From "No Difference" to "Trending": misinterpretation of statistical analysis in animal studies.

The frequentist approach, with its emphasis on statistical significance, remains predominant in animal research. However, the interpretation and reporting practices associated with this approach have not been evaluated in the context of in vivo pharmacology. In this study, 100 recent publications from 5 journals were analyzed. Of these, 61 studies (95% CI 51-71) either misinterpreted the results of hypothesis testing or used terminology that implied a gradation of p-values, rather than adhering to a binary distinction between significant and non-significant outcomes. Additionally, only 1 study (95% CI 0.03-5%) reported both effect sizes and confidence intervals, while 99 (95% CI 95-100%) relied exclusively on p-values for interpretation. Furthermore, none of the studies had a publicly available protocol or registration, precluding clear identification of primary outcomes, planned sample size calculations, and strategies for controlling family-wise error rates. These findings highlight the need for researchers in preclinical pharmacology to report effect sizes and appropriate precision metrics, such as confidence intervals, and to develop and register study protocols. This would support more comprehensive interpretation of results, improve methodological transparency, and reduce reliance on dichotomizing results alone, which can potentially inflate false positive findings.