*2.4. Statistical Analysis*

The statistical analysis included descriptive statistics: mean (M), standard deviation (SD), 95% of confidence interval (CI) with lower limit (LL) and upper limit (UL). Subsequently, a one-way analysis of variance (ANOVA) was performed to test the differences in the mean scores of depression and anxiety between university students from the nine countries: Slovenia, Czechia, Germany, Poland, Ukraine, Russia, Turkey, Israel, and Colombia. The effect size for ANOVA was assessed using η<sup>p</sup> <sup>2</sup> (a value of η<sup>p</sup> <sup>2</sup> = 0.01 is considered to be a small effect size, 0.09 a medium effect, and 0.25 a large effect). Tukey's honest significant difference (HSD) test was used to find means that are significantly different from each other. Furthermore, Pearson's χ<sup>2</sup> independence test was conducted to examine relationships between depression and anxiety and other variables in each of the nine countries. A 2 × 2 contingency table was provided in each country separately, for depression and anxiety as independent variables, as well as such predictor variables as gender, place of residence, level of study, physical activity, exposure to the COVID-19 pandemic, the total impact of COVID-19 on students' well-being, as well as impact in the domain of qualifications, economic status, and social relationships, self-rated physical health, and comparative self-rated physical health (Comparative PH). However, all Colombian students (100%) were assigned to the Town/City category, and 97% (*n* = 155) to the first cycle study. Therefore, place of residence and level of study were excluded from the statistical analysis in the Colombian sample. The effect size for Pearson's χ<sup>2</sup> independence test was assessed using ϕ statistic (a value of ϕ = 0.1 is considered to be a small effect, 0.3 a medium effect, and 0.5 a large effect). Next, multivariate logistic regression analysis was performed in each country separately to test the adjusted odds ratio (AOR), in order to assess potential risk factors (gender, place of residence, exposure to COVID-19, PIC, PA, PH, Comparative PH) as predictors of depression and anxiety in each country. All predictors were entered into the model simultaneously. The following statistics were calculated for estimation: coefficient estimates, 95% confidence intervals (CI) for the regression coefficient, standard errors of the regression coefficient, odds ratio, *z*-values, and their corresponding *p*-values.

The bias-corrected accelerated bootstrapping (BCa) method of estimating regression coefficient was also applied, with the number of replications set to 5000 (if the bias-corrected 95% confidence intervals (CIB) did not include the null value, then a statistically significant effect was considered). Goodness of fit of the regression model was assessed using pseudo *R*2, including Cox and Snell *R*<sup>2</sup> CS, McFadden *R*<sup>2</sup> McF, and Nagelkerke *R*<sup>2</sup> N. All analyses were performed using Statistica Version 13.1, StatSoft Polska (Cracow, Poland) [68] and the open-source statistical software JASP Version 014.1 [69].
