Implications for cisgender female underrepresentation, small sample sizes, and misgendering in sport and exercise science research

Abstract

A sex-data gap, from testing primarily males, results in a lack of scientific knowledge for other groups (females, transgender individuals). It is unknown whether typical recruitment and participant characterization causes incorrect statistical decisions, and three factors were evaluated: 1) underrepresenting cisgender females, 2) recruiting small sample sizes, 3) misgendering. Data from the National Health and Nutrition Examination Survey (2003–2004) were evaluated for sex differences after removing missing values (N = 3,645; F = 1,763). Disparities were determined by utilizing sample sizes common in sport and exercise science research; mean sample size N = 187, median sample size N = 20. Participants were randomly allocated into datasets in an imbalanced manner (33.5% females, 66.5% males). Potential effects of misgendering were determined at rates of 2% and 5%. Differences between the complete data set and expected decisions were conducted through Chi-squared (χ2) goodness of fit with significance at p < .05. When the entire dataset was evaluated as if a sex testing disparity was present, decisions were not altered (χ2 = .52, p = .47). Differences were observed for mean sample size (χ2 = 4.89, p = .027), median sample size (χ2 = 13.52, p < .001), and misgendering at 2% (χ2 = 13.52, p = < .001) and 5% (χ2 = 13.52, p = < .001). Recruitment practices in sport and exercise science research should be revisited, as testing primarily cisgender males has consequences, particularly in small sample sizes. Misgendering participants also has consequences on ultimate decisions and interpretations of data, regardless of sample size. Inclusiveness is needed in helping all individuals feel valued and respected when participating in sport and exercise science research.