The total sample consisted of 246 adolescents (mean age=15.83 years, SD=1.04); 93.9% (N=231) provided complete data. See Table 1 for demographic information.
Table 1. Demographic Characteristics of Adolescent Survey Respondents (N = 246)
|
|
Mean (SD)
|
n (%)
|
Age, years
|
|
|
15.83 (1.04)
|
|
Ethnicity
|
|
American Indian or Alaska Native
|
|
39 (15.9)
|
|
Asian
|
|
2 (0.8)
|
|
Black or African American
|
|
71 (28.4)
|
|
Hispanic or Latino
|
|
4 (1.6)
|
|
Native Hawaiian or Other Pacific Islander
|
|
3 (1.2)
|
|
White
|
|
129 (52.4)
|
|
Other Race/Mixed Race
|
|
3 (1.2)
|
|
Prefer not to answer
|
|
1 (0.4)
|
Gender Identity
|
|
Female
|
|
114 (46.3)
|
|
Male
|
|
130 9 (52.8)
|
|
Non-binary
|
|
2 (0.8)
|
Primary Caregiver Education Attainment (n = 242)
|
|
Did not complete high school
|
|
2 (0.8) (2)
|
|
High school diploma
|
|
3 (1.2)
|
|
Postsecondary vocational certificate
|
|
4 (1.6)
|
|
Associate degree
|
|
10 (4.1)
|
|
Bachelor's degree
|
|
52 (21.1)
|
|
Master's degree
|
|
91 (37)
|
|
Doctoral degree
|
|
80 (32.5)
|
n, number of cases; SD, standard deviation
|
Psychosocial outcomes
Table 2 provides means and standard deviations (SDs) for each psychosocial variable by gender identity, as well results from t-tests for mean differences between male and female adolescents.
Significant binary gender differences were found between male and female adolescents for the Wakefulness subscale of the ASWS (t=-3.09, p < 0.01), indicating more wakefulness for males than for females, and for the Anxious Anticipation subscale of the SCREEN (t=-2.25, p=0.03), indicating more anxious anticipation for males than for females. There were no other significant binary gender differences between male and female adolescents for any other measures (p-values > 0.05). No significant binary gender differences emerged for psychological stress, peer relationships, or pain (p-values > 0.29).
Table 2. Psychosocial Variables by Gender Identity
|
|
|
|
|
Total
(N = 246)
|
Female
(n = 114)
|
Male
(n = 130)
|
Non-binary
(n = 2)
|
Binary Gender Difference
|
|
|
Mean (SD)
|
N
|
Mean (SD)
|
n
|
Mean (SD)
|
n
|
Mean (SD)
|
n
|
t
|
p
|
Age
|
15.83 (1.04)
|
239
|
15.88 (1.09)
|
111
|
15.80 (.993)
|
126
|
14.50 (.707)
|
2
|
0.54
|
0.59
|
School Functioning Total
|
45.01 (18.94)
|
232
|
42.66 (18.76)
|
114
|
46.85 (19.09)
|
130
|
49.50 (2.12)
|
2
|
-1.67
|
0.10
|
|
Anxious Anticipation
|
12.78 (6.43)
|
241
|
11.80 (6.31)
|
111
|
13.67 (6.45)
|
128
|
10.50 (4.94)
|
2
|
-2.25
|
0.03*
|
|
Difficult Transition
|
11.01 (4.98)
|
241
|
10.33 (4.88)
|
109
|
11.50 (5.00)
|
130
|
16.50 (4.94)
|
2
|
-1.83
|
0.07
|
|
Interpersonal Discomfort
|
10.45 (3.92)
|
245
|
9.93 (4.18)
|
113
|
10.83 (3.63)
|
130
|
15.00 (2.82)
|
2
|
-1.78
|
0.08
|
|
School Avoidance
|
10.41 (5.16)
|
240
|
9.44 (5.08)
|
109
|
10.77 (5.17)
|
129
|
7.50 (4.94)
|
2
|
-1.98
|
0.05
|
Sleep Quality
|
3.42 (.778)
|
245
|
3.42 (.731)
|
113
|
3.46 (.789)
|
130
|
1.51 (.314)
|
2
|
-.401
|
0.69
|
|
Wakefulness
|
4.01 (1.14)
|
245
|
3.80 (1.40)
|
113
|
4.24 (.739)
|
130
|
1.00 (00)
|
2
|
-3.09
|
<0.01*
|
|
Falling Asleep
|
3.16 (1.22)
|
245
|
3.30 (1.25)
|
113
|
3.06 (1.18)
|
130
|
2.20 (1.41)
|
2
|
1.53
|
0.13
|
|
Going to Bed
|
3.09 (.888)
|
245
|
3.15 (.767)
|
113
