Table 2. Results from the univariable and multivariable linear regression analyses predicting CSA performance from the scores from the three selection measures (CPS, SJT, and SC) on the multiply imputed study data (m = 10) for the whole sample and for different subgroups of applicants. The last three rows also report the results from analysis of the non-imputed data

Selection assessment	Coefficient (β)	P value	Lower 95% CI	Upper 95% CI	R² for the model^a
Univariable results
Clinical problem solving	0.17 (0.51)	<0.001	0.16	0.17	0.26
Situational judgement test	0.18 (0.56)	<0.001	0.17	0.19	0.31
Selection centre	0.91 (0.39)	<0.001	0.83	0.98	0.15
Multivariable results
All applicants
Clinical problem solving	0.09 (0.28)	<0.001	0.08	0.10	0.40
Situational judgement test	0.11 (0.34)	<0.001	0.10	0.12
Selection centre	0.43 (0.18)	<0.001	0.33	0.53
Applicants scoring below the first quartile on the MSRA
Clinical problem solving	0.06 (0.15)	<0.001	0.05	0.07	0.12
Situational judgement test	0.08 (0.21)	<0.001	0.07	0.10
Selection centre	0.42 (0.21)	<0.001	0.32	0.51
Results from non-imputed data
Clinical problem solving	0.10 (0.27)	<0.001	0.09	0.11	0.30
Situational judgement test	0.12 (0.32)	<0.001	0.10	0.13
Selection centre	0.47 (0.17)	<0.001	0.37	0.56

^aThis is the mean R ² for the models derived from the imputed data. CPS = clinical problem-solving. CSA = Clinical Skills Assessment. SC = selection centre. SJT = situational judgement test.