Effect of A-level Subject Choice and Entry Tariff on Final Degree and Level 1 Performance in Biosciences

Abstract

Following the publication of the higher education white paper increasing entry tariff and widening participation have become even more important issues for universities. This report examines the relationship between entry tariff and undergraduate achievement in Biosciences at the University of Exeter. We show that, whilst there is a significant correlation between A-level tariff and both level 1 and final degree marks, the magnitude of the correlation decreases with increasing A-level tariff. It was also found that, contrary to anecdotal evidence, there is no link between A-level maths or chemistry and degree success in biosciences. However gender was found to have a marked effect on degree success independent of A-level tariff. Female students consistently obtained marks more than one third of a degree classification higher than male students with equivalent entry qualifications. These results are compared to investigations in other disciplines, notably medicine, and in the context of the drive to increase entry tariff to allow unrestricted recruitment of the top students.

Keywords:

• A-level tariff
• bioscience
• gender
• degree success
• level 1 marks

Introduction

In the recent white paper on higher education (Department for Business Innovation and Skills, 2011) the coalition government outlined proposals for uncapped recruitment of the highest achieving students, those with the at least AAB at A-level, or the equivalent. The aim was to increase competition for the most selective courses, institutions and most able students and to enable more of the top students to go to their first choice of institution. The proportion of students achieving ‘good honours’ degrees (1^st and upper seconds) is also included in league tables, but is allowing more high achieving 16–19 years olds into institutions a guarantee of more high achieving graduates? There have been many accusations that pupils are being ‘spoon fed’ at school (Woolcock, 2009) and entering university unable to think for themselves and with a poorer grasp of their subject as a whole; rather their understanding is in ‘bite-sized’ chunks for examination (Bassett et al., 2009). So is A-level achievement a good indicator of degree success?

The 2003 report by the Higher Education Funding Council for England (HEFCE), Schooling effects on higher education achievement, still identifies A-level grades as the most important factor in determining HE achievement, measured by completion rates and class of degree. The report showed that for all types of school, ranked according to performance, the proportion of students receiving a 2:1 or higher increased approximately linearly with increasing A-level tariff but that the effect of A-level points was most marked when they looked at the proportion of students achieving 1^st class degrees, which increased rapidly for the highest achieving students at A-level. This broadly agrees with other research carried out within disciplines. The review by Ferguson et al. (2002) into the factors associated with success in medical school found that 23% of the difference in achievement amongst undergraduate medical students could be accounted for by previous academic performance but that other effects were complex and merited further study. Similarly a report commissioned by the Sutton Trust in 2001 found that whilst A-levels were a ‘modest’ predictor of degree success they were the best available indicator and other measures such as additional aptitude tests were less conclusive. This correlated with their findings from the US that high school reports were better predictors of success than the SAT (Scholastic Aptitude Test).

In 2010 a report for the Department for Business Innovation and Skills stated that A-level tariff is the ‘best’ predictor of degree class and that additional aptitude tests offer no further information and do not help to distinguish between those students at the highest achieving end of the A-level spectrum nor help identify students from educationally disadvantaged backgrounds who have the ability to achieve at degree level. In contrast Huws et al. (2005) looked at A-level and GCSE grades and subject choice as predictors of success in psychology and found that whilst A-level grades did not appear to predict performance, GCSE grades in science and English did show a correlation with degree success.

The aim of this study is to examine the relationship between prior academic achievement (measured by A-level tariff on entry) and the success of students in Biosciences at the University of Exeter, and to examine whether as entry tariff increases the effect on degree success changes.

Biosciences at the University of Exeter currently recruits about 200 students per year on to a range of undergraduate programmes, some taught solely within Biosciences and others which are joint programmes with other disciplines. The majority of students (>85%) enter at level 1 with traditional A-levels and the current typical offer is AAB-ABB, including Biology and one other science at A2. Over the last 12 years Biosciences has undergone enormous change, with the establishment of a new campus near Falmouth in Cornwall, 75% of academic staff recruited since 2005 and a significant move towards higher entry tariff as the University continues its upward trajectory in the league tables (from 40th in the The Times Good University Guide in 1999 to 10th in 2011). The average A-level tariff for students on all Biosciences programmes has increased from 241 points in 19991 (equivalent to CCC) to 338 points in 2010 (equivalent to AAB). Retention rates on Biosciences programmes are good with a drop out rate of just 2.7% in year 1.

