**3. Results**

A total of 919 adolescents (55.6% girls) with a mean age of 15.5 ± 1.3 years and a mean BMI of 22.0 <sup>±</sup> 3.9 kg/m<sup>2</sup> participated. With regard to the additional classification into three weight groups, 5.3% (n = 49) of students were in the underweight group, 75.3% (n = 692) in the normal weight group, and 19.4% (n = 178) in the overweight/obese group, respectively.

Regarding self-rated health status, 21.8% of adolescents stated very good health, 52.6% good, 21.9% moderate, and 3.7% poor/very poor health.

In total, 68% stated to have a TV or computer in the bedroom, and 42% participated in sports clubs. Additionally, 8.7% reported smoking, 59.8% consumed alcohol, and 42.2% reported suffering from back pain.

Mean reported daily electronic media use was 2.7 ± 1.7 h, mean reported weekly leisure time sports activity was 9.8 ± 4.9 h, and mean physical fitness (Z-value) of the cohort was 105.6 ± 6.2.

In Table 1, results of the univariate comparison of the four different health groups are presented. Significant differences between groups were found with regard to sex, BMI, weight status, electronic media in the bedroom, daily electronic media use, sports club participation, weekly leisure time sports activity, physical fitness, smoking, and prevalence of back pain.

**Table 1.** Group differences in factors between adolescents with a very good, good, moderate, and poor/very poor perceived health. Values are means with SD or prevalence (%).


Notes: Data are displayed as means <sup>±</sup> standard deviations or relative frequencies, as appropriate. <sup>a</sup> : Kruskal– Wallis H-test, <sup>b</sup> : chi-square-test, <sup>c</sup> : significant differences specified according to Bonferroni-corrected pairwise comparisons. BMI: body mass index, VH: group with very good perceived health, GH: group with good perceived health, MH: group with moderate perceived health, PH: group with poor/very poor perceived health. Bold values indicate significant differences.

In Table 2, results of the multiple multinomial regression model are presented. Compared to the group with very good perceived health (reference group), being female (OR = 1.57) significantly increased the odds for being in the group with good perceived health. Conversely, being older (OR = 0.84), having electronic media in the bedroom (OR = 0.57), showing higher physical fitness (OR = 0.93), and not reporting back pain (OR = 0.66) significantly decreased the odds for being in the group with good perceived health compared to the group with very good perceived health.

**Table 2.** Results of the multiple multinomial regression analysis with the dependent variable perceived health status.


Notes: R<sup>2</sup> (Nagelkerke) = 0.26, model chi-square (21) = 243.45, *p* < 0.001, B: unstandardized regression coefficient, SE: standard error, OR: odds ratio, 95% CI: 95% confidence interval, lb: lower bound, ub: upper bound. Bold values indicate *p* < 0.05. VH: group with very good perceived health (reference group), GH: group with good perceived health, MH: group with moderate perceived health, PH: group with poor/very poor perceived health.

Compared to the group with very good perceived health (reference group), the likelihood for being in the group with moderate perceived health significantly increased with higher BMI (OR = 1.11) and smoking (OR = 3.17). Conversely, being older (OR = 0.81), participating in a sports club (OR = 0.41), higher weekly leisure time sports activity (OR = 0.89), higher physical fitness (OR = 0.88), and not reporting back pain (OR = 0.44) was associated with a decreased likelihood to be in the group with moderate perceived health.

Similarly, the odds for being in the poor/very poor health group was significantly larger in adolescents with higher BMI (OR = 1.22) compared to the group with very good perceived health. Age (OR = 0.70), sports club participation (OR = 0.23), higher weekly leisure time sports activity (OR = 0.88), higher physical fitness (OR = 0.86), and not reporting

back pain (OR = 0.20), on the other hand, was associated with decreased odds for being in the group with poor/very poor health perceived health compared to the group with very good perceived health. Additionally, the chance for an increased daily electronic media use was significantly higher in the poor/very poor health group (OR = 1.30).
