# Behavioral science of voting in hypothetical utopia and dystopia scenarios: a predictive modeling approach

## TL;DR

This study explores how hypothetical utopian and dystopian scenarios influence how Finnish voters rate politicians, finding that personal and political factors have stronger effects than the scenarios themselves.

## Contribution

The paper introduces a predictive modeling approach combining machine learning and statistical analysis to study political judgment in hypothetical future scenarios.

## Key findings

- Party, gender, and familiarity of the candidate were the strongest predictors of voter valuations.
- Both utopian and dystopian scenarios generally reduced valuations of politicians, with a notable exception for right-wing politicians in the dystopian scenario.
- Liberality was linked to higher sensitivity to opinion change among women.

## Abstract

We investigated how near-future utopia and dystopia scenarios affect electorates' valuation of the leading Finnish politicians in the multiparty democracy. Using an online questionnaire with 1,109 respondents, we collected valuations across six attributes (e.g., trustworthiness, suitability as prime minister) both at baseline and after random assignment to a scenario. Valuations were first made for the present day and second in a near-future scenario. The strongest predictors for the electorates' valuations were the party, gender, and familiarity of the candidate, as well as liberal or conservative attitudes of the voter. Conversely, the effect of the scenarios was found to be weaker, and both scenarios tended to reduce the valuations of politicians. Devoted voters were less likely to change their opinion about their representative than non-devoted voters. The only exception to this was a notable increase in valuation of the right-wing politician in the dystopian scenario. In addition, we found that liberality was associated with higher sensitivity to opinion change among women. This study reveals the interplay between voter-, politician-, and context-related factors in political judgment. It also demonstrates the value of combining predictive machine learning with inferential statistical modeling to analyze complex behavioral data.