How does informing laypersons about the “status quo” affect preferences regarding the objectives?: A study on anchoring bias in environmental multi-criteria decision analysis

Abstract

Multi-criteria decision analysis (MCDA) has been a powerful tool to support people in making complex decisions, especially in making environmental decisions. These decision problems often involve unfamiliar objectives, tradeoffs between objectives and translation from feelings to values, which means that preferences are most likely constructed. To familiarize decision-makers with the decision problem and assist them in constructing preferences, previous researchers recommend providing additional objective information (e.g., the range information, and the “status quo” information) to decision-makers. However, while it seems logical to inform novice decision-makers (laypersons) about the current situation, the so-called “status quo”, it is not confirmed yet whether the “status quo” information helps or biases the construction of preferences by introducing an anchor. This master thesis bridged the gap by conducting an experiment to explore the anchoring bias caused by “status quo” information in the weight elicitation process of environmental MCDA. The experiment followed a 4×2 between-participant design with two varying factors: “status quo” factor and the elicitation method factor. In addition, to explore a better design for online preference elicitation, three innovative features (i.e., attention question, consistency check questions, eliciting value functions) were designed and tested in the online survey. According to the results, the “status quo” information helped subjects understand the decision problem in both SWING elicitation and tradeoff elicitation. Besides, a significant deviation in weights was not found between the groups with “status quo” information and without “status quo” information. However, the involved levels did not cover the whole attribute range, meaning the results were not conclusive about the occurrence of an anchoring bias to the “status quo” information. Despite of the non-significant deviation between groups with and without “status quo”, the anchoring was observed with different “status quo” levels. The anchoring bias might have different forms in SWING elicitation (where the “status quo” is an indirect anchor) and tradeoff elicitation (where the “status quo” is a direct anchor). The anchoring in SWING elicitation could be associated with the adjustment in perceived range, while the anchoring in tradeoff elicitation could be associated with the direct anchoring in “tradeoff judgment” step. Among the innovative features, attention questions and consistency check questions were proved effective, while the elicited value functions showed relatively low reliability. Five recommendations were given for future weight elicitation: 1) providing the “status quo” information and using a substitute reference point when needed, 2) designing attention questions based on the type of information to emphasize in online surveys, 3) designing consistency check questions with appropriate instructions in online surveys, 4) using plain language in the online survey, and 5) designing a user-friendly online interface.