Table 7.
Comparisons of Fit Indices for Models of Depression, Social Anxiety, and E/PE
| Sample | Model | RMSEA | SRMR | AIC | DIC | χ2 (1) | p |
|---|---|---|---|---|---|---|---|
| Students | Higher-order E/PE | .092 | .080 | 63451.48 | 63675.71 | -- | |
| Lower-order E/PE | .090 | .074 | 63360.89 | 63597.24 | -- | ||
| Constrained Models | |||||||
| Sociability paths* | .091 | .080 | 63394.55 | 63628.87 | 35.7 | <.01 | |
| PE paths | .090 | .074 | 63359.31 | 63593.64 | 0.33 | .565 | |
| Ascendance paths* | .092 | .082 | 63426.32 | 63660.65 | 61.9 | <.01 | |
| Fun-Seeking paths* | .091 | .077 | 63376.08 | 63610.41 | 16.0 | <.01 | |
| Psychiatric Patients | Higher-order E/PE | .098 | .091 | 39146.52 | 39310.83 | -- | |
| Lower-order E/PE | .094 | .083 | 39062.74 | 39235.92 | -- | ||
| Constrained Models | |||||||
| Sociability paths | .094 | .085 | 39065.69 | 39237.40 | 3.65 | .056 | |
| PE paths | .094 | .085 | 39063.82 | 39235.53 | 3.70 | .055 | |
| Ascendance paths* | .096 | .098 | 39094.26 | 39265.99 | 32.8 | <.01 | |
| Fun-Seeking paths | .094 | .083 | 39061.26 | 39232.99 | 0.46 | .497 | |
| Combined | Higher-order E/PE | .095 | .086 | 102599.20 | 103142.16 | -- | |
| Lower-order E/PE | .092 | .080 | 102423.32 | 102981.16 | -- | ||
| Constrained Models | |||||||
| Sociability paths* | .092 | .083 | 102453.55 | 103008.91 | 25.9 | <.01 | |
| PE paths | .092 | .081 | 102422.42 | 102977.78 | 0.96 | .327 | |
| Ascendance paths* | .093 | .088 | 102525.48 | 103080.84 | 92.2 | <.01 | |
| Fun-Seeking paths* | .092 | .081 | 102427.13 | 102982.49 | 4.75 | .029 | |
Note. N = 350 students, 204 psychiatric patients. Higher-order E/PE = relations between symptoms and higher-order E/PE factor only (without any facets); Lower-order E/PE = relations between symptoms and each E/PE facet; PE = Positive Emotionality; RMSEA = Root mean square error of approximation; SRMR = Standardized root mean square residual; AIC = Akaike's information criterion; DIC = Draper's information criterion.
Models with facets constrained indicate that the correlations of the facet with Depression and with Social Anxiety were constrained to be equal. Combined models are multiple group analyses where all lower-order estimated correlations were constrained to be equal between student and psychiatric patient samples.
signifies a significant nested χ2 difference test, higher AIC, and higher DIC as compared to the (unconstrained) lower-order E/PE model for each sample. This indicates that the correlations of a given facet with Depression and with Social Anxiety are not equal.