Skip to main content

Table 4 Comparing shape of growth for interest and anxiety

From: Understanding why apprentices consider dropping out: longitudinal prediction of apprentices’ workplace interest and anxiety

LGM model tested Δχ2
χ2 df p-value CFI ΔCFI AIC
Interest       
Intercept only 188.428** 16   0.749   22,230
Linear (vs intercept) 29.092** 13 0.000 0.977 0.228 22,027
Quadratic (vs linear) 29.330** 12 0.626 0.975 0.002 22,029
Cubic (vs linear) 31.878** 11 0.248 0.970 0.007 22,035
Anxiety       
Intercept only 98.381** 16   0.784   25,011
Linear (vs intercept) 35.275** 13 0.000 0.942 0.158 24,931
Quadratic (vs linear) 37.383** 12 0.147 0.933 0.009 24,937
Cubic (vs linear) 35.520** 11 0.885 0.936 0.003 24,938
Piecewise 1 (vs linear)a 28.045** 13 0.961 24,921
Piecewise 2 (vs piecewise 1)b 23.618* 13 0.972 24,915
  1. Both quadratic models did not converge due to negative variance on linear slope, resolved by fixing the variance to zero. Both cubic models did not converge, resolved by holding the linear and quadratic variances to zero. Italicised entries indicate improved fit
  2. **p < .01, * p < .05
  3. aPiecewise model with linear slope1 at T1–T5 and slope2 at T6 = 0
  4. bPiecewise model with slope1 at T1 and T2 = 0; linear slope2 at T3–T5 and slope 3 at T6 = 0