Abstract
Attitudes and expectations towards others are major drivers of political polarization. However, there is limited understanding of their relevance when decisions with high stakes are taken. In this study, we compare self-reported attitudes against economically incentivized estimates of data coming from official sources and offer participants financial rewards for accuracy. Our methodology yields three principal findings. (i) Extreme attitudes from a small partisan subgroup primarily account for the observed partisan divide; this subgroup diminishes when incentivized estimates are considered. (ii) There is a weak correlation between economically incentivized and unincentivized measures within individual respondents. (iii) We introduce a novel metric for assessing perceived polarization. This metric allows participants to estimate data points for those with opposing political views, rewarding accurate predictions financially. Interestingly, this measure of perceived polarization correlates with attitudes but not with incentivized data estimates. This is in line with the concept of 'false polarization', attributing polarization more to expectations towards others than to genuine differences. These findings challenge the reliability of standard attitude surveys and suggest avenues for mitigating perceived polarization in contentious issues.