Elicitation Bias in Multiple Price Lists: A Field Experiment

Abstract of the Paper

Multiple Price Lists (MPLs) or Choice Lists (CLs) are widely used to elicit risk and time preferences, yet are prone to cognitive biases, particularly among respondents with limited numeracy skills. This paper compares three elicitation approaches; row-by-row from the top, from the bottom, and a Rapid Elicitation (RE) method using random starting points; in a field experiment with 906 rural Malawian farmers. With 20 MPLs per subject, we estimate starting point and order biases in switch points using nonparametric and parametric methods. Row-by-row elicitation from the top or bottom introduces significant bias in preference elicitation, with effect sizes of up to 0.4 standard deviations. In contrast, the RE approach yields significantly lower starting point bias (Cohen’s d of 0.08 or less). Order effects were present but smaller in magnitude. RE also reduced cognitive load and shortened response time. These findings underscore the importance of the elicitation method in experimental design, particularly in low-literacy settings. The RE method offers a more reliable and scalable tool for eliciting behavioral preferences in development economics.