Skip to main content
. 2020 Sep 1;20:503. doi: 10.1186/s12884-020-03205-2

Table 4.

Fit indices for multiple-group measurement invariance tests of the Reproductive Agency Scale 17 (RAS-17) across Qatari and non-Qatari Arab pregnant women (N=684) attending prenatal appointments at Hamad Medical Corporation Maternity Hospital in Doha, Qatar

Model Free params χ2 (df) RMSEA 90% CI RMSEA CFI TLI ∆χ2 p-value χ2
M1: Configural invariance 116 577.88 (232) .07 .06–.07 .94 .93
M2: Full metric invariance 102 594.36(246) .06 .06–.07 .94 .93

M3: Partial metric invariance

(removed: fommovie)

103 575.76 (242) .06 .06–.07 .94 .94 20.80 .08
M4: Full scalar invariance 68 679.41 (280) .07 .06–.07 .93 .93

M5: Partial scalar invariance

(thresholds freed: univedu, wkouthm, dleavehm)

73 616.95 (275) .06 .05–.07 .94 .94 47.65 .02*
M6: Full strict invariance 90 603.55 (258) .06 .06–.07 .94 .94

M7: Partial strict invariance

(residual freed: wkouthm)

74 610.34 (274) .06 .05–.07 .94 .94 20.86 .18

Note: Params = parameter. Statistics ∆χ2 and p-valueχ2 are associated with difference testing. CFA uses WLSMV estimation, probit link. Mplus handles missingness with multiple imputation using Bayesian analysis

*p < .05