Does statistical language constitute a "significant" roadblock to readers' interpretations of research results?

Abstract

Critics of statistical hypothesis testing mention how such tests mislead unsuspecting readers into believing that a statistically significant outcome is important and that a nonsignificant outcome is unimportant. Specifically, they mention how the term significant is misleading. In 3 experiments, the authors investigated whether statistical language influences readers' interpretations of research results. For statistically naive college students reading brief research abstracts, the term significant contributed somewhat to an overinterpretation of the results' importance. But for reading lengthier abstracts and articles, language per se was associated with no negative outcomes. Although the authors argue that significant language changes should not be mandated for quantitative research studies in scientific journals, if such changes are mandated, then use of the term statistical is recommended as a straightforward solution to the problem.