Voting bias: switches in the neighbour effect as a function of vote valence

<p>Abstract: Background/Aims:- We demonstrated at SABE, Atlanta, that contestants on the TV game-show The Weakest Link were prone to a particular and profound voting bias; contestants significantly avoided picking their direct neighbour(s) when called upon to identify which one of their fellow contestants should be labelled as the worst performer, the so called weakest link (Goddard, Hylton, Parke & Noh, 2013). We set out to test whether the neighbour effect would emerge in other voting paradigms too, or if it was just a peculiar artefact of the hothouse, rarefied atmosphere of the TV game-show.Procedure:- In order to resolve this, the voting choices of ‘freshers’ were recorded during one of their induction/orientation lectures ( first year undergraduates, n=233). Each participant was asked to vote for just one of their fellow participants in the same seating row and block in the lecture theatre. Their vote conferred lottery tickets to the candidate for entry into a prize draw to win course related materials. However, vote valence varied as to whether their vote increased, decreased or had no effect on the candidate’s stack of lottery tickets. The observed frequencies of votes were counted for each voter-candidate spatial relationship. These observed frequencies were then compared with the frequencies that would be expected purely due to chance alone.Findings:- When voters issued a negative vote, they demonstrated a significant neighbour effect by avoiding voting for their nearest neighbours. This neighbour effect reversed polarity however, when the vote valence switched to positive. So in the condition when their vote benefitted the candidate, voters became significantly more likely to target their neighbours.Discussion:- Given that i.) when asked, participants expressed that they were making vote choices at random, and ii.) participants’ votes were closed (private), it was something of a surprise that the ensuing significant biases demonstrated such a large effect (?2 (10) = 49.31 p < .001, Cramer’s V = 0.15 (medium effect size)). Insodoing,we confirmed that the neighbour effect is a robust and strong bias in decision-making, operating at an unconscious, implicit level. We consider the difference in voting in the negative and positive valence conditions in terms of the do-no-harm principle (Baron, 1995).</p>