The educational background and qualifications of UK medical students from ethnic minorities
© McManus et al; licensee BioMed Central Ltd. 2008
Received: 20 July 2007
Accepted: 16 April 2008
Published: 16 April 2008
UK medical students and doctors from ethnic minorities underperform in undergraduate and postgraduate examinations. Although it is assumed that white (W) and non-white (NW) students enter medical school with similar qualifications, neither the qualifications of NW students, nor their educational background have been looked at in detail. This study uses two large-scale databases to examine the educational attainment of W and NW students.
Attainment at GCSE and A level, and selection for medical school in relation to ethnicity, were analysed in two separate databases. The 10th cohort of the Youth Cohort Study provided data on 13,698 students taking GCSEs in 1999 in England and Wales, and their subsequent progression to A level. UCAS provided data for 1,484,650 applicants applying for admission to UK universities and colleges in 2003, 2004 and 2005, of whom 52,557 applied to medical school, and 23,443 were accepted.
NW students achieve lower grades at GCSE overall, although achievement at the highest grades was similar to that of W students. NW students have higher educational aspirations, being more likely to go on to take A levels, especially in science and particularly chemistry, despite relatively lower achievement at GCSE. As a result, NW students perform less well at A level than W students, and hence NW students applying to university also have lower A-level grades than W students, both generally, and for medical school applicants. NW medical school entrants have lower A level grades than W entrants, with an effect size of about -0.10.
The effect size for the difference between white and non-white medical school entrants is about B0.10, which would mean that for a typical medical school examination there might be about 5 NW failures for each 4 W failures. However, this effect can only explain a portion of the overall effect size found in undergraduate and postgraduate examinations of about -0.32.
In the UK there is a concern that medical students and doctors from ethnic minorities underperform both in undergraduate and postgraduate examinations. In this paper we look firstly at a large population study of the educational achievement of all UK sixteen year-olds (the Youth Cohort Study), which reflects the pool from which all future UK medical students will be drawn, and we relate those results to a separate study based on data from UCAS (Universities and Colleges Admissions Service), looking at the educational qualifications of all university applicants, some of whom applied to medical school, and a proportion of whom were then accepted to study medicine. The primary interest throughout is whether, where, when and why ethnic minority students underperform relative to white students, and the extent to which any differences might explain the underperformance of medical students from ethnic minorities.
Medical students in UK universities are ethnically diverse, so that in recent years about 30% of home students (i.e. those resident in the UK, with UK nationality) come from ethnic minorities. That proportion has risen , surveys by one of us finding that about 10% of 1981 entrants were non-white, a figure which rose to 14% for those entering in 1986, and 22% for those entering in 1991 [2–5]. Official data from UCAS for 1996, 2001 and 2005, show that 30%, 33% and 30% of entrants for medicine and dentistry were non-white. (It should be noted that the statistics provided at the UCAS website  do not allow a separate analysis of medicine and dentistry in relation to ethnic group. Based on the datafile described below, the proportion of non-white entrants for medicine alone in 2005 was 31.2%, a value very similar to the UCAS figure for medicine 29.7% for medicine and dentistry combined. In the 2001 census, about 13% of those aged 15, 17% of those aged 16–19 and 19% of those aged 20–24 were from ethnic minorities, suggesting that ethnic minorities are somewhat over-represented amongst medical students compared with their proportion in the population as a whole.
In 1995, controversy erupted in the UK because of a higher failure rate in clinical examinations of non-white students in the University of Manchester [7, 8]. As a result of that controversy, a re-analysis of data from the 1981 and 1986 cohorts found that amongst students taking University of London final examinations, the non-white students performed less well overall than did white students, the difference being found in all types of examination (multiple-choice questions, essay, oral and clinical), and in all the five main subjects (Medicine, Surgery, Obstetrics and Gynaecology, Pathology and Pharmacology) . The reasons for the differences were not clear, and could not be explained away in terms of a range of background measures, be they demographic, educational or psychometric.
In the past ten years, further data have accumulated on the relatively poorer performance of UK non-white medical students and doctors in undergraduate and postgraduate examinations, non-whites performing less well in undergraduate clinical examinations [9–11, 11–14], as well as in the postgraduate examinations of the Royal Colleges of Physicians (the MRCP(UK) [15, 16]), and the Royal College of General Practitioners (MRCGP [17, 18]).
The published studies comparing the undergraduate and postgraduate performance of white and non-white medical students and doctors can be compared by calculating d, a standard meta-analytic measure of effect size techniques (see Method, below, for details). Effect sizes are a standard statistical method for comparing disparate results across studies, by converting all results to a standard scale similar to z scores. When there are two groups, then the effect size is a measure of the distance apart of the two means measured in standard deviation units. A conventional interpretation of effect sizes, due to Cohen, is that effect sizes of .2, .5 and .8 can be described as small, medium and large . Considering the published studies described above, the effect sizes were -.097 and -.283 , -.237, -.291, -.580 and -.273 , -.665 , -.579 , -.492 , and -.277, -.344 and -.391 . We are currently preparing a formal meta-analysis of these and other unpublished data, but for present purposes it suffices to say that the mean effect size found in these studies is -.375, with a median of -.318. Viewing the data conservatively, for the present paper we will use the median effect size of -.32 as typical of published studies. Using Cohen's terminology, it is between medium and small.
Although the problem of the underperformance of ethnic minorities in medical education has been much discussed, the underperformance is not unique to medical studies, Richardson in a recent analysis of HESA (Higher Education Statistics Authority) data for graduates of UK degree-giving bodies in 2005 , found a consistent under-performance of non-white students, which for the cohort aged 21–24 at graduation gave an odds ratio comparing the likelihood of NW students with W students gaining > good degrees = (i.e. I or II.i) of 0.518, equivalent to an effect size of B0.363, similar to the averaged effect reported in medicine. Similar results have also been reported by Connor et al , who have also emphasised that the participation rate of ethnic minorities in higher education is higher than for the white population, although there are differences between the various groups. Neither is underperformance of ethnic minorities restricted to university performance. Kirkup et al , in a study for the Sutton Trust, with what was probably not an entirely representative sample of UK students taking A-levels, found differences between white and non-white students, with estimated effect sizes of -0.159 for GCSEs, -0.239 for A-level grades, and -.514, -.141, -.571, -.588 and -.199 for the Critical reading, Mathematics, Writing, Writing-Multiple Choice, and Writing: essay assessments of the SAT test (giving a mean effect size of -.403 for the five SAT components); overall the seven measures had a mean effect size of -.345. The summary data from the Research Study of the Department for Education and Skills (DfES)  for pupils taking assessments in 2005 allow the calculation of effect sizes for differences between white and non-white pupils of -.215 at Key Stage 1 (age 7), -.211 at Key Stage 2 (age 11), -.178 at Key Stage 3 (age 14), and -.028 at Key Stage 4 (GCSEs) (effect sizes being calculated separately from the percentages of pupils attaining expected levels of achievement in the three separate components of each Key Stage, using the data provided in table 8 of the Research Report [, p.42], with all non-white pupils (excluding 'Unclassified') being combined together, and weighted by the Ns provided for the data at the (former) DfES website ) In pre-school children, the second sweep of the Millennium Cohort Study of a large, representative group of UK three-year olds, found lower performance in ethnic minorities on two cognitive scales, with an effect size of -.28 for the Naming Vocabulary scale of the British Ability Scales, and an effect size of -.11 for the Bracken Basic School Readiness Scale .