|
3.07 (.963)
|
130
|
1.33 (.471)
|
2
|
.690
|
0.49
|
Psychological Stress
|
55.24 (12.05)
|
246
|
54.71 (11.55)
|
114
|
55.50 (12.45)
|
130
|
68.65 (7.99)
|
2
|
-.512
|
0.61
|
Peer Relationships
|
46.69 (8.59)
|
246
|
47.48 (10.2)
|
114
|
46.21 (6.61)
|
130
|
33.5 (12.09)
|
2
|
1.16
|
0.24
|
Pain Intensity
|
5.58 (2.74)
|
43
|
5.05 (2.91)
|
19
|
6.08 (2.60)
|
23
|
4.00 (0)
|
1
|
-1.21
|
0.23
|
Note: n=number of cases; SD=standard deviation; N=total number of cases; Sleep was measured by Adolescent Sleep-Wake Scale (Essner et al., 2015); Stress and Peer Relationships were measured by Patient Reported Outcomes Measurement Information System (Cella et al., 2010) Pediatric Psychological Stress (Bevans et al., 2018) and Pediatric Peer Relationships (Dewalt et al., 2013), respectively; School functioning was measured by the School Refusal Evaluation Scale (SCREEN) (Gallé-Tessonneau & Gana, 2019); pain was measured by the Pediatric Pain Screening Tool (Rau et al., 2021). *p<.05
|
Correlation analysis
School functioning and overall sleep quality were strongly positively correlated (r=0.73,
p<.001), indicating that better overall sleep quality was related to better school functioning. School functioning was also positively correlated with peer relationships (r=0.30, p<.001). In addition, peer relationships were negatively correlated with age (r=-0.22, p<.001). Average pain intensity was negatively correlated with school functioning (r=-0.76, p<.001), indicating increased pain levels were related to worse school functioning. There was a strong negative correlation between sleep quality and psychological stress, indicating worse sleep quality was associated with more psychological stress (r=-0.72, p<.001). Moreover, sleep quality was positively correlated with peer relationships (r=0.23, p<.001). Psychological stress was strongly positively correlated with pain levels (r=0.63, p<.001) and strongly negatively correlated with total school functioning (r=-0.60, p<.001). Further, psychological stress was negatively correlated with age (r=-0.16, p=.014). Psychological stress was not correlated with peer relationships (r=0.01, p=.916). Figure 1 presents the correlation heatmap. For full correlational output for all variables including their subscales, please see the Supplement.
Figure 1. Correlation Heatmap
Note: Sleep was measured by Adolescent Sleep-Wake Scale (Essner et al., 2015); Stress and Peer Relationships were measured by Patient Reported Outcomes Measurement Information System (Cella et al., 2010) Pediatric Psychological Stress (Bevans et al., 2018) and Pediatric Peer Relationships (Dewalt et al., 2013), respectively; School functioning was measured by the School Refusal Evaluation Scale (SCREEN) (Gallé-Tessonneau & Gana, 2019); pain was measured by the Pediatric Pain Screening Tool (Rau et al., 2021). To ease interpretation of correlations, we reverse coded the SCREEN so that higher scores indicated better school functioning. In doing this, for all measures presented above, higher scores represent more of the measured construct (i.e., higher sleep quality, more stress, better peer relationships, better school functioning, and more pain). Plot was created using corrplot (Wei, 2021).
