Identifying Bias in Self-Reported Pro-Environmental Behavior

Abstract

Research on pro-environmental behavior (PEB) informs social policies and interventions, so the quality of PEB measurement is critical. To estimate PEB, researchers often use self-reported measures. However, these measures, which are derived using survey data, often contain non-negligible measurement error, which can bias estimates and lead to incorrect research findings. Given the potential presence of error, we hypothesized that changes to the way self-reported PEB is measured might affect measurement error and therefore the validity of the results. Study 1 ( N = 951) showed that priming participants with related scales like environmentalist identity did not change reported behavior substantially (all d s ≤ .12). To investigate the presence of a particular type of systematic measurement error, namely social desirability bias, Study 2 ( N = 385) measured littering prevention behavior using the Unmatched Count Technique. A standard questionnaire format led to much higher reported behavior compared to the more anonymous covert condition, d = 0.52, and this effect seemed to be driven by participants who reported a stronger environmentalist identity, confirming the presence of social desirability bias. These results may help to explain the observed error in PEB measures and suggest that researchers could reduce social desirability effects with indirect questioning techniques.

Keywords: pro-environmental behavior, measurement error, question-behavior effect, social identity, social norms, social desirability bias