A range of explanations have been put forward for the underperformance of ethnic minorities in medical school, often revolving around discrimination of some form or another (e.g. Wass et al ), although such explanations have problems in explaining underperformance in computer-marked multiple-choice examinations (e.g. at the undergraduate level in  and at postgraduate level in Part 1 and Part 2 of the MRCP(UK) ). A detailed analysis of a large number of candidates taking PACES, the MRCP(UK) clinical examination, also finds that there is little association between the performance of candidates in relation to their own ethnicity and that of their examiners .
There are relatively few consistent predictors of educational outcome at university level and beyond, although one measure which does continue to make a statistical prediction, is A level examination results, the > Advanced level = examination taken by would-be university entrants in the UK, particularly in England, Wales and Northern Ireland (most applicants in Scotland taking Highers). A level results have been shown to predict performance at university, both in general [26, 27] and in medicine for both undergraduate [14, 28, 29] and post-graduate medical examinations [28, 29]. If there were differences in performance of non-white entrants at A level examinations, then that might in part explain some of the differences found in performance of ethnic minorities at medical school.
In the UK, school education is compulsory until the age of 16, when the GCSE (General Certificate of Secondary Education) exams are taken by almost all students in schools, the only exception being a small minority with severe educational, intellectual or behavioural problems (and for a detailed study of them and their characteristics see Cassen & Kingdom ). At the age of 16, students can leave education, and many choose to do so. Those who do stay on, either at school (Sixth Form), or at colleges of further education, can take a range of qualifications, of which the main ones relevant to the question of medical school applications, are the A level examinations (or Highers in Scotland), with only a few medical school applicants taking instead the International Baccalaureate or the Welsh Baccalaureate, which is currently being piloted).
Education after the age of sixteen is often called post-compulsory education, with the important implication that students choose to stay on, and also, in staying on, they choose which subjects to study and for what purposes. The element of choice in post-compulsory education complicates all statistical analyses of differences between groups, ethnic or otherwise, because one is not dealing with an entire population sample, but instead a sample that is self-selected, having chosen to be doing what it is doing, and which therefore, in statistical terms, is potentially biased. Any differences between groups may therefore result either from differences in true ability or from differences in the choices that have been made, for whatsoever reason.
In this paper we wish to look carefully at academic qualifications in relation to ethnic origin, initially in the Youth Cohort Study (YCS), a representative, population sample of 16-year olds in England and Wales, in whom performance at GCSE is well-described, as also are many aspects of performance in post-compulsory education. Our principle interest will be in factors relevant to the pool of potential medical students, and we note that more general aspects of the attainment gap in minority ethnic pupils in school have been reviewed comprehensively elsewhere [23, 31]. Because medical students are a relatively small proportion of the total population, we will also compare the analysis of the representative sample of students from YCS with a parallel analysis of a separate database of all UK applicants to university, who have applied through UCAS (Universities and Colleges Admissions Service). The data in the latter are of course far less comprehensive in relation to the population as a whole, because many of the population do not apply to university, but they allow a analysis of applicants and acceptances to a range of university subjects, and in particular to medicine. A joint analysis of the YCS data and the UCAS data allows a more complete picture to appear.
i) Youth Cohort Study (YCS)
The YCS has been carried out annually or biennially since 1985, and the current analysis is for Cohort 10, who were aged 16 or 17 when the first sweep of the survey was carried out in Spring 2000, mostly having taken GCSEs the summer before, in May 1999 [32, 33]. Further sweeps were carried out in November 2000 (Sweep 2), and March 2002 (Sweep 3), the latter being of particular importance as A level results are available. The initial sample consisted of a representative sample of 25,000 pupils from all schools in England and Wales (excluding special schools and those with fewer than 20 pupils) who were of school-leaving age during the academic year 1998–1999, and reached the age of 16 by Aug 31st 1999. Questionnaires for sweeps 2 and 3 were only sent to those replying to previous sweeps. Details of the survey are available in the survey codebook , which includes details of all three sweeps. The YCS provides weighting variables to take into account any biasses in responding in the surveys, and these have been applied in all analyses described here, so all values are representative of population values.
ii) UCAS dataset
Data were provided by UCAS for all applicants applying for admission to university in the autumn of 2003, 2004 and 2005. Information was available on the university subject for which the applicant had applied, as well as specific information on whether they had applied for medicine, and whether they had been accepted for medicine. Demographic information was also available on age, sex, ethnicity, place of residence, parental occupation, and parental education, and the educational data included the subject and grade of all A levels and AS levels taken, as well as the standard UCAS tariff score (see below).
Statistical analysis used SPSS 11.5 for most analyses, with LISREL 8.54 being used for structural equation modelling. Effect sizes for continuous variables are calculated as d = (mean W-mean NW)/SD W. Odds ratios, calculated as OR = p W × (1-p NW )/(p NW × (1-p W )), are converted to effect sizes using the method of Chinn . Negative effect sizes should be interpreted as NW students performing less well or having a lower mean than W, or NW having a lower proportion or probability of an event than W.
Definition of ethnicity
The definition of ethnicity is complex, with different studies using different criteria and classifications. It is also probable that there are differences between varying groups of ethnic minorities. However for our present purpose we are primarily interested in those studies which, for convenience and other reasons, simply compare > white = (W) students with > non-white = (NW) students. In the medical school population the majority of non-white students are from the Indian sub-continent (primarily Indian, Pakistani, Bangladeshi and other), although there are students from other groups. Supplementary table 1 [see Additional file 1] provides a complete description of all individuals using the classificatory schemes of YCS and UCAS. Although so doing inevitably simplifies a complex process, for ease of explication, we will divide students into just two groups, in order to be able to see the bigger picture more straightforwardly. We acknowledge in advance that many further, more detailed, analyses could be carried out.
The Youth Cohort Study
Performance at GCSE
GCSE achievement of W and NW students (YCS).
