AI Distribution by Incentive Level: Implications for Online Survey Research

TLDR

Findings show that higher payments (exceeding $1.20) may reduce bot participation and boost human responses, and lower payments (exceeding$ 1.20) may reduce bot participation and boost human responses.

Abstract

The rise of bot-generated survey responses threatens data integrity in online social science research. This study explores how incentive levels impact bot level activity on Amazon Mechanical Turk. Findings show that higher payments (exceeding $1.20) may reduce bot participation and boost human responses. Lower incentives may attract bots. Carefully calibrated incentives may improve data reliability. This research contributes to the academy by highlighting how incentive structures influence data quality and the importance of balancing cost and response authenticity.