Popular wisdom suggests avoiding votes for parties unlikely to win. Yet many voters do so—often supporting 'underdogs'. This study argues voters estimate their preferred party's chances via social networks.
Simulated Social Network Experiments
Researchers designed voting experiments where participants received simulated network signals about popular support, mimicking real-life political echo chambers. Two types of simulated networks were tested: highly homogeneous (homophilic) and random connections.
Key Findings
• Participants in homophilic networks overestimated their preferred party's electoral chances by 30-45%
• These groups showed significantly increased support for the underdog candidate, compared to those in random network conditions or control groups
• The findings mirror anecdotal evidence about echo chambers boosting fringe political candidates' viability perceptions
• This effect highlights how social media algorithms might unintentionally distort electoral behavior by creating filter bubbles
This Matters For Political Science
Our results demonstrate a previously undocumented psychological mechanism: voters use their networks to gauge probability, making them unreliable predictors.
This discovery helps explain why many citizens support parties they perceive as having less chance of winning.