White Mean (SD) N = 11273
Non-White Mean (SD) N = 1776
Number of GCSEs taken
t = 2.07,
p = .039
d = -.053
Total number of GCSE points
t = 5.81,
p < .001
d = -.148
Mean GCSE points
t = 5.89,
p < .001
d = -.156
Number of GCSEs at A grade
t = 1.80,
p = .072
d = -.046
Average difficulty of GCSEs taken
t = 1.93,
p = .054
d = -.048
White N (%)
Percent with 6 or more GCSEs at grade A
1206/11272 = 10.7%
174/1602 = 9.8%
χ2 = 1.32, 1df, p = .251
OR = 0.907 (CI .77 – 1.07) d = -.054
Percent with A grade in particular subjects:
1580/10798 = 14.6%
216/1666 = 13.0%
.190 (1) p = .190
1561/9272 = 16.8%
196/1391 = 14.1%
6.62 (1) p = .010
1321/9485 = 13.9%
208/1168 = 15.1%
1.41 (1) p = .236
1219/5798 = 21.0%
151/783 = 19.3%
1.27 (1) p = .260
969/4678 = 20.7%
110/513 = 17.7%
3.17 (1) p = .075
916/3647 = 25.1%
122/369 = 24.8%
.017 (1) p = .897
799/3427 = 23.3%
123/560 = 22.0%
.494 (1) p = .482
869/4508 = 16.2%
101/828 = 12.2%
8.54 (1) p = .003
521/2508 = 20.8%
55/303 = 18.2%
1.14 (1) p = .286
251/1910 = 13.1%
50/346 = 14.5%
.43 (1) .510
1141/8125 = 14.0%
158/1182 = 13.4%
.39 (1) p = .531
Difficulty of GCSEs
Although almost all students take GCSE at age 16, there is an element of choice in the particular subjects chosen for study. There is also good evidence that not all subjects are equally difficult, some such as Chemistry, Physics, Biology and Latin being harder subjects in which to gain top grades than Art, Drama, and Sociology . We therefore calculated a measure of the average difficulty of all the GCSE subjects that had been taken by a student, based on Coe = s estimates of the Rasch difficulty for an A* grade. The choice of A* for this calculation has little impact overall, since the Rasch difficulties for different grades are, to a large extent, parallel , and A* has the advantage for present purposes that it is the grade which most medical students would be expected to achieve. Table 1 shows that there is no significant difference between W and NW students, meaning that NW students take GCSEs of similar difficulty to W students.
Demographic measures for W and NW students (YCS).
White N = 11273
Non-White N = 1776
χ2 = .089 (1) p = .765
χ2 = 25.71 (4) p < .001
III (Skilled manual)
II (Other non-manual)
χ2 = 75.72 (2) p < .001
χ2 = .015 (1) p = .093
Correlation and regression analyses
Pearson correlations between achievement at GCSE and background variables. Weighted N for cells varies between 11,544 and 13,049. Significance levels are shown as: *** p < .001; ** p < .01; * p < .05.
Total points at GCSE
Number of GCSEs taken
Difficulty of GCSEs taken
Higher socio- economic
Total points at GCSE
Number of GCSEs taken
Difficulty of GCSEs taken
Higher socio- economic group+
Greater parental education
The model was fitted by firstly using a saturated model in which all variables to the left of a variable could cause all variables to the right of the variable, through the BETA matrix in LSIREL. Paths which were non-significant at the 0.05 level were dropped sequentially from the model, with the least significant first, until all paths remaining were significant.
The final fitted model is shown in Figure 2. The overall goodness of fit was excellent, with χ2 = 8.92, 8 df, p = .349, a goodness of fit index of 1.000, and an adjusted goodness of fit of 0.999. Figure 2 shows a number of readily interpretable effects. Higher parental education results in a higher parental socio-economic group, and both factors result in a child being more likely to take part in private schooling. Private schools are more likely to put children in for more difficult GCSEs, and the children also gain more points at GCSE. Total GCSE points is particularly dependent on number of GCSEs taken, and the difficulty of the GCSEs taken, and there are also direct influences of private schooling, as well as effects of sex, parental social class, and parental education which are not explained by schooling. Finally it should be noticed that ethnicity has multiple effects, as a result of NW students having less parental education and coming from a lower socio-economic group, but after taking those into account, ethnic minorities are somewhat more likely to attend private schools.
To summarise the results so far, NW students do achieve slightly less well at GCSE, but that is largely, but not entirely, mediated by a number of background variables such as parental education and type of schooling. In particular, the numbers of NW students attaining good grades at GCSE (6 or more As) is similar to that in W students, both in a simple analysis, and also after taking background factors into account. From the point of view of understanding the potential pool of medical school applicants B those performing particularly well at GCSE, in other words B W and NW students do not show significant differences in GCSE achievement.
Choosing to take A levels
A levels are taken as a part of post-compulsory education in the UK, so that students firstly must choose to take A levels, and then they must choose which particular A level subjects to take. Medical schools typically require science subjects to have been studied, at least in part, and a majority of schools require a good grade in A level Chemistry. A levels undoubtedly differ in their relative difficulty, and Chemistry is widely perceived as more difficult than most other subjects, a perception which it would seem is correct . To enter the pool of potential medical students a student must therefore in most cases take A level sciences in general, and A level Chemistry in particular.
A-levels in W and NW students (YCS).
White Mean (SD) N
Non-White Mean (SD) N
Total number of points achieved at A level
17.71 (9.17) N = 2376
16.42 (9.15) N = 355
t = 2.474 p = .013
White N (%)
Odds Ratio (95% CI)
Taking one or more A-levels (excluding General Studies)
4666/11273 = 41.4%
738/1777 = 41.5%
χ2 = .012 (1) p = .911
OR = 1.01 (.91–1.11) d = .005
Taking one or more science A-levels
2701/11273 = 24.0%
515/1777 = 29.0%
χ2 = 20.84 (1) p < .001
1.30 (1.16–1.45) d = .145
Taking A-level Chemistry
744/11273 = 6.6%
202/1776 = 11.4%
χ2 = 51.98 (1) p < .001
1.82 (1.54–2.14) d = .331
Regression analysis and path analysis
Although overall NW students have poorer achievement at GCSE, a previous analysis here had suggested that NW students are equally likely as W students to achieve very high GCSE grades (6 or more A grades or better). The logistic regressions described in the previous section were therefore repeated only for students with 6 or more A grades at GCSE. Overall there was now no effect of ethnicity on taking one or more A-levels, although that largely reflects the fact, seen at the right hand end of Figure 3, that 97% of such a high achieving group go on to take A levels. However, a similar analysis for the taking of science at A level and the taking of Chemistry showed that even among GCSE high-achievers, NW students were 2.85 times more likely to take a science A level (p < .001; effect size = .579), and 2.45 times more likely to take Chemistry A level (p < .001; effect size = .495).
A level difficulty
As with GCSEs, not all A level subjects are equally difficult. We therefore used the data of the CEM Centre at the University of Durham  to calculate a difficulty score for the subjects being taken by each student at A level. Multiple regression of A level difficulty on the six background variables and two measures of GCSE attainment shown in Figure 2 found that more difficult A levels were taken by students with more GCSE points (p < .001), but who had taken more difficult (P < .001) but somewhat fewer GCSEs (P < .001), were female (p < .001), and were non-white (p < .001). However, since Chemistry is one of the most difficult of A levels, the latter effect is perhaps not unexpected. Repeating the analysis but including a variable indicating whether or not students were studying chemistry at A level meant that the effect of ethnicity was no longer significant (p = .207).
