hypothetical Bias: How to Improve Models for Meta-Analyses, Journal of Agricultural and Resource Economics, Volume 45, Issue 2, May 2020, Pages 378-397
Using a set of 462 observations from 87 public and private goods economic valuation studies, this study reviews and updates meta-analyses on hypothetical bias using a metaregression hierarchical mixed-effect (MRHME) model that corrects the effects of the unobservable characteristics, within-study error correlation, and potential heteroskedasticity specific to each study. The findings indicate that the MRHME model is more efficient than the log-linear models used in previous meta-analyses. Moreover, this modeling approach and the use of interaction variables by type of goods highlight significant differences relative to previous meta-analyses in the explanatory variablesÕ effects, significance levels, magnitudes, and signs.