Predictive power of UKCAT and other pre-admission measures for performance in a medical school in Glasgow: a cohort study

Table 3 Bivariate associations (multiple regression analyses) between each pre-admission measure and composite course performance indicators* in models with and without adjustment for confounders – beta (95% confidence intervals, significance) and R ² of unadjusted model

	Educational performance measure		Honours & commendation^#
	Unadjusted	Adjusted^	Unadjusted	Adjusted^
UKCAT – Total score
Beta (95% CIs)	0.216 (0.074-0.357)	0.212 (0.067-0.358)	0.251 (0.110-0.392)	0.217 (0.070-0.364)
(sig)	*(0.003)*	*(0.005)*	*(0.001)*	*(0.004)*
R ²	0.047		0.063
UKCAT – Verbal reasoning
Beta (95% CIs)	0.213 (0.071-0.354)	0.219 (0.076-0.361)	0.201 (0.059-0.344)	0.170 (0.024-0.315)
(sig)	*(0.003)*	*(0.003)*	*(0.006)*	*(0.023)*
R ²	0.046		0.041
UKCAT – Quantitative reasoning
Beta (95% CIs)	0.219 (0.079-0.360)	0.237 (0.089-0.385)	0.216 (0.074-0.358)	0.205 (0.054-0.356)
(sig)	*(0.002)*	*(0.002)*	*(0.003)*	*(0.008)*
R ²	0.049		0.047
UKCAT – Decision analysis
Beta (95% CIs)	0.125 (-0.020-0.270)	0.097 (-0.049-0.243)	0.174 (0.030-0.319)	0.141 (-0.006-0.288)
(sig)	(0.091)	(0.192)	*(0.018)*	(0.059)
R ²	0.016		0.030
UKCAT – Abstract reasoning
Beta (95% CIs)	0.014 (-0.131-0.159)	0.012 (-0.132-0.156)	0.074 (-0.072-0.219)	0.053 (-0.093-0.198)
(sig)	(0.849)	(0.869)	(0.319)	(0.476)
R ²	0.000		0.005
Total science score
Beta (95% CIs)	0.136 (-0.012-0.283)	0.259 (0.092-0.427)	0.087 (-0.062-0.236)	0.148 (-0.024-0.320)
(sig)	(0.071)	*(0.003)*	(0.249)	(0.092)
R ²	0.018		0.007
Interview score ^#
Beta (95% CIs)	0.155 (0.012-0.297)	0.148 (0.006-0.291)	0.131 (-0.013-0.274)	0.119 (-0.026-0.264)
(sig)	*(0.034)*	*(0.041)*	(0.074)	(0.107)
R ²	0.024		0.017

*All pre-admission and course performance measures standardised as z-scores.
^#Honours & Commendation and interview scores transformed to reduce skew prior to standardisation.
^Adjusted for gender, age, ethnicity and deprivation. Note that R² is not included for the adjusted models, since our focus is on the variance explained by the pre-admission measures, not the additional variance explained by gender, age, ethnicity and deprivation.

ISSN: 1472-6920