Download data and study materials from OSF
Field period: 11/24/2008-6/26/2009
Models of voting behavior often assume that (1) the uncertainty surrounding candidates and the outcomes of voting decisions lead voters behave in a risk-averse manner (Davis, Hinich, and Ordeshook 1970; Enelow and Hinich 1981; Bartels 1986; Alvarez 1997), and (2) that this risk aversion creates a powerful advantage for the more well-known incumbent and makes the status quo nearly invincible (Berger et al. 2000). Despite these theoretical expectations, challengers to the status quo (in the form of candidates and policies) often do garner substantial support. This project employs a unique 7-question risk-acceptance battery to test the idea that support for challengers to the status quo can be explained at least in part by individuals' attitudes towards risk. We attempt to add to prior survey research in two major ways. First, our risk acceptance battery provides a measure that is more reliable and valid than the latent scales and smaller batteries used by authors. Secondly, our experimental design isolates the causal link between risk acceptance and incumbency bias by controlling for environmental factors and characteristics besides risk acceptance that may have driven prior results. Ultimately, we fail to find a definitive connection between risk orientation and preference for an incumbent, which implies that the incumbency label alone is not enough to draw the support of the risk averse.
H1a: Preference for a candidate will increase when the incumbency label is present, regardless of economic conditions.
H1b: Incumbency preference depends upon economic conditions. Incumbents will be rewarded in good times and punished in bad.
H2a: The risk averse will prefer an incumbent, whereas the risk accepting will prefer a challenger.
H2b: The effect of risk orientations will depend upon the reference point. The effect of risk orientations on preference for the incumbent will be even more pronounced in poor economic times.
A pre-stimulus questionnaire assessed risk orientations. The experiment consisted of a 3 (candidate type: Incumbent A and Challenger B, Challenger A and Incumbent B, or Open Seat) x 3 (economic conditions: Good, Poor, or No Information) factorial design; the stimulus presented subjects with short biographical sketches of two hypothetical candidates running for a nonpartisan city council seat. Respondents were randomly assigned to read one of nine possible combinations of the two factors being manipulated.
Following exposure to the treatment, subjects were asked to express which candidate they preferred.
Our results first suggest that simply being labeled as an incumbent does not induce support for a candidate. Simple incumbency bias, to the extent that it exists, does not derive from a label (the way that partisan or ideological bias might); instead if probably arises from certain knowledge that experiences, familiarity, etc. accrued from residing in a place where the incumbent has governed. Our results do, however, show that incumbents, even in a hypothetical scenario, do get punished for poor economic conditions and rewarded for good economic conditions. Finally, we are unable to determine the role that risk acceptance plays in the interaction between incumbency status and economic conditions. If anything, our results suggest that providing citizens with the mere label of "incumbent" is not enough to draw support from the risk averse.
Kam, Cindy D. and Elizabeth N. Simas. "Risk Orientations and Electoral Decision Making." Presented at the 2009 Annual Meetings of the American Political Science Association, Toronto, ON Canada.