The Academic Backbone in the UCLMS Cohort Study. This figure, and Figures 2, 3, 4, and 5, which show path analyses of the Academic Backbone in the various cohorts, have the same structures and conventions, and are also very similar to the figures in the additional material (see Additional files). Pale blue boxes indicate measures obtained prior to medical school, usually at secondary school, pale green boxes indicate performance at medical school, and pale purple boxes indicate post-graduate performance. The path model was fitted using multiple regression, each variable being regressed on all variables to its left (that is, causally prior), using backwards regression, variables being eliminated sequentially until all remaining variables were significant with P<0.05. Path coefficients are shown as β coefficients (that is, they are standardized), and arrow thickness is proportional to effect size. Solid black arrows indicate positive β coefficients. Solid arrows entering or leaving secondary school measures are in grey to indicate that they are not accurate estimates of the true effect in the non-selected population. No paths were found which were significant and had negative β coefficients. When interpreting path models, it should be remembered that any analysis towards the right of the diagram takes account of prior effects occurring to the left of the diagram. For this figure, that means, for instance, that the effect of BMS marks on MRCP(UK) Part 1 mark takes into account and is additional to the effect of clinical marks on MRCP(UK) Part 1 mark. Abbreviations: A-level, Advanced level; GCSE, General Certificate of Secondary Education; MRCP(UK), Membership of the Royal College of Physicians of the United Kingdom; PACES, Practical Assessment of Clinical Examination Skills; UCLMS, University College London Medical School.
McManus et al. BMC Medicine 2013 11:242 doi:10.1186/1741-7015-11-242