*2.3. Statistical Analyses*

Shapiro–Wilk Normality tests (*p* > 0.05) revealed that all data were normally distributed. We used an ANOVA to test our prediction that males would be stronger than females and that the dominant hand (defined here as the hand used to write) would be significantly stronger than the non-dominant hand (dominant hand asymmetry) using, first, absolute grip strength and, second, relative grip strength (i.e., grip strength/hand area). A Levene's test was performed to test the homogeneity of variance between males and females and for both hands. An ANOVA was also used to assess the difference in absolute grip strength between both hands across age categories within (1) males, (2) females, and (3) right and left-handers (sexes pooled) (DA).

Next, we fitted four linear multiple regressions to predict the four outcome measures of: (1) male dominant hand, (2) male non-dominant hand, (3) female dominant hand, (4) female non-dominant hand. Our predictor variables were age, occupation, hand shape, hand preference, playing music, and playing sport. These six predictor variables were considered as fixed effects and grip strength was considered a random effect. The function "predictorEffects" from the package "effect" [85] was used to graphically represent the model effects. An ANOVA was performed for each model to statistically test the effect of the predictor variables on grip strength. Tukey corrections were used for post-hoc analyses. All tests were performed with R 3.6.3 [86] with level of significance set at *p* ≤ 0.05.
