Implementing statistical equating for MRCP(UK) parts 1 and 2

BMC Medical Education

Table 3 Correlations of Part 1, Part 2 and PACES results, before and after equating of Parts 1 and 2

		Part 1 not equated and Part 2 not equated			Part 1 equated and Part 2 equated
		r	SE	N	r	SE	N	Significance
Part 1 predicting Part 2	UK	0.618	0.008	6347	0.651	0.009	3865	P = .007
	non-UK	0.553	0.008	7361	0.614	0.010	4150	P < .001
	*All*	*0.605*	*0.005*	*13708*	*0.623*	*0.007*	*8015*	*P = .040*
Part 1 predicting PACES	UK	0.305	0.010	7875	0.320	0.019	2172	P = .493
	non-UK	0.276	0.010	8648	0.283	0.027	1183	P = .807
	*All*	*0.385*	*0.007*	*16523*	*0.311*	*0.016*	*3355*	*P < .001*
Part 2 predicting PACES	UK	0.285	0.012	5957	0.270	0.020	2180	P = .516
	non-UK	0.233	0.012	5822	0.202	0.028	1195	P = .306
	*All*	*0.309*	*0.008*	*11779*	*0.211*	*0.016*	*3375*	*P < .001*

Pearson correlations (r) are shown separately for UK candidates, non-UK candidates and All Candidates (shown in italics), for candidates where both Part 1 and Part 2 were passed before statistical equating was introduced and for candidates for whom both Part 1 and Part 2 were statistically equated.
The significance of the difference in correlations is computed using a standard test [35], and rows where before and after correlations are significant are shown in bold.

ISSN: 1472-6920