Methods

Data on entry grades, A-level subjects, level 1 and final degree marks were obtained from the student record system for the 2010/11 graduates (2008 entry; n = 130) and 2010 entrants (n = 93) for all Biosciences-only BSc programmes at the Streatham campus of the University of Exeter (Biological Sciences; Biological Sciences with Study Abroad; Biochemistry; Biological and Medicinal Chemistry; Biological and Medicinal Chemistry with Study Abroad; Biological and Medicinal Chemistry with Industrial Experience; Molecular Biology; Molecular Biology with Study Abroad; Molecular Biology with Industrial Experience). Programmes shared with other disciplines were excluded from this study (BSc Human Biosciences, BSc Biology and Animal Behaviour, Bachelor of Clinical Sciences). Of the students included in the study, 74% were following Biological Sciences (including those with Study Abroad programmes). Excluded from the study were students with entry qualifications other than A-levels, such as International Baccalaureate or Access courses. These alternative qualifications accounted for 11.5% of the total number of students.

Data on entry grades, A-level subjects, level 1 and final degree marks were also obtained from student records for the 222 BSc Biological Sciences graduating in 1999/2000, 2000/2001 and 2001/2002 (1997, 1998 and 1999 entry). Data for the 1997–1999 entry cohorts were combined, as year of entry had no significant effect on the final degree mark (ANCOVA, F = 2.13, P = 0.121) and was independent of the A-level score (entry year x A-level score interaction: F = 1.50, P = 0.225).

A-level entry tariff was calculated based on the three best A-levels achieved by students (excluding General Studies) using the UCAS Tariff (A*=140, A=120, B=100, C=80, D=60, E=40). 1997–1999 entry UCAS points based on A=10, B=8, C=6, D=4, E=2 were recalculated to the current UCAS Tariff. Level 1 and final degree marks were expressed as percentages to two decimal places. Final degree marks for 3-year programmes are calculated from level 2 and level 3 marks in a 1:3 ratio (1997–1999 entry) or 1:2 ratio (2008 entry); for 4-year (with Study Abroad; with Industrial Experience) programmes final degree marks are calculated from levels 2, 3 and 4 marks in a 2:1:4 ratio. The average entry tariff for the 1997, 1998 and 1999 cohorts was 241 points, for 2008 entry it was 308 points and for 2010 entry it was 338 points.

The data were analysed using regression analysis to assess the relationship between A-level tariff and a) Level 1 marks and b) final degree mark. The relationship between Level 1 marks and subject choice at A-level was also examined. Significance was analysed by ANOVA. Variation in final degree marks was further analysed by Analysis of Covariance (ANCOVA) with gender or A-level Mathematics or A-level Chemistry as covariate. Statistical analysis was carried out using Microsoft Excel 2007 (2008 and 2010 cohorts) and SPSS v10.0 for Windows (1997/8/9 cohorts).

Results

Entry tariff and level 1 marks

Level 1 marks in Biosciences programmes increased with increasing A-level tariff for all three cohorts studied (Figures 1a, b & c). This correlation was significant for all three cohorts (P <0.05) however the impact of A-level tariff on level 1 marks was found to decrease with increasing entry tariff (Table 1). In 1997/8/9 30% of the variation in mark was attributable to entry tariff (R² = 0.30, average entry tariff 241), compared to just 11% for the 2010 cohort (R² = 0.11, average entry tariff 338). For the three cohorts an increase in one A-level grade is associated with an average increase of 1.7%, 1.8% and 2.0% in level 1 mark respectively (0.086, 0.088, 0.10% per tariff point).

Table 1 Summary of data for regression analysis of marks at level 1 on entry tariff

YEAR OF ENTRY	YEAR OF GRADUATION	AVERAGE TARIFF	R^2	P	m	c
1997/8/9	1999/2000/2001	241	0.30	<0.001	0.083	35.3
2008	2010/11	308	0.24	0.0036	0.088	39.1
2010	2012/13	338	0.11	0.0015	0.10	32.2

Entry tariff and final degree marks

Final degree marks in Biosciences programmes increased with increasing A-level tariff for both cohorts studied who have reached graduation (Figures 2a and b). In agreement with the data presented above the impact of A-level tariff on final degree marks was also found to decrease with increasing entry tariff (Table 2). For both the 1997–1999 and the 2008 entry cohorts the correlation between entry tariff and grades drops slightly between the first year and the final degree mark (0.30 to 0.24 and 0.23 to 0.17 respectively) and so it may be expected that for the 2010/11 cohort the correlation will drop further, but this will require further analysis in the summer of 2013. This demonstrates that less of the variation in final degree mark can be accounted for by entry tariff than the variation in level 1 marks obtained. For each of these cohorts an increase in one A-level grade is associated with an average increase of 1.6% and 1.5% in final degree mark respectively (0.079, 0.077% per tariff point).

