Exploring sociodemographic subgroup differences in multiple mini-interview (MMI) performance based on MMI station type and the implications for the predictive fairness of the Hamburg MMI

Table 4 Linear regression model predicting OSCE communication station performance (N = 400)

	Model 1			Model 2
Predictors¹	b	95%-CI	p	b	95%-CI	p
zMMI overall	2.15	0.94; 3.36	0.001	2.98	1.00; 4.96	0.003
Male	−0.004	−2.15; 2.14	0.997	0.84	−1.49; 3.16	0.481
21 and older	−0.59	−2.78; 1.60	0.597	0.03	−2.48; 2.53	0.984
At least one parent is a physician	−1.70	−3.94; 0.55	0.138	−2.36	−4.89; 0.16	0.067
Interactions
zMMI overall * Male				−1.91	−4.35; 0.54	0.127
zMMI overall * 21 and older				−1.10	−3.65; 1.45	0.396
zMMI overall * At least one parent is a physician				1.08	−1.48; 3.64	0.409
ΔR²	.035			.044

¹As only 43 out of 400 students did not speak German as their first language, the product term of zMMI overall and German as first language was too redundant to zMMI overall resulting in multicollinearity. Therefore, German as first language had to be excluded from the model

ISSN: 1472-6920