A psychometric evaluation of the Gender Bias in Medical Education Scale

BMC Medical Education

Table 5 Model fit: confirmatory factor analysis measurement invariance across studies, gender, and program of study

	χ²	df	RMSEA	CFI	TLI	Δχ²	Δdf	ΔRMSEA	ΔCFI
Study Invariance
Configural	167	64	.072	.963	.948
Weak: FL	172	71	.067	.964	.954	5	7	−.005	−.001
Strong: FL + I	200	78	.071	.956	.949	28*	14	.001	.007
Strict: FL + I + R^a	209	88	.066	.957	.956	37*	23	.006	.006
Mean: FL + I + M	242	81	.080	.942	.936	75***	17	.008	.021
Gender Invariance
Configural	158	64	.070	.963	.948
Weak: FL	165	71	.066	.964	.954	7	7	−.004	−.001
Strong: FL + I	182	78	.066	.959	.953	24*	14	−.004	.004
Strict: FL + I + R^a	195	88	.063	.958	.957	37*	23	−.007	.005
Mean: FL + I + M	213	81	.073	.949	.943	55***	17	.003	.014
Program of Study Invariance
Configural	191	64	.080	.955	.937
Weak: FL	198	71	.076	.955	.943	7	7	−.004	.000
Strong: FL + I^a	238	78	.081	.944	.935	49**	14	.001	.011
Strict: FL + I + R	256	88	.078	.940	.939	65***	23	−.002	.015
Configural	282	81	.089	.929	.921	91***	7	.009	.026

Notes. Δ = Difference from configural model. ** p < .01; *** p < .001. ^a indicates best model. FL factor loadings, I item intercepts, R item residuals, M latent means

ISSN: 1472-6920