Figure 1 Regression analysis of level 1 mark against entry tariff for a) 1997, 8, 9 cohort, b) 2008 cohort and c) 2010 cohort

Table 2 Summary of data for regression analysis of final degree marks on entry tariff

Year of Entry	Year of Graduation	Average Tariff	R^2	P	m	c
1997/8/9	2000/01/02	241	0.23	<0.001	0.079	44.7
2008	2011	308	0.17	0.0015	0.077	41.6
2010	2013	338	Not yet graduated

Gender

Gender was found to affect final degree marks for both the 1997–1999 cohort (ANCOVA, F = 18.1, P < 0.001) and the 2008 cohort (ANCOVA, F = 21.4, P < 0.001) in a way that was independent of A-level score (F = 1.18 and 0.37, P = 0.28 and 0.54 respectively). The magnitude of this effect was great and was similar for both cohorts: female students on average gaining 5.1% and 4.9% higher final degree marks than males despite the gender difference for A-level grades on entry being less than one grade. Adjusting for differences in entry qualifications, female students on average gained 4.0 % and 3.8 % higher marks than males.

For the 2008 and 2010 entry cohorts further detailed analysis was undertaken to investigate the effect of A-level subjects choice on level 1 marks. This study was only conducted for BSc Biological Sciences and BSc Biological Sciences with Study Abroad students as those studying on other degree programmes had specific subject requirements on entry and so would not prove a fair comparison. This accounted for 79% of the total cohort. The A-level choices of these two cohorts are detailed in tables 3, 4 and 5.

Table 3 Percentage of students taking specific subjects at A-level for the 2008 and 2010 entry cohorts

A-level subject	2008 cohort	2010 cohort
Chemistry	55 %	50 %
Maths	27 %	36 %
Physics	10 %	20 %
Geography	31 %	24 %
Psychology	15 %	9 %

Table 4 Percentage of students taking ‘harder’ science A-levels (i.e. Maths, Physics or Chemistry) in addition to Biology

Number of ‘harder’ science A-levels in addition to Biology	2008 cohort	2010 cohort
0	34 %	29 %
1	41 %	43 %
2	23 %	22 %
3	2 %	6 %

Table 5 Percentage of students taking more than 3 A-levels and average Biology tariff (A = 120, B = 100)

	2008 cohort	2010 cohort
More than 3 A-levels	13 %	15 %
Average Biology tariff	101	114

Figure 2 Regression analysis of final degree mark against entry tariff for a) 1997/8/9 cohort and b) 2008 cohort

Effect of A-level Biology tariff

It is possible to see from the data in table 6 that the effect of A-level Biology grade is very small: it can account for 6.8% and 4.5% of the variation in level 1 marks for the two cohorts. The effect was significant for the 2008 cohort but not for the higher tariff 2010 cohort.

Table 6 Effect of A-level Biology grade on level 1 marks

	2008 cohort	2010 cohort
Average Biology tariff	101	114
R^2	0.068	0.045
P	0.009	0.082

Effect of A-level Chemistry

Possession of A-level chemistry was not found to affect level 1 marks for either the 2008 cohort (ANCOVA, F = 0.03, P = 0.86) or the 2010 cohort (ANCOVA, F = 0.39, P = 0.53) and was independent of overall A-level score (F = 0.01 and 0.05, P = 0.92 and 0.82 respectively).

Effect of A-level Maths

Possession of A-level maths was also not found to affect level 1 marks for either the 2008 cohort (ANCOVA, F = 0.78, P = 0.38) or the 2010 cohort (ANCOVA, F = 3.73, P = 0.06) and was independent of overall A-level score (F = 1.03 and 0.00, P = 0.31 and 1.00 respectively).

Effect of number of ‘harder’ science A-levels

Neither cohort showed any significant relationship between number of ‘harder’ science A-levels and level 1 marks (R² = 0.004 and 0.027, P = 0.547, 0.175 for 2008 and 2010 respectively).

Discussion

The results of this analysis show that there is a correlation between A-level tariff on entry and both level 1 and final degree success across all three cohorts studied, and whilst the magnitude of the correlation was variable it was significant in all cases. This is broadly in agreement with previous studies (Ferguson et al., 2002; Burrell, 2010; and McDonald et al., 2001) which showed that although not perfect, prior qualifications, in particular A-levels, were the best available predictor of success. This is in contrast to the findings of Huws et al. (2005) that in psychology at Aston University there was no correlation between A-levels and degree performance, however prediction of science subjects is generally more accurate than for arts and social sciences (McDonald et al. 2001).

