. Author manuscript; available in PMC: 2017 Jun 21.

Published in final edited form as: Psychol Test Assess Model. 2016;58(1):183–219.

Table 9. PROMIS anxiety short form item set: Differential item function (DIF) results. Gender and age subgroup comparisons.

Item description	IRTPRO			Lordif			Magnitude (NCDIF)			Effect Size T1
	Gender	Age		Gender	Age		Gender	Age		Gender	Age
	Gender	21-49 vs. 50-64	21-49 vs. 65-84	Gender	21-49 vs. 50-64	21-49 vs. 65-84	Gender	21-49 vs. 50-64	21-49 vs. 65-84	Gender	21-49 vs. 50-64	21-49 vs. 65-84
I felt fearful	NU	U^*	NU^; U^	NU; U^*	U^*	NU^; U^	0.0087	0.0112	0.0213	0.0770	0.0929	0.1219^†
I felt anxious		U^*	U^*	U^*	U^*	U^*	0.0040	0.0079	0.0069	0.0448	0.0733	0.0669
I felt worried		U^*	U^*	U^*	U^*	U^*	0.0030	0.0081	0.0178	0.0473	0.0819	0.1175^†
I found it hard to focus on anything other than my anxiety	NU			U^*	U	U^*	0.0068	0.0052	0.0072	-0.0564	-0.0547	-0.0604
I felt nervous				NU^; U^		NU; U^*	0.0001	0.0010	0.0044	0.0007	-0.0179	-0.0484
I felt uneasy		U	NU; U	U^*		U^*	0.0049	0.0052	0.0092	-0.0464	-0.0423	-0.0629
I felt tense		NU^; U^	NU^; U^	NU; U^*	U	U^*	0.0016	0.0017	0.0024	0.0213	0.0260	0.0244
My worries overwhelmed me			U	U^*		U^*	0.0011	0.0007	0.0008	0.0187	0.0210	0.0204
I felt like I needed help for my anxiety				U^*	U	U^*	0.0017	0.0062	0.0038	-0.0282	-0.0577	-0.0404
Many situations made me worry		NU	NU	U^*		U^*	0.0006	0.0014	0.0020	0.0108	-0.0223	-0.0288
I had difficulty calming down	U^*	U^*	U	NU	U^*	U^*	0.0126	0.0122	0.0182	-0.0810	-0.0809	-0.0991

All non-compensatory differential item functioning (NCDIF) values were smaller than the threshold (0.0960)

^†

Indicates value above threshold of 0.10; bolded values are above 0.15.

Asterisks indicate significance after adjustment for multiple comparisons.

NU= Non-uniform DIF involving the discrimination parameters; U=Uniform DIF involving the location parameters.

For the lordif analyses, the Uniform and non-uniform DIF was determined using the likelihood ratio chi-square test. Uniform DIF is obtained by comparing the log likelihood values from models one and two. Non-uniform DIF is obtained by comparing the log likelihood values from models two and three. DIF was not detected using the pseudo R2 measures or the change in Beta criterion.