Hedges, mottes, and baileys: Causally ambiguous statistical language can increase perceived study quality and policy relevance.

Abstract

There is a norm in psychology to use causally ambiguous statistical language, rather than straightforward causal language, when describing methods and results of nonexperimental studies. However, causally ambiguous language may inhibit a critical examination of the study's causal assumptions and lead to a greater acceptance of policy recommendations that rely on causal interpretations of nonexperimental findings. In a preregistered experiment, 142 psychology faculty, postdocs, and doctoral students (54% female), ages 22-67 (M = 33.20, SD = 8.96), rated the design and analysis from hypothetical studies with causally ambiguous statistical language as of higher quality (by .34-.80 SD) and as similarly or more supportive (by .16-.27 SD) of policy recommendations than studies described in straightforward causal language. Thus, using statistical rather than causal language to describe nonexperimental findings did not decrease, and may have increased, perceived support for implicitly causal conclusions.