In summary, despite having somewhat poorer overall attainment at GCSE than W students, NW students are more likely to go on to take A levels, and in particular science A levels and Chemistry A level, which tend to be more difficult. The same was also true of science A levels and Chemistry A level amongst the high achieving students with 6 or more A grades at GCSE.
Attainment at A level
YCS was carried out in several sweeps, with information on GCSE results and intentions to take A level obtained at sweep 1. Information on actual attainment at A level was obtained only at sweep 3, and inevitably the response rate was much poorer. Of the original 25,000 individuals in the sampling frame, 13,699 responded to sweep 1, and only these individuals were sent sweep 2, of whom 10,100 responded. Only these 10,100 individuals were sent sweep 3, of whom 7,971 responded to sweep 3 (and of course many of these students did not take A levels). Nevertheless, because the sampling frame is well characterised, YCS provides appropriate weighting factors to take response biases into account, and the weighting has been applied in the following analyses, so that estimates are likely to be applicable to the population as a whole. Information at sweep 3 is less comprehensive, and little apart from examination achievement is directly relevant to the study of medicine as such.
A level grades in relation to GCSE grades
Actual taking of A levels
Of the 13,049 individuals in sweep 1 of YCS, only 6692 (51.3%) responded on sweep 3 when A level results were asked for, and of these only 2731 had actually taken one or more A levels. (Note that these figures are slightly different from those cited at the beginning of the section on A level achievement as those figures are raw counts of number of respondents, whereas the figures quoted here, and subsequently, refer to weighted responses). Of the 3478 individuals at sweep 1 who said they did not intend to take A levels, only 74 (2.1%) actually did, with no difference in proportion between W and NW (2.1% and 2.4% respectively). However of the 3214 individuals at sweep 1 who said they intended to take A levels, 2657 (82.7%) actually did, and the proportion was higher in W students (2314/2780 = 83.2%) than in NW students (342/433 = 79.0%), a significant difference (chi-squared = 4.73, 1df, p = .030; odds ratio = 0.760, effect size = B.152), suggesting that more NW than W students had decided their aspirations were inappropriate or impractical, although the data cannot decide between those possibilities. Those not actually taking A levels who had said they would, had significantly lower GCSE point scores (mean = 29.1, SD = 10.24, N = 557) than those who did take A levels as intended (mean = 48.56, SD = 8.68, N = 2657, t = 22.76, p < .001). Similarly, although fewer NW students actually took A levels than said they would, the NW students who actually took A levels had significantly lower GCSE points than the W students (W: mean = 38.62, SD 8.63, N = 2376; NW: mean = 36.32, SD = 10.39, N = 355; t = 4.56, p < .001).
A level achievement
The UCAS dataset on applicants and entrants to universities
The YCS is a useful database for looking at the entire population of individuals taking GCSEs and then A levels. However, the numbers involved are too small to be useful for looking at applicants and entrants for particular university subjects. The UCAS database is extremely large, the three years from 2003–2005 having a total of 1,484,650 applicants, 476,467 in 2003, 486,028 in 2004 and 522,155 in 2005, of whom 52,557 applied for medicine, and 23,443 were accepted at medical school. However the database is weak in terms of the richness of the background variables, having only A levels, and not GCSEs, for instance, and having only a limited number of demographic variables. And of course, of necessity, it can only look at those individuals who have applied to university. Nevertheless, the UCAS data complements and supports the YCS data.
A level scores
Although most of UCAS = published analyses are in terms of its > tariff score = , the tariff score has several disadvantages when thinking primarily about medical school applicants. The tariff combines different qualifications which are on different metrics, but that has the disadvantage that the origin of any particular tariff is far from clear. For instance, a candidate gaining B grades at three A levels will receive 3 @ 100 = 300 points, as also will a candidate gaining two As and a D (120+120+60), as also will a candidate taking six AS levels and gaining Bs in all of them (6 @ 50). The tariff also considers all such candidates equally, irrespective of the level or number of qualifications or the subjects in which they were taken. In particular, the tariff score includes General Studies A level on an equal footing with all other A levels, whereas many medical schools do not regard it as a full A level. For the present study we have therefore recalculated attainment scores de novo, and have considered only A levels, scoring them in the same way as for the YCS, with 10 = A, 8 = B, etc.. For most medical schools it is A levels that matter, and not combinations of A and AS levels. UCAS was unable to provide data on Scottish Highers, and therefore our analysis is restricted to those taking A levels, being typically candidates outside of Scotland.
For convenience and ease of interpretation we have confined our analyses to applicants who were aged under 21 at the time of application (making them broadly equivalent to the YCS data), and who were resident in the UK, giving a total sample of 976,007. Some applicants had not taken A levels, for various reasons, in particular being resident in Scotland or Wales, and taking Highers or Baccalaureate, and we therefore considered only the 531,333 candidates who had results for at least three A level subjects (excluding General Studies). A very small number of candidates, 2642, had five or more A levels, and for practical reasons we excluded them also, leaving 528,691 applicants, 473,781 with three A levels and 54,910 with four A levels. For simplicity, for candidates with four A levels we have considered only the grades from their three best results, meaning that for all candidates the maximum number of points is 30, which is equivalent to AAA. Although this might seem to devalue the additional qualification obtained by those with four A levels, it is worth noting that while 10.5% of the applicants with three A levels had attained the maximum score of 30 points, that score was attained by 48.9% of the applicants with four A levels, showing that those taking four A levels tended to be amongst the very highest achieving of all applicants.
A level results of UCAS applicants in relation to the YCS population
A level results in UCAS applicants in relation to ethnicity
Comparison of A-level achievement in White and non-White applicants in the UCAS data, for all applicants, applicants taking Chemistry, applicants applying to medical school, and applicants accepted for medical school.
White Mean A-level scores (SD)
Non-White Mean A-level scores (SD)
21.92 (5.93) N = 443,038
21.09 (6.31) N = 74,262
t = 34.397 p < .001
d = -.140
All applicants taking Chemistry
23.83 (6.03) N = 72,925
22.92 (6.35) N = 24,626
t = 20.26, p < .001
d = -.151
All medical school applicants
27.29 (3.56) N = 15,073
26.29 (4.41) N = 8,174
t = 18.81 p < .001
d = -.282
All medical school entrants
28.77 (1.97) N = 9,945
28.54 (2.38) N = 4,373
t = 5.89 p < .001
d = -.114
White N (%)
All applicants: percent gaining ABB or higher
158,373/443,038 = 35.7%
24,074/74,262 = 32.4%
χ2 = 308.8 (1) p < .001
OR = .862 (CI .848 – .877) d = -.082
All applicants: proportion taking Chemistry
72,925/443,038 = 16.5%
χ2 = 11,593 (1) p < .001
OR = 2.52 (CI 2.48 – 2.56) d = .-511
Medical school applicants: percent gaining ABB or higher
11,977/15,074 = 79.5%
5815/8174 = 71.1%
χ2 = 204.3 (1) p < .001
OR = 0.634 (CI .60 – .68) d = -.252
Medical school applicants: percent gaining AAA higher
6,893/15,074 = 45.7%
3098/8174 = 37.9%
χ2 = 132.6 (1) p < .001
OR = .724 (CI .69–.77) d = -.178
Medical school acceptances: percent gaining ABB or higher
9509/9945 = 95.6%
4109/4373 = 94.0%
χ2 = 17.9 (1) p < .001
OR = .714 (CI .61 – .84) d = -.186
Medical school acceptances: percent gaining AAA higher
6083/9945 = 61.2%
2586/4373 = 59.1%
χ2 = 5.24 (1) p = .022
OR = .919 (CI .854–.988) d = -.047
Applicants taking A level chemistry
Applicants to medical school
Entrants to medical school
Exploring the censored distribution of A level attainment in W and NW students
The distribution of A level attainment in university applicants in general, shown in Figure 7, is an important one that is capable of further analysis. It can be seen that to a first approximation the distribution is normal, with a mode at about 20 points (i.e equivalent to about BCC) but with a wide distribution around that, and, in particular a clear second mode at the maximum of 30 points (AAA). It should also be noticed that the distribution tails away towards the left-hand end of the distribution, the very lowest points possible with three A levels being 6 (i.e. EEE).
