Correlation between description of results and effect size/statistical power.

<p>Description scores refer to the mean score given by 14 neuroscience researchers who rated terms as “weak” (0), “neutral” (1) or “strong” (2) in the case of those describing significant differences, or as “similar” (0), “neutral” (1) or “trend” (2) in the case of those describing non-significant ones. (A) Correlation between normalized effect size and description score for significant results. r = -0.05, p = 0.48 (n = 195). (B) Correlation between statistical power and description score for significant results. r = 0.03, p = 0.73 (n = 155). (C) Correlation between normalized effect size and description score for non-significant results. r = 0.28, p = 0.0002* (n = 174). (D) Correlation between upper-bound estimate of statistical power and description score for non-significant results. r = 0.03, p = 0.74 (n = 146). Asterisk indicates significant result according to Holm-Sidak correction for 28 experiment-level correlations.