**3. Results**

*3.1. Variations in the Average Level of LS among Adolescents*

Table 1 shows the distribution of demographic characteristics. The distribution of sex was quite equal across the three measurement points, whereas the distribution of age groups varied more across time points. When looking at the total mean scores on LS across time points (Table 2), the highest score was in 2016, followed by 2022, and the lowest was in 2011. When looking at the mean scores on LS for sex (Table 2), they were all above the neutral point of the scale (≥20) at all time points, where boys reported higher scores than girls. When looking at the age groups, the highest mean scores on LS were in the age group of 13–14-year-olds at all time points, followed by the age groups 15–16 years old and 17–20 years old. The mean stress scores were moderately high across all three time points; however, the stressors related to school performance, school/leisure conflict, and peer pressure showed the highest scores. When looking at self-rated health, the mean scores were above the neutral point of the scale at all time points.

#### *3.2. Associations between LS and Time, Sex, Age, and Stress Domains, Controlled for Self-Rated Health*

Table 3 presents the results from the multiple linear regression analysis for associations between sex, age, time point and stress domains and the criterion variable LS. When looking

at time, those participating in 2016 reported significantly higher LS than those in 2011 in the unadjusted model (β = 0.12) and in the model adjusting for sex and age (β = 0.16). No significant difference in LS was found between 2011 and 2016 when controlling for self-rated health and stress domains (β = 0.04). No differences were found in LS between timepoint 2022 and 2011 in the unadjusted or adjusted results. Sex was significantly associated with LS in both the unadjusted model (β = 0.14) and when controlling for age (β = 0.14), where boys scored higher than girls. However, no sex differences were found when controlling for time, stress domains, and self-rated health (β = 0.03). Age showed a negative curvilinear association with LS in all the regression models, where levels of LS declined from 13 to 18 years and increased from the age of 19 to 20 years. The results showed that all stressor domains were significantly and negatively associated with LS in all the models; however, the strongest associations were found for stress of peer pressure (β = −0.27) and home life (β = −0.26), followed by school performance (β = −0.23) and school attendance (β = −0.23), controlled for sex, age, time, and self-rated health.

**Table 3.** Multiple linear regression analysis of associations between sex, age, time, stressor domains, and life satisfaction.


Note. Unadjusted analyses present bivariate estimates. a Model: adjusted for sex/age. b Model: adjusted for sex, age, time, stress, and self-rated health. Sex: girls—0 and boys—1. Time dummy—2011 is the reference category. Age-squared test curve linearity of age. n = 1943, R<sup>2</sup> = 0.32; Cohen's f2 = 0.47; \* *p* ≤ 0.05; \*\*\* *p* ≤ 0.001.

#### *3.3. Interaction Effects of Sex* × *Time, and Sex* × *Stress Domains in Association with LS*

When looking at the interaction effects between sex × time and sex × stressor domains, there were significant interactions between sex × peer pressure, sex × home life, sex × school attendance, and sex × school performance, with stronger associations for girls (Table 3). The interaction between sex and time was not significant.
