Table 2.

Differential item functioning with logistic regression for CBD items.

Item	χ2 (M2-M1)	χ2 (M3-M2)	R2Nagelkerke(M2-M1)	R2Nagelkerke M3-M2
1	< 0.01	< 0.01	.0018	.0004
2	< 0.01	< 0.01	.0028	.0012
3	< 0.01	< 0.01	.0015	.0040
4	< 0.01	< 0.01	.0029	.0004
5	0.083	0.142	.0020	.0006
6	< 0.01	< 0.01	.0004	.0001
7	< 0.01	< 0.01	.0017	.0004
8	0.001	0.003	.0039	.0004
9	< 0.01	< 0.01	.0001	.0025
10	< 0.01	< 0.01	.0012	.0026
11	< 0.01	< 0.01	.0037	.0020
12	< 0.01	< 0.01	.0005	< .0001
13	< 0.01	< 0.01	.0045	.0009
14	0.001	< 0.01	.0040	.0025
15	< 0.01	< 0.01	.0003	.0002
16	< 0.01	< 0.01	.0026	.0001
17	0.146	0.303	.0020	.0002
18	< 0.01	< 0.01	.0073	< .0001
19	< 0.01	< 0.01	.0067	.0006
20	< 0.01	< 0.01	.0087	.0006
21	0.012	0.004	.0013	.0030

Note. Models of logistic regression were adjusted. M1 = model 1; M2 = model 2; M3 = model 3. In all of them the dependent variable was sex. In M1 the predictor variable was total score in the test, in M2, response to the item and total score and in M3, total score, response to the item and the interaction between them.