Skip to main content

Table 3 Fit indices of 1-factor BSEM measurement invariance models for the Urbanicity Scale

From: Longitudinal measurement invariance in urbanization index of Chinese communities across 2000 and 2015: a Bayesian approximate measurement invariance approach

Model specification Prior variance pD 2.5%
PPL
97.5%
PPL
RMSEA CFI ΔRMSEA ΔCFI
2000–2015 (N = 301):
Configural / 389 631.5 1101.6 0.000 1.000 / /
 Metric Exact 309 1123.6 1584.5 0.024 0.975 + 0.024 −0.025
Metric V = 0.01 323 819.4 1276.3 0.012 0.994 + 0.012 − 0.006
 Scalar Exact 268 2505.7 3004.3 0.050 0.887 + 0.038 − 0.107
 Scalar V = 0.01 280 1791.9 2250.2 0.039 0.933 + 0.027 −0.061
 Scalar V = 0.04 302 1070.4 1518.8 0.022 0.978 + 0.010 −0.016
Scalar V = 0.09 311 897.6 1350.6 0.016 0.989 + 0.004 −0.005
2004–2011 (N = 295):
Configural / 214 264.0 564.1 0.012 0.996 / /
 Metric Exact 173 450.5 732.2 0.027 0.979 + 0.015 −0.017
Metric V = 0.01 181 329.5 619.5 0.018 0.990 + 0.006 −0.006
 Scalar Exact 144 1062.4 1368.9 0.052 0.920 + 0.034 −0.070
 Scalar V = 0.01 160 580.9 869.3 0.034 0.966 + 0.016 −0.024
Scalar V = 0.04 171 380.3 674.0 0.023 0.985 + 0.005 −0.005
2006–2009 (N = 220):
Configural / 84 41.6 170.5 0.046 0.972 / /
Metric Exact 72 66.9 192.1 0.050 0.965 + 0.004 −0.007
 Scalar Exact 61 220.9 339.1 0.071 0.928 + 0.021 −0.037
Scalar V = 0.01 66 103.2 229.1 0.056 0.956 + 0.006 −0.009
  1. PPL posterior predictive limit, pD Estimated number of parameters, RMSEA root mean square error of approximation, CFI comparative fit index; Δ = change in fit index over the previous nested (bolded) model