Path analyses
As reported in our correlation results, associations of stress with sleep, pain, and school functioning were significant. However, peer relationships were not significantly associated with stress, therefore it was excluded from further path analysis (Baron, 1986). The structural equation model is displayed in Figure 2 and results are reported in Table 3. Stress displayed a significant negative effect on school functioning (b=0.57, p<0.001) and pain (r=0.26, p<0.001). Sleep quality accounted for 78.4% of the total effect of stress on school functioning (b=-0.45, p<0.001) and 54.2% of the effect of stress on pain (b=0.14, p=0.002). For both outcomes, the remaining direct effect of stress was no longer significant once the indirect effect of sleep quality was accounted for (school functioning: b=-0.12, p=0.07; pain: b=0.11, p=0.14).
Figure 2. Path Diagram Including All Genders (N=226) Note: Sleep was measured by Adolescent Sleep-Wake Scale (Essner et al., 2015); Stress and Peer Relationships were measured by Patient Reported Outcomes Measurement Information System (Cella et al., 2010) Pediatric Psychological Stress (Bevans et al., 2018) and Pediatric Peer Relationships (Dewalt et al., 2013), respectively; School functioning was measured by the School Refusal Evaluation Scale (SCREEN) (Gallé-Tessonneau & Gana, 2019); pain was measured by the Pediatric Pain Screening Tool (Rau et al., 2021). To ease interpretation of parameter estimates, we reverse coded the SCREEN in these analyses so that higher scores indicated better school functioning. In doing this, for all measures presented above, higher scores indicate more of the measured construct (e.g., higher sleep quality, more stress, better peer relationships, better school functioning, and more pain). Displayed parameter estimates are standardized and correspond to correlation coefficients. P-values were adjusted for multiple comparisons using Benjamini-Hochberg method. *p-value<0.05; **p-value<0.01; *** p-value<0.001.
Gender Specific Effects
In total, 101 female and 123 male identifying participants were included in the analysis of gender-specific effects. Male participants experienced a larger total effect of stress on school functioning than female participants (b=-0.31, p=0.002). Sleep quality accounted for 85.0% of the effect of stress for males and 87.7% of the effect for females.
There was no significant difference between male and female participants in the total effect of stress on pain (b<0.01, p=0.975). However, binary gender did significantly moderate the effect of sleep quality on pain such that the indirect effect of sleep quality on pain was significant for female participants (b=0.18, p=0.007; 69.6% of total effect), but not for male participants (b=0.08, p=0.39). The remaining direct effect of stress on pain was not significant for females (b=0.08, p=0.317).
Table 3. Effects of Stress on School Functioning and Pain: Indirect Effect of Sleep Quality
|
|
|
Total Effect [95% CI]
|
Direct Effect [95% CI]
|
Indirect Effect [95% CI]
|
% Accounted by Sleep Quality
|
School Functioning
|
|
Whole Sample
|
-0.57 [-0.66, -0.46]***
|
-0.12 [-0.26, 0.02]
|
-0.45 [-0.56, -0.35]***
|
78%
|
|
Female
|
-0.38 [-0.54, -0.17]***
|
-0.05 [-0.23, 0.19]
|
-0.33 [-0.48, -0.22]***
|
88%
|
|
Male
|
-0.70 [-0.81, -0.58]***
|
-0.10 [-0.29, 0.06]
|
-0.59 [-0.73, -0.44]***
|
85%
|
Pain
|
|
Whole Sample
|
0.25 [0.12, 0.39]***
|
0.12 [-0.04, 0.29]
|
0.14 [0.06, 0.23]**
|
54%
|
|
Female
|
0.26 [0.07, 0.45]*
|
0.08 [-0.07, 0.23]
|
0.18 [0.08, 0.31]*
|
70%
|
|
Male
|
0.26 [0.08, 0.46]*
|
0.18 [-0.09, 0.50]
|
0.08 [-0.11, 0.25]
|
30%
|
Note: Confidence Intervals and p-values estimated by bootstrapping. P-values are adjusted for multiple comparisons using Benjamini-Hochberg method. Significant effects are bolded and marked by asterisk: * p-value<0.05; ** p-value<0.01; ***p-value<0.001.