. Author manuscript; available in PMC: 2015 Feb 1.

Published in final edited form as: Qual Life Res. 2013 Jul 4;23(1):239–244. doi: 10.1007/s11136-013-0463-0

Table 2.

Results of testing measurement invariance of the pain behavior items across PROMIS Wave 1 and ACPA using MG-CFA

Measurement invariance models	Overall fit indices						Comparative fit indices

	χ²	df	CFI	TLI	RMSEA	RMSEA 90 % CI	Model comparison	Δ χ²	Δdf	p	ΔCFI
1. Configural	3,453.968	648	0.904	0.896	0.086	0.083–0.089	1 versus 2	428.170	27	<.01	Δ 0.00
2. Metric	3,486.512	675	0.904	0.900	0.084	0.081–0.087	n.a.	n.a.	n.a.	n.a.	n.a.
3. Scalar	9,085.440	771	0.716	0.742	0.135	0.133–0.138	n.a.	n.a.	n.a.	n.a.	n.a.

Measurement invariance tests included 27 items after deleing PB2, PB9, PB16, PB23, PB24, PB29, PB31, PB43, and PB53 items from the initial 36 items CFI comparative fit index, TLI Tucker–Lewis index, RMSEA root mean square error of approximation, CI confidence interval, n.a. not applicable