Skip to main content

Table 3 Linear mixed models regression of national and parental unemployment in relation to life satisfactiona of 11, 13 and 15-year olds (n = 386,402) in 27 countries across 3 waves (2002, 2005, 2009)b

From: The interplay between national and parental unemployment in relation to adolescent life satisfaction in 27 countries: analyses of repeated cross-sectional school surveys

  n (%)Empty modelModel 1p-valueModel 2p-valueModel 3p-value
Constant  7.589.13< 0.0019.12< 0.0019.12< 0.001
Fixed effects
 Country overall unemployment  −0.040.011−0.040.014− 0.040.016
 Year-specific unemployment  0.010.2130.010.1270.010.119
 Sex (girl) 197,892 (51.2) −0.18< 0.001−0.18< 0.001− 0.18< 0.001
 Age11 years127,083 (32.9) 1 1 1 
13 years132,581 (34.3) −0.45< 0.001− 0.45< 0.001−0.45< 0.001
15 years126,738 (32.8) −0.77< 0.001−0.77< 0.001− 0.77< 0.001
 Wave2001/02107,786 (27.9) 1 1 1.00 
2005/06142,024 (36.8) 0.050.2120.050.2290.050.234
2009/10136,592 (35.4) 0.040.3270.040.3370.040.350
 Family situationParents separated97,219 (25.1)   −0.46< 0.001−0.46< 0.001
 Unemployed parentsNone365,848 (94.7)   1 1 
Father7156 (1.9)   −0.38< 0.001− 0.170.007
Mother12,294 (3.2)   −0.21< 0.001−0.25< 0.001
Both parents1104 (0.3)   −0.49< 0.001−0.080.636
Interactions
 Country unemployment * Father unemployed     −0.020.001
 Country unemployment * Mother unemployed     0.0040.547
 Country unemployment * Both parents unemployed     −0.040.005
Random effects
 Country variance (% change from previous model)d0.0760.049 (−34.9%)0.049 (−1.1%)0.049 (−0.2%)
 Country-year variance (% change from previous model)d0.0150.016 (8.7%)0.016 (−0.8%)0.016 (−0.4%)
 School variance (% change from previous model)d0.1640.068 (−58.6%)0.067 (−0.5%)0.067 (0%)
 Individual variance (% change from previous model)d3.2963.217(−2.4%)3.213(−0.1%)3.213 (0%)
 Log likelihood− 784,253.4− 779,415.8− 776,821.08−776,811.3
 Likelihood ratio testcp < 0.001p < 0.001p < 0.001p < 0.001
  1. aLife satisfaction is self-reported on a scale of 1–10, with higher values representing better life satisfaction
  2. b27 countries (n: average 14,311.2, range 7737-41,906); 74 country waves (n: average 5221.6, range 2804-14,497); 14,344 schools/classes (n: average 26.9, range 1–5389)
  3. cLikelihood ratio test compares each nested model to the previous model to see if it is a significantly better fit. The empty model is compared to a single-level regression
  4. dThe percentage change is obtained by dividing the difference in variance between a preceding model and a nested model by the variance in the preceding model
Back to article page

BMC Public Health

ISSN: 1471-2458

Contact us