Skip to main content

Table 3 Goodness-of-fit statistics for the SF-12 mental health (MH) and physical health (PH) sub-scales with and without correlated residual variances, stratified by type of joint replacement

From: Differential item functioning of the SF-12 in a population-based regional joint replacement registry

Model Joint Type RMSEA (90% CI) CFI TLI WRMR
MH Sub-scale
 No residual variance correlations THA 0.22 (0.21, 0.23) 0.97 0.94 4.77
TKA 0.21 (0.21, 0.22) 0.97 0.95 5.18
 With residual variance correlations: items M1 and M2, M3 and M5 THA 0.15 (0.14, 0.15) 0.99 0.97 2.47
TKA 0.14 (0.13, 0.15) 0.99 0.98 2.76
 With residual variance correlations: items M1 and M2, M3 and M5, M3 and M4, M4 and M6 THA 0.05 (0.04, 0.07) 1.00 1.00 0.62
TKA 0.03 (0.02, 0.04) 1.00 1.00 0.38
PH Sub-scale
 No residual variance correlations THA 0.19 (0.18, 0.20) 0.98 0.97 4.49
TKA 0.15 (0.14, 0.15) 0.98 0.97 3.95
 With residual correlations: items P2 and P3, P4 and P5 THA 0.03 (0.02, 0.04) 1.00 1.00 0.61
TKA 0.02 (0.01, 0.03) 1.00 1.00 0.52
  1. THA Total hip arthroplasty, TKA Total knee arthroplasty, RMSEA Root mean square error of approximation, CFI Comparative fit index, TLI Tucker-Lewis index, WRMR Weighted root mean square residual, CI Confidence interval; values in boldface font represent the fit statistics for the best-fitting models