Whilst there is some variation in the extent to which prior academic achievement can be used to predict degree success it is generally accepted that A-levels are the best available measure and that they outperform other assessments, such as aptitude testing (Department for Business Innovation and Skills, 2010). In agreement with the other studies our data showed that the correlation between level 1 marks and A-levels was closer than for final degree marks, demonstrating how during the journey through their programme students pick up numerous other skills which contribute towards success. For the 1997, 8 and 9 cohorts it was found that 30 % of the variation in level 1 mark could be attributed to A-level achievement (R² = 0.30) whereas when it came to the final degree mark only 23 % of the variation was down to A-level achievement. For the 2008 cohort the effect was the same with the contribution of A-levels similarly dropping by 7% from 24% to 17% between level 1 and final degree marks. When looking at the contribution of increasing A-level tariff to final degree mark the observed increases in degree mark of 1.6% and 1.5% per A-level grade for the 1997–1999 and the 2008 cohorts respectively represent a small impact but one which would have a significant effect on the degree outcomes for some students. For the 2008 cohort it can be estimated that adding 1.5% to the grades of all students would have improved the degree classification of 10 students out of the 130 graduating.

It is also possible to observe in these data an interesting trend not examined elsewhere. As average A-level tariff on entry has increased over the period studied, from 241 (CCC) from 1997–1999, to 338 (AAB) in 2010 there has been a reduction in the correlation between tariff and success, implying that for higher achieving students A-levels have less of an impact on degree success than for other students. Previous studies (Roddan 2002; Yorke and Longden, 2007; Scott and Graal, 2007) have looked at the factors affecting student retention and failure rates in higher education and demonstrated that both are due to a complex range of factors, including poor choice and advice in advance of choice of programme, lack of personal commitment, teaching quality and finance. Over the three years studied here the correlation between level 1 marks and A-level tariff has dropped from 30% of the variation being attributable to tariff in 1997–1999 to 24% in 2008 and just 11% in 2010. Whilst between the 1997–1999 cohorts and the 2008 entrants there was a massive change in circumstance and teaching staff within Biosciences which may have altered the approach to teaching and the methods employed, and there may also be A-level range restriction effects, the same cannot be said for the 2010 entrants for whom the teaching staff and methodology was broadly the same as the 2008 cohort and for which there is drop in correlation double that for the earlier cohorts compared (13% cf. 7%). This trend could be seen to reflect the observation that students with lower A-levels were at increased risk of poor performance on undergraduate medical courses (Yates and James 2007), where medical courses traditionally have high tariff requirements for incoming students, i.e. for weaker students entry tariff is a more significant predictor of level 1 success than for higher achieving students. One could extrapolate this trend to final degree success but due to the changes at Exeter in the period between the two cohorts studied who have already graduated it would be prudent to wait until the 2010 entrants have graduated in 2013 in order to comment in more depth on the continuation of this correlation. It is clear however from the work of Roddan (2002), Yorke and Longden (2007) and Scott and Graal (2007) that the factors affecting achievement and failure are many and complex.

As highlighted recently by Morgan (2011) the drive to increase entry tariff in order to be able to accept more students on to university courses contrasts with the accompanying policy of increasing social mobility and encouraging students from the most disadvantaged backgrounds to apply to the most selective institutions (Department for Business Innovation and Skills, 2011). Schemes such as Realising Opportunities, which aims to increase access to 12 top institutions for students in the ‘most able, least likely’ category through mentoring, skills development and ultimately alternative offers through UCAS, encourage universities to look beyond simply the A-level grades achieved by students when accepting them on to their courses. The University of Exeter participates in both the Realising Opportunities scheme and the Sutton Trust Academic Routes programme in which students can also receive an alternative offer through UCAS. The question of whether students with a lower entry tariff are less likely to gain top marks at university is one which will be of interest not just to Exeter but also many other top institutions as they attempt to balance the move to higher entry tariff and the need to widen access. Whilst this study gives food for thought we have not looked at the effect of school performance or school type on entry tariff, level 1 or final degree mark but this is certainly an area for further study as the higher education market undergoes significant change in the next few years.

