Good Is More Causal Than Bad: Correlation Framing Affects Perceived Causality
(A2026-132969)

AUTHORS

Yue Zhang, Erasmus University Rotterdam; Gabriele Paolacci, Rotterdam School of Management, Erasmus University

ABSTRACT

In the era of big data, vast amounts of information have become accessible to business analysts and scientists. Statistical associations involving individual characteristics (e.g., health conditions) are frequently disseminated through advertising and newsletters. However, many of these correlations are spurious, as they do not reflect causal relationships. Understanding when consumers believe these correlations reflect causality is important. Across ten studies (N = 6,907), we demonstrate that people are more likely to believe a correlation reflects causality when it is framed as presenting desirable outcomes (e.g., frequent jazz music listeners tend to have better memory than frequent pop music listeners) than when it is framed as presenting negative outcomes (e.g., frequent pop music listeners tend to have worse memory than frequent jazz music listeners). This effect occurs because people are more likely to construe a positive outcome as reflecting a broader pattern, leading them to perceive a broader range of outcomes as associated with the plausible cause, which further increases perceived causality. We demonstrate the underlying mechanism through two-stage causal-chain studies and by attenuating the effect through presenting additional correlations with opposite-valence outcomes. In addition, we show that the correlation framing effect extends to product choices.