In interpreting this distribution, a clear distinction should be made between distributions that statisticians describe as censored and those described as truncated. In a clinical survival analysis, patients may be followed up for, say, five years. During that time, some will have died, and their precise survival will be known. Others, however, will have survived for the whole of the five years but will, inevitably, succumb at some future time, but all that can be said of them is that their survival is at least five years; these data are censored, the number of observations being known but their precise value being unknown. In contrast, a study may look only at a group of patients whose blood pressure is known to be at least 160 mmHg, and the distribution of blood pressure for those individuals is truncated, no individuals being included below the threshold level. The key difference is that for the censored data, all of those surviving for beyond the measured time are included in the final bin of the histogram, whereas in the case of truncation, those not included are not shown anywhere in the histogram.
The A level data of Figures 7 to 11 are truncated at the lower bound, as individuals are only included who have gained a minimum of 6 points from 3 A levels. In contrast, at the upper end of the distribution, individuals who could have gained more than 30 points (AAA) if the marking scheme had allowed it, by, for instance, including hypothetical grades of A*, A** and A***, gaining 12, 14 and 16 points respectively, are forced into the highest bin of the histogram, AAA, because they cannot achieve any higher result given the marking scheme. The A level distribution is therefore censored at the top end (and that accounts for the clear second mode in Figures 7 to 11 at 30 points).
Parameter estimates for the latent (uncensored) distributions in all applicants, and in medical school applicants and acceptances, separately for W and NW candidates.
Medical school applicants and acceptances
Estimated true mean in applicants
Estimated true SD in applicants
Selection process: Slope
Selection process: Intercept
Estimated true mean in acceptances
Estimated true SD in acceptances
The effect size of -.129 for the A level performance of all NW applicants is similar both to the simple, unadjusted, effect size of -.141 found for applicants in the UCAS data (Table 5), and the effect size of -.140 found for A level points in the YCS data (Table 4). In contrast the true effect size of -0.592 for medical school applicants shown in Table 6 is considerably larger than the effect size calculated from the raw data that is shown in Table 5 of -.282, showing that the censoring of the data has distorted the true picture. Similarly, the raw effect size of -.114 for medical school acceptances in Table 5 is larger than the adjusted effect size of -.068 in Table 6. Censoring of the data by the artificial ceiling of AAA in A levels therefore distorts the estimates of effect size.
This paper has analysed two very different databases, in order to answer questions about the academic background and qualifications of medical students from ethnic minorities. The questions it has asked have, inevitably, been simplified, particularly in so far as the studies have looked only at A level grades, in those students who have taken three or four A levels, and they have looked only at a division of the population into just two groups, White and non-White. No doubt far more subtle analyses could carried out, but those presented here are sufficient for a basic analysis of the issue.
The starting point for the paper is that in a range of studies, medical students and doctors from ethnic minorities have underperformed in different types of examination, both undergraduate and postgraduate. Although it is inevitably tempting to attribute underperformance in some examinations, such as clinical assessments, to factors such as direct or indirect discrimination on the part of examiners, perhaps because of cultural insensitivity or other reasons, all such explanations have problems in explaining why candidates from ethnic minorities underperform in assessments of knowledge which use multiple-choice questions and are marked by a computer.
A crucial but implicit assumption of many studies of underperformance by students from ethnic minorities is that W and NW students are initially equivalent in their academic ability. Since selection has taken place on the basis of an academic criterion, then it might seem reasonable to infer that equality of intellectual ability is present in students from different groups. A particularly clear example comes from Esmail, commenting on the original findings in Manchester in 1995 that a higher proportion of NW students were failing their final examinations.
AWe found a significant statistical association between men with Asian names and failing clinical exams. ... These were not substandard students. They were accepted to Manchester University on the basis of the same A level criteria as everyone else.@  (our emphasis)
It is a seductive and powerful argument, and one that in some form or another, many of us have used. However, the data in the present analysis suggest it is not correct, and that finding has a number of important implications, particularly in an area where research is quoted in the context of Aracial discrimination≅  and Aracism in medicine≅ . Despite being selected Aon the basis of the same A level criteria as everyone else≅, there are several mechanisms by which NW students may still underperform at medical school.
i) Different distributions of educational achievement in the W and NW populations
The YCS data are important because they show the achievement of a large, representative population sample on an examination, the GCSE, which is taken by almost all students in the UK population at the age of 16, and for which teachers and schools are rewarded for their students performing as well as possible. There is therefore every reason to believe it is taken very seriously. Figure 1 shows that NW students have both a lower mean attainment and a somewhat greater variance of attainment at GCSE than do W students. The reasons for that are complex, and as Figure 2 shows, it is in large part a reflection of NW students tending to come from lower socio-economic groups, and their parents being less likely to have progressed into post-compulsory education at A level or degree standard, both factors influencing GCSE achievement, factors which are probably still important in the cohort of UK children born in 2000 . Noteworthy, though, in Figure 2, is that ethnicity, despite its clear effect in a simple analysis, has little effect after other variables are taken into account, its effects mainly being mediated via parental variables. It is also important that ethnic minority students are equally likely to attain 6 or more grade As at GCSE, suggesting that performance at the top end of the distribution, at a level which is probably needed to achieve medical school admission, is equivalent in white and non-white students. The overall difference in mean GCSE performance, coupled with a wider variance in non-white students, suggests that differences in socio-economic and parental educational factors are mainly responsible for non-white students underperforming. As more non-white students enter higher education and subsequently have children, that difference is likely to diminish.
The YCS data also show that there is a very clear relationship between performance at GCSE and subsequent performance at A level (Figure 5), and that the structural form of the relationship is very similar in W and NW students, suggesting that GCSEs meet the Cleary test for equivalent predictive validity  B although see Chung-Yan and Cronshaw for criticisms of the Cleary test . Good GCSE results therefore should predict good A level results in a similar way in both W and NW students. Although the simple correlation between GCSEs and A levels is strong, with a correlation of about 0.63, it is inevitably an underestimate of the true correlation, since measurement error of the two examinations has not been taken into account, and neither has the effect of restriction of range (A levels only being taken by those with better GCSE results). The true structural (disattenuated) correlation is likely to be of the order of 0.7 to 0.8, or so.
