Preprint - The prevalence of marginal significance in psychology over time
We examined the proportion of p-values (.05<𝑝≤.1) reported as marginally significant in 44,200 articles across 9 psychology disciplines, published in 70 journals belonging to the American Psychological Association (APA) between 1985 and 2016. Using regular expressions we extracted 42,504 p-values between .05 and .1. Almost 40% of p-values between .05 and .1 were reported as marginally significant, though there were considerable differences between disciplines. The practice is most common in organizational psychology (45.4%) and the least common in clinical psychology (30.1%). Contrary to what was reported by Pritschet, Powell, and Horne (2016), we found no evidence of an increasing trend in any discipline; in all disciplines the percentage of p-values reported as marginally significant was decreasing or constant over time. The 'Journal of Personality and Social Psychology' (JPSP), also examined by Pritschet et al., was an exception to the general trend and showed an increase over time. The degree to which reporting results as marginally significant is problematic depends largely on individual interpretation. Due to the low evidential value of p-values between .05 and .1 we recommend against reporting these results as marginally significant.