Skip to main content
. 2018 Mar 22;13(3):e0194021. doi: 10.1371/journal.pone.0194021

Table 4. Fit statistics for invariance assessment.

Model X2 df p Δ X2 Δ df p RMSEA (90% CI) CFI ΔCFI TLI ΔTLI Pass?
Measurement Model Estimates
Non-psychiatric 178.24 150 .042 (.000-.065) .948 .934
Schizophrenia 169.89 150 .034 (.000-.057) .960 .949
Configural Invariance 343.84 297 .032 .038 (.012-.054) .955 .942 Yes
Weak Invariance 385.63 337 .035 .036 (.011-.052) .953 .002 .947 .005 Yes
Strong Invariance 403.26 354 .036 .036 (.010-.051) .952 .001 .949 .002 Yes
Latent Model Estimates
Latent Variance 405.017 357 .040 1.757 3 .624 .035 (.008-.051) .953 .001 .950 .001 Yes
Latent Means 597.823 360 < .001 192.806 3 < .001 .077 (.066-.088) .769 .184 .757 .193 No

CFI = comparative fit index; TLI = Tucker-Lewis; RMSEA = root mean square error of approximation; df = degrees of freedom