Because NW students do less well at GCSE, it is therefore to be expected that on average they are also likely to do less well at A level, solely on that basis.
ii) Choice of GCSEs, choice of A levels and inappropriate aspirations
Although the taking of GCSEs is, in effect, compulsory, the particular subjects chosen to be taken, as well as their number, is in part a matter of choice in a decision made jointly by students, their schools, and their parents. Some subjects are undoubtedly easier than others , and some students choose, for whatever reason, to take harder subjects, a decision, as shown in Figure 2, which is influenced by parental education and occupation, and by schools. At GCSE, there is evidence that students choosing to take harder subjects in fact perform better, probably because these students are already higher achievers and schools, parents, or they themselves, wish them to be stretched.
Similar considerations apply at A level, although the choice of A level subjects, and indeed the decision to take A levels at all, are both made in the light of GCSE grades already attained. Of particular importance are the results shown in Figures 3 and 4. It is clear that, for any particular level of GCSE attainment, NW students are much more likely to take A levels in general, and science and chemistry A levels in particular. Since GCSEs are seen to be powerful predictors of A level attainment, the implication is that NW students will, on average underperform yet further at A level than would be predicted on the basis of their relatively poorer overall performance at GCSE. That can be seen in the YCS data, where the mean A level attainment of NW students is significantly less than that in W students. Inappropriate aspirations are partly involved, as seen in the significantly higher proportion of NW students who intend to take A levels but in fact do not do so, those non-takers also having lower GCSE attainment than those who do go on to take A levels. Connor et al  have noted that NW students were more likely to cite parental influence in their educational choices, and that Asian parents aspire to their children having Aprofessional≅ careers, such as in medicine or law, and that this may in part explain the increased proportion of Asians studying medicine.
iii). Poorer NW A level attainment in the pool of UCAS applicants
The majority of applicants with reasonable A levels currently apply to university (as can be inferred from Figure 7), and there are other similarities between the YCS and UCAS data, which suggest that the data sets are equivalent in many ways. In particular, it is important to note that in the YCS data, and using measures of effect size, the NW students performed 0.141 SDs below the W students, and in the UCAS data the NW students performed 0.140 SDs below the white students (see Figure 7), showing the datasets are very similar in the effects they are finding, with the advantage that the UCAS dataset is extremely large, and has information on the small minority of students who apply to and are accepted by medical school.
iv). The A level achievement of applicants and acceptances at medical school
Figures 9 and 10 make very clear that medical school applicants, and particularly medical school acceptances, are a very highly selected academic elite from amongst the entire pool of those applying to university, the modal A level attainment being at the maximum possible score of three As at A level. Despite that, it is also clear, a) that medical acceptances have higher attainment than those not accepted, and b) that both in applicants and in acceptances, the NW candidates have significantly lower levels of achievement than the W candidates. The mechanisms for that in applicants cannot be explored directly, but it is likely that differential aspirations are involved. Likewise, the reasons for the difference in those accepted cannot be analysed further, but it might, for instance, be argued that extenuating circumstances are more likely to be important in NW students than W students, and hence they are admitted with somewhat lower achievement than W students. However, that explanation is not compatible with the clear finding, in the selection functions shown in Figure 13b, that selection for non-white applicants is stricter than for white applicants, the 50% point of the logistic function being at 26 points for W applicants and 30 points for NW applicants. Similar differences have been reported previously, in a study of selection for UK medical schools in 1996 and 1997 , the likelihood of a medical school offer being lower for NW applicants at all levels of A level achievement.
It is harder for NW applicants to enter medical school, even when they achieve equivalent A level grades to W applicants, although the net result of selection is that the A level distributions of White and non-white entrants are much more similar than in applicants. Nevertheless, W acceptances still have significantly higher A level grades than do NW, although the effect size is relatively small (but see below for further discussion).
v) Selection from groups of different mean levels of ability
The mean A level achievement of NW university applicants and entrants is lower than that of W students, for whatever reason (see Figures 7, 8, 9, 10, 11, 12, 13). That fact alone means that in a fair selection system in which all individuals of whatever group who are over some critical threshold are accepted, the mean achievement of those selected from the group with the lower average performance, will also be significantly lower than in the higher achieving group. As an example, consider the entire distribution of candidates in Figure 9 who have taken A level Chemistry and hence can be considered as in the > real pool = of potential medical students. Simply select all of those who have 26 or more points at A level i.e. grades of ABB or better. Altogether, 25.0% of those taking A level chemistry are NW (compared with the 14.4% of UCAS applicants overall who are NW). However, only 23.0% of those meeting the 26 point criterion are NW, compared with 27.0% of those in the W group, an odds ratio of 1.235 (equivalent to an effect size of -.117). More crucially, amongst those selected on the 26-point criterion, and despite the extreme narrowness of the range of A levels selected (only 26, 28 or 30 points), the mean score of the W students (28.67, SD = 1.61) is higher than the mean score of the NW students (28.53, SD = 1.64), a highly significant difference (t = 8.390, p < .001, effect size = -0.090). Of particular interest is that 54.6% of the 37752 W students had the maximum 30 points, whereas only 50.1% of the 11294 NW students had 30 points (p < .001; odds ratio = 1.198, effect size = -0.100). To some extent that difference in entry qualifications is likely to perpetuate itself as students continue through their education, particularly given that A levels have been shown to be continuing predictors of undergraduate and postgraduate medical school performance [28, 29].
Throughout this paper we have presented differences in attainment of white and non-white students in terms of standard effect sizes, with d being the difference between the groups in terms of standard deviations of the white students (the larger group). Perhaps the most interesting result concerns acceptances at medical school, where, in Tables 5 and 6, effect sizes have been calculated in four different ways, giving values of B.186, B.114, B.068 and B.047, with a mean of -.10, or one tenth of a standard deviation, which can be generally regarded as a small effect (although it is much larger than some of the very small effects found in therapeutically important clinical trials of new drugs ). The effect is a little smaller, albeit broadly similar, to the overall effects found at GCSE (B.151) and A level (B.141) in the YCS, and at A level in the UCAS data for all applicants (B.140), for applicants taking chemistry (B.151), although it is substantially lower than in medical school applicants (B.282). There seems therefore no doubt that the under-attainment of NW students applying to university, be it for medicine or in general, is part of a broader pattern of under-attainment throughout pre-school education, compulsory and post-compulsory education, perhaps exaggerated at A level by greater educational aspirations in NW students. However, and it is an important > however =, the effect found in medical school acceptances of about B.10 is substantially smaller than the effect found in undergraduate and postgraduate medical examinations of about B.32. It seems unlikely, therefore, that the earlier underachievement of NW students at GCSE and A level alone can entirely explain the poorer performance of NW students at doctors at medical school and beyond. Nevertheless, the difference is still likely to explain some of the effects found. If NW students come from a normal distribution with an effect size of -.10 relative to the W students, then it can be expected that amongst those performing 1 SD below the white mean there will be 17% more NW students, and amongst those 2SDs below the white mean, there will be 27% more NW students (so that if about 5% of W students were failing, then about 6.2% of NW students would be likely to fail, a 23% excess, with about 5 NW failures for every 4 W failures).
