*2.3. Analysis*

Intercorrelations between items were weak; so, all items were used separately as variables. However, using factor analysis, we were able to calculate two composite variables. The results of a principal axis factoring with six affective variables showed that a two-factor structure was consistent with the data, as only two of the eigenvalues were above 1.0 (i.e., 2.289 and 1.140). The factor analysis was repeated with a promax rotation to obtain rotated factor loadings. Only loadings greater than 0.4 were considered nontrivial. Stress, worry, sadness, and anger had nontrivial loadings (ranging from 0.416 to 0.788) on the first factor (variance explained = 28.330%). These four variables were averaged to form an index of negative experience (Kuder–Richardson 21 reliability coefficient = 0.589). Factor 2 had only two nontrivial loadings (both = 0.646), laughter and enjoyment (variance

explained = 9.513%). The two items for positive affect were averaged to form an index of positive experience (Kuder–Richardson 21 reliability coefficient = 0.592). Data were analysed using regression analysis to examine which variables significantly predict life satisfaction and ANOVA to examine group differences in life satisfaction. We used a standard or simultaneous regression model. In this model, all predictors enter the regression equation simultaneously; each is evaluated as if it entered the regression after all other predictors. In other words, each predictor is evaluated based on its unique contribution to predicting dependent variables, after other predictors' contributions are controlled for [42]. We used stepwise regression as a supplementary tool to filter out the best predictors from our long list of potential predictors. Stepwise regression helps reduce a long list of potential predictors to a manageable number of significant predictors to facilitate interpretation. While the simultaneous method retains all entered variables regardless of significance level, this method combines the forward and backward approaches to remove nonsignificant predictors. The variance explained by each predictor changes as more predictors are added to the equation. As more predictors enter the equation, a variable that has qualified for inclusion may lose some of its predictive power. In this case, the variable with "weakened" predictive power is removed using the stepwise procedure [43].

#### *2.4. Ethical Considerations*

This study used publicly available data and as such did not require further ethical approvals.
