Using involvement to reduce inconsistencies in risk preference elicitation

Empirical research aiming to elicit risk attitudes faces problems of within- and between-method inconsistencies, which reduce the explanatory and predictive power of risk research. In this paper, we investigate the relevance of context and task involvement on these inconsistencies. Our analysis is based on a sample of 244 German agricultural sciences students, which were performing an iterative multiple price list (iMPL) and a simple self-assessment question on risk preferences. We find that using a real life and subject context specific (here, agricultural) framing of the iMPL is leading to fewer within- and between-method inconsistencies. This is due to the fact that contextual framing has an increasing effect on task involvement (proxied with the time spent in the iMPL). Additionally, we find that contextual framing triggers the role of subjects' context involvement (proxied using an indicator for students' involvement in the agricultural domain). More specifically, both higher task and context involvement are found to decrease within-method inconsistency in the iMPL task. While also between-method inconsistency is decreasing in subjects' task involvement, we found no effect of context involvement. In conclusion, our results suggest that by framing a risk elicitation method according to the subjects' specific context, involvement can be triggered and inconsistencies and misspecifications can be reduced.