Skip to main content

Table 2 Summary of construct validity coefficients

From: Construct-level predictive validity of educational attainment and intellectual aptitude tests in medical student selection: meta-regression of six UK longitudinal studies

Outcome

Predictor

A-levels

GCSEs/O-levels

Aptitude tests

First year BMS

.809 n = 3 (CI .501; .935)

.332 n = 3 (CI .024; .583)

.245 n = 1 (CI .207; .276)

BMS overall

.744 n = 4 (CI .518; .872)

.361 n = 4 (CI .305; .413)

.065 n = 1 (CI −.049; .180)

All BMS

.772 n = 7 (CI .627; .865)

.338 n = 7 (CI .205; .459)

.164 n = 2 (CI −.031; .347)

Finals

.625 n = 5 (CI .449; .754)

.400 n = 4 (CI .274; .513)

.226 n = 2 (CI −.093; .503)

All Undergraduate except first year BMS

.684 n = 9 (CI .561; .778)

.379 (n = 8) (CI .316; .439)

.147 (n = 3) (−.065; .346 )

All Undergraduate

.723 n = 12 (CI .616; .803)

.359 n = 11 (CI .255; .455)

.181 n = 4 (CI .055; .302)

MRCP(UK) Part 1 (written)

.661 n = 2 (CI .523; .765)

.433 n = 2 (CI .098; .680)

.168 n = 1 (CI .044; .308)

MRCP(UK) Part 2 (written)

.502 n = 2 (CI −.030; .812)

.372 n = 2 (CI .153; .555)

.358 n = 1 (CI .174; .559)

MRCP(UK) Clinical

.303 n = 2 (CI .010; .547)

.498 n = 2 (CI .068; .772)

.226 n = 1 (CI .007; .422)

All MRCP

.506 n = 6 (CI .301; .666)

.447 n = 6 (CI .300; .573)

.226 n = 3 (CI .108; .339)

Specialist Register

.627 n = 4 (CI .450; .756)

.119 n = 3 (CI −.044; .276)

.258 n = 2 (CI -.171; .605)

All postgraduate

.556 n = 10 (CI .426; .663)

.316 n = 9 (CI .148; .466)

.243 n = 5 (CI .090; .385)

All undergraduate and postgraduate

.656 n = 22 (CI .574; .726)

.342 n = 20 (CI .258; .420)

.208 n = 9 (CI .124; .289)

  1. Construct validities are combined meta-analytically as Fisher’s Z-transforms and then back-transformed to the conventional correlation scale. The number of construct validities and the 95% confidence interval (CI) for the construct validities are also shown. Where n = 1 the confidence intervals are those for the single estimate.