A levels and selection
There seems little doubt that A levels are still the major component of selection in most medical schools, with typical requirements being AAA, AAB or occasionally ABB . Most schools however also say that suitable personality, motivation, experience and other factors are also necessary. Although A levels are clearly important, it can be seen in Figure 13 that success is not guaranteed, even for white applicants with grades of AAA (of whom only 88% were accepted in the UCAS data). Although it is tempting to believe that this reflects medical schools using other criteria for selection, that conclusion is far from secure. In particular it should be remembered that most applicants to medical school are applying pre-A level, and hence offers and rejections are made not on A levels attained, but on GCSEs attained, and teachers = estimates of A levels that are likely to be gained. If a medical school believes, for whatever reason, that a candidate will not attain their likely offer (say, AAB), then they may well reject a candidate, particularly since it is a particular problem for a school to have more rather than less candidates achieving their offers, schools being legally bound to accept all applicants who meet their entry requirements but also having strict government quotas on the numbers of medical students who can be accepted. Subsequently such candidates may well achieve grades of AAA but in the absence of an offer they cannot at that time gain a place (although they may gain one in the next year). The result, even in the absence of selection on the basis of personality, or whatever, would be that not all AAA candidates would be accepted. The UCAS data do not allow these possibilities to be distinguished.
The implications of differences in educational attainment of ethnic minority students
Although NW students have been selected on similar criteria to W students, they not only have lower educational attainment than W students, but the population from which they are drawn, the pool of all NW medical school applicants, has lower educational attainment than those in the pool of all W medical school applicants. As a result, and using the estimated parameters of the censored normal distributions provided in table 6, while the mean score of W acceptances is 0.43 standard deviations from the mean of W medical school applicants, the mean score of W acceptances is 0.88 standard deviations above the mean of NW medical school applicants. In such a situation there is likely to be differential regression to the mean, with NW applicants performing less well than W applicants.
Other factors contributing to underperformance of NW students and doctors
Although this study has identified a small but real effect of educational background which is likely to mean that some NW medical students and doctors underperform in examinations, it clearly cannot account for the entire effect found in earlier studies. What other factors also account for the underperformance is still very unclear. NW students may have somewhat different motivations , but it is not clear as yet if these differences predict outcome. Differences in learning style and a range of other factors have been looked at previously , and typically any differences do not account for differences in W and NW examination performance. Further studies are needed of this phenomenon, which provides a serious challenge to the discipline of medical education.
Non-white medical students have somewhat lower A level grades than white medical students, the effect size being about -0.10 standard deviations, which is similar, if a little smaller, than the relative underperformance of non-white students at GCSE and A level. Although the effect is reliable, it is substantially smaller than the relatively large effect size found in undergraduate and postgraduate medical education, of about -0.32 standard deviations, although it probably contributes something to that effect.
List of abbreviations used
General Certificate of Secondary Education
Universities and Colleges Admissions Service
Youth Cohort Study
We thank UCAS for providing data for secondary analysis.
- Bedi R, Gilthorpe S: Ethnic and gender variations in university applicants to United Kingdom medical and dental schools. British Dental Journal. 2000, 189: 212-215. 10.1038/sj.bdj.4800725a.Google Scholar
- McManus IC, Richards P: An audit of admission to medical school: 1. Acceptances and rejects. British Medical Journal. 1984, 289: 1201-1204.View ArticleGoogle Scholar
- McManus IC, Richards P: Admission to medical school [letter]. British Medical Journal. 1985, 290: 319-320.View ArticleGoogle Scholar
- McManus IC, Richards P, Maitlis SL: Prospective study of the disadvantage of people from ethnic minority groups applying to medical schools in the United Kingdom. British Medical Journal. 1989, 298: 723-726.View ArticleGoogle Scholar
- McManus IC, Richards P, Winder BC, Sproston KA, Styles V: Medical school applicants from ethnic minorities: identifying if and when they are disadvantaged. British Medical Journal. 1995, 310: 496-500.View ArticleGoogle Scholar
- UCAS: Statistical services. Accessed 4th April; 2008., [http://www.ucas.com/he_staff/stat_services1/]
- Dillner L: Manchester tackles failure rate of Asian students. British Medical Journal. 1995, 310: 209.View ArticleGoogle Scholar
- Esmail A, Dewart P: Failure of Asian students in clinical examinations: The Manchester experience. Race and Higher Educcation: Experiences, challenges and policy implications. Edited by: Modood T, Acland T. 1995, London: Policy Studies Institute, 134-143.Google Scholar
- McManus IC, Richards P, Winder BC, Sproston KA: Final examination performance of students from ethnic minorities. Medical Education. 1996, 30: 195-200.View ArticleGoogle Scholar
- Haq I, Higham J, Morris R, Dacre J: Effect of ethnicity and gender on performance in undergraduate medical examinations. Med Educ. 2005, 39: 1126-1128. 10.1111/j.1365-2929.2005.02319.x.View ArticleGoogle Scholar
- Woolf K, Haq I, McManus IC, Higham J, Dacre J: Exploring the underperformance of male and minority ethnic medical students in first year clinical examinations. Advances in Health Sciences Education. 2007, [http://www.springerlink.com.libproxy.ucl.ac.uk/content/yh5r716074g54265/fulltext.pdf]Google Scholar
- Wass V, Roberts C, Hoogenboom R, Jones R, van der Vleuten C: Effect of ethnicity on performance in a final objective structured clinical examination: qualitative and quantitative study. British Medical Journal. 2003, 326: 800-803. 10.1136/bmj.326.7393.800.View ArticleGoogle Scholar
- Yates J, James D: Predicting the 'strugglers': A case-control study of students at Nottingham Universityt Medical School. British Medical Journal. 2006, 332: 1009-1013. 10.1136/bmj.38730.678310.63.View ArticleGoogle Scholar