Gender was found to have a marked effect on degree marks: women obtaining on average about 3.9 % higher degree marks than men with equivalent entry qualifications. This is huge difference — more than one third of a degree class — and also cries out for further study. To put it another way, male bioscience students achieve on average the same as female students with A-levels two grades lower. These findings are consistent with those of Ferguson et al. (2002) that women medical students perform better than men in medical training. Such a sex differential is not of course peculiar to higher education: under-performance of boys is well documented in schools, and it could be that some factors (e.g. motivational) are common. More detailed analysis was undertaken for BSc Biological Sciences students in order to investigate the effect of subject choice at A-level. Research (Scott, 2006) has shown that enjoyment of the subject was the key reason that students studied Biological Sciences to degree level but whilst 69% of students agreed that it was their best subject at school, only 8% said this was the most important factor influencing their choice, indicating that students do not simply study the subject in which they gain the highest marks. It is therefore interesting to see that whilst the correlation between a student’s A-level Biology grade and their level 1 marks was significant for the 2008 entrants, it was not significant for 2010 entrants and that in both cases the correlation was small, with just 6.8% and 4.5% of the variation in performance attributable to Biology grade alone for the two years in question. This demonstrates that whilst overall academic achievement at A-level is important in determining success at university in biosciences, the same cannot be said for achievement in A-level Biology alone.

Whilst A-level Chemistry is not a requirement for admission to BSc Biological Sciences at the University of Exeter, students at Exeter often cite a lack of A-level Chemistry as a reason for struggles in term 1 (where level 1 biochemistry is taught) and subsequent lower than anticipated achievement. However the data analysed here show no significant correlation between A-level Chemistry grade and overall level 1 achievement in Biological Sciences, so this may be an issue of perception rather than a reality. There was also found to be no correlation between A-level Maths grade and achievement at level 1, justifying the fact that neither of these subjects is a pre-requisite for admission to Biological Sciences. There is a requirement for students to obtain A-levels in Biology ‘plus one other science’, for which the list is wide and varied, and whilst the total number of students taking Biology A-level nationally has remained fairly constant there has been a drop in those taking Chemistry, Maths and Physics (Select Committee on Science and Technology, 2006) in favour of ‘easier’ A-levels such as Physical Education, Psychology and Environmental Science, which are all accepted as the second science subject for admission to Biosciences at Exeter.

In the Select Committee on Science and Technology (2006) report it is reported that Chemistry, Biology, Maths and Physics do appear more difficult subjects at A-level; comparison of average A-level points obtained for a range of subject for students with comparable GCSE achievement showed that students taking these subjects could expect a significantly lower grade than those taking other subjects at A-level. Coe et al. (2008) used five different methodologies and arrived at a similar conclusion for Chemistry, Physics and Biology A-levels. It might therefore be reasonable to suggest that students with the same overall tariff on entry but who had studied these ‘harder’ sciences at A-level might be better placed to achieve more highly at degree level in a science subject at a top institution as higher grades in these subjects are harder to achieve. However there was found to be no correlation between number of ‘harder’ science A-levels (i.e. Maths, Physics or Chemistry), in addition to Biology, and level 1 achievement in either of the two cohorts studied, implying that this is a perceived rather than real concern. This mirrors the lack of correlation between success at level 1 and A-level grades in Chemistry and/or Maths.

It should be recognised that as A-level entry tariff increases there is a compression effect as students’ entry qualifications are bunched towards the top end of the A-level spectrum (3 x A*, 420 UCAS points). Despite this (standard deviation in A-level tariff for the 1997–1999 cohort was 49.7 points, and for 2008 it was 35.2 points) the findings in this study show that as entry tariff increases the contribution of A-level tariff to university marks decreases, implying that for the most able students there are many complex factors which determine overall degree success. This study has also shown that whilst A-level entry tariff does show a correlation with both level 1 and degree success, the contribution of A-levels to marks achieved at university decreases as the students progress from level 1 through to their final degree mark. Interestingly there is little correlation between the subjects chosen at A-level and university level achievement, counter to popular opinion on the difficulty of different A-level subjects and the way they prepare students for a university education. In order to further identify students who may not fulfil their potential and who may be at risk of under-achievement it is important to continue this research further to look at ‘softer’ factors which may influence achievement, such as student engagement, financial and social pressures and career objectives, though these are necessarily more difficult to measure. It will also be important to re-examine the data upon graduation of the 2010 entrants in order to assess whether the trends reported here are continued, and it may be predicted that the correlation between A-level tariff and degree success for those graduating in 2013 will continue to weaken.

Acknowledgements

Thanks must go to colleagues in Biosciences and central University administration for their help collating the data used in this research.

Notes

1 UCAS points recalculated from 1997,8 and 9 system, which was based on A=10, B=8, C=6 points etc, to current UCAS Tariff in order to allow direct comparison.