Hypothetical Bias Mitigation Techniques in Choice Experiments: Do Cheap Talk and Honesty Priming Effects Fade with Repeated Choices?

Abstract

We design a choice experiment comparing policies that reduce agricultural nutrient pollution and harmful algal blooms in Lake Erie and administer it to Ohio residents using an online survey panel. We compare two treatments that have been found to mitigate hypothetical bias, cheap talk and honesty priming, with a neutral priming control. We find greater sensitivity to price among respondents during choices made immediately following the cheap talk intervention. As additional choices are made, the “wedge” between price sensitivity in the control and honesty prime treatments diminishes and eventually loses statistical significance. We find this effect in both our online choice experiments and among respondents to face-to-face choice experiments conducted by de-Magistris, Gracia, and Nayga. Our online implementation of an honesty priming intervention yields no significant change in price sensitivity compared to a control. While de-Magistris et al. implement an honesty priming intervention that fully mitigates hypothetical bias in a face-to-face setting, we show that this effect is also transient, and in later choice exercises we cannot reject the null hypothesis of equality between honesty priming and the control. Our results are robust to multiple specifications and suggest additional work is required to adapt priming interventions for online settings and to extend the effectiveness of popular hypothetical bias mitigation techniques when respondents face multiple choice tasks.