- Lumb AB, Vail A: Comparison of academic, application form and social factors in predicting early performance on the medical course. Medical Education. 2004, 38: 1002-1005. 10.1111/j.1365-2929.2004.01912.x.View ArticleGoogle Scholar
- Dewhurst NG, McManus IC, Mollon J, Dacre JE, Vale JA: Performance in the MRCP(UK) Examination 2003-4: Analysis of pass rates of UK graduates in the Clinical Examination in relation to self-reported ethnicity and gender. BMC Medicine. 2007, 5 (8): [http://www.biomedcentral.com/1741-7015/5/8/abstract-doi:10.1186/1741-7015-5-8]Google Scholar
- Bessant R, Bessant D, Chesser A, Coakley G: An analysis of predictors of success in the MRCP(UK) PACES examination in candidates attending a revision course. Postgraduate Medical Journal. 2006, 82: 145-149. 10.1136/pmj.2005.035998.View ArticleGoogle Scholar
- Wakeford R, Farooqi A, Rashid A, Southgate L: Does the MRCGP examination discriminate against Asian doctors?. British Medical Journal. 1992, 305: 92-94.View ArticleGoogle Scholar
- Wakeford R: MRCGP: 2006 Annual report. 2007, London: Royal College of General Practitioners, [http://www.rcgp.org.uk/PDF/Annual%20Report%20for%202006%2029%20July%202007.pdf]Google Scholar
- Cohen J: A power primer. Psychological Bulletin. 1992, 112: 155-159. 10.1037/0033-2909.112.1.155.View ArticleGoogle Scholar
- Richardson JTE: The attainment of ethnic minority students in UK higher education. Studies in Higher Education. 2008, 33: 33-48. 10.1080/03075070701794783.View ArticleGoogle Scholar
- Connor H, Tyers C, Modood T, Hillage J: Why the difference? A closer look at higher education minority ethnic students and graduates. 2004, Department for Education and SkillsGoogle Scholar
- Kirkup C, Schagen I, Wheater R, Morrison J, Whetton C: Use of an aptitude test in university entrance – A validity study: Relationships between SAT scores, attainment measures and background variables. 2007, London: Department for Education and Skills, Research Report RR846Google Scholar
- Department for Education and Skills: Ethnicity and education: The evidence on minority ethnic pupils aged 5-16. 2006, London: Department for Education and Skills, [http://www.dfes.gov.uk/rsgateway/DB/RRP/u014955/index.shtml]Google Scholar
- Department for Innovation UaS: National Curriculum Assessment, GCSE and Equivalent Attainment and Post-16 Attainment by Pupil Characteristics, in England 2005. Accessed 18th March 2008., [http://www.dfes.gov.uk/rsgateway/DB/SFR/s000640/]
- Hansen K, Joshi H: Millennium Cohort Study: Second Survey: A user's guide to initial findings. 2007, London: Centre for Longitudinal Studies, Institute of EducationGoogle Scholar
- Choppin BHL, Orr L, Kurle SDM, Fara P, James G: The prediction of academic success. 1973, Slough: NFER Publishing Company LtdGoogle Scholar
- Anonymous: Schooling effects on higher educationachievement. 2003, [http://www.hefce.ac.uk/pubs/hefce/2003/03_32.htm]Google Scholar
- McManus IC, Smithers E, Partridge P, Keeling A, Fleming PR: A levels and intelligence as predictors of medical careers in UK doctors: 20 year prospective study. British Medical Journal. 2003, 327: 139-142. 10.1136/bmj.327.7407.139.View ArticleGoogle Scholar
- McManus IC, Powis DA, Wakeford R, Ferguson E, James D, Richards P: Intellectual aptitude tests and A levels for selecting UK school leaver entrants for medical school. British Medical Journal. 2005, 331: 555-559. 10.1136/bmj.331.7516.555.View ArticleGoogle Scholar
- Cassen R, Kingdon G: Tackling low educational achievement. 2007, York: Joseph Rowntree FoundationGoogle Scholar
- Department for Education and Skills: Ethnicity and education: The evidence on minority ethnic pupils (Research Topic Paper: RTP01-05). 2005, London: Department for Education and Skills, [http://www.dfes.gov.uk/research/programmeofresearch/projectinformation.cfm?projectid=14488&resultspage=1.]Google Scholar
- Anonymous: Youth Cohort Study: Activities and Experiences of 16 Year Olds: England and Wales 2002. 2003, [http://www.dfes.gov.uk/rsgateway/DB/SFR/s000382/index.shtml]Google Scholar
- Russell N, Phelps A: Youth Cohort Study: Cohort 10 Sweep 1 (C10S1). Technical Report. 2001, [http://www.data-archive.ac.uk/doc/4571%5Cmrdoc%5Cpdf%5C4571userguide.pdf]Google Scholar
- Chinn S: A simple method for converting an odds ratio to effect size for use in meta-analysis. Statistics in Medicine. 2000, 19: 3127-3131. 10.1002/1097-0258(20001130)19:22<3127::AID-SIM784>3.0.CO;2-M.View ArticleGoogle Scholar
- Coe R: Relative difficulties of examinations at GCSE: an application of the Rasch model: Working Paper. 2006, [http://www.cemcentre.org/Documents/News/subjectdifficultiesbyrasch.pdf]Google Scholar
- Anonymous: A-Level Subject Difficulties. 2006, [http://www.cemcentre.org/renderpage.asp?linkid=11625001]Google Scholar
- UCAS: The UCAS tariff. 2006, [http://www.ucas.com/students/ucas_tariff/]Google Scholar
- McManus IC, Woolf K, Dacre JE: Even one star at A level could be "too little, too late" for medical student selection. BMC Medical Education. 2008, 8: 16-10.1186/1472-6920-8-16.View ArticleGoogle Scholar
- McKenzie KJ: Racial discrimination in medicine [Editorial]. British Medical Journal. 1995, 310: 478-479.View ArticleGoogle Scholar
- Coker N: Racism in medicine: An agenda for change. 2001, London: King's FundGoogle Scholar
- Hansen K, Joshi H, [Editors]: Millennium Cohort Study, Second Survey: A user's guide to initial findings. 2007, London: Centre for Longitudinal Studies, Institute of Education, [http://www.cls.ioe.ac.uk/studies.asp?section=0001000200010004]Google Scholar
- Cleary TA: Test bias: Prediction of grades of Negro and white students in integrated colleges. Journal of Educational Measurement. 2008, 5: 115-124. 10.1111/j.1745-3984.1968.tb00613.x.View ArticleGoogle Scholar
- Chung-Yan GA, Cronshaw SF: A critical re-examination and analysis of cognitive ability tests using the Thorndike model of fairness. Journal of Occupational and Organizational Psychology. 2008, 75: 489-509. 10.1348/096317902321119709.View ArticleGoogle Scholar
- McManus IC: Factors affecting likelihood of applicants being offered a place in medical schools in the United Kingdom in 1996 and 1997: retrospective study. British Medical Journal. 1998, 317: 1111-1116.View ArticleGoogle Scholar
- Rosenthal R: How are we doing in soft psychology?. American Psychologist. 1990, 45: 775-777. 10.1037/0003-066X.45.6.775.View ArticleGoogle Scholar
- Parry J, Mathers J, Stevens A, Parson A, Lilford R, Spurgeon P, Thomas H: Admissions processes for five year medical course at English schools: review. British Medical Journal. 2006, 332: 1005-1009. 10.1136/bmj.38768.590174.55.View ArticleGoogle Scholar
- McManus IC, Livingston G, Katona C: The attractions of medicine: the generic motivations of medical school applicants in relation to demography, personality and achievement. BMC Medical Education. 2006, 6: 11-10.1186/1472-6920-6-11.View ArticleGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6920/8/21/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.