A Level qualifications are an important requirement for university entry and other post-secondary destinations. If a candidate fails to achieve the required grade, then they may wish to re-sit the qualification. Recent reforms to the structure of A Levels, such as the removal of the January examination session and the move to linear A Levels, may increase the proportion of candidates choosing to re-sit an entire qualification. In this Data Byte we look at the re-sitting of A Level qualifications prior to the introduction of these reforms.
What does the chart show?
The chart shows the observed qualification re-sit rate, i.e. the number of re-sits divided by the total number of entries in that subject, for 94 A Level subjects (the black dots). We define a re-sit candidate as one who entered an A Level subject more than once in different sessions, irrespective of the choice of subject specification and/or awarding body.
Based on the observed data, we then estimated the probability that a candidate would re-sit a subject given its entry size using a beta binomial regression. The dark grey ribbon shows the 95% prediction interval for this probability and the dashed line indicates the median predicted re-sit probability. We have assumed that the re-sit probability is proportional to a subject's entry based on the intuition that the more often a subject is stipulated as a requirement for university entrance and other post-secondary destinations, the more entries we would expect to see in that subject and the higher the stakes riding on the assessment's outcome.
The data come from the National Pupil Databases for the 2013/14 and 2014/15 academic years and allow us to analyse the full cohort of students who started A Level study in the autumn of 2012. These individuals sat a total of 736,264 A Levels with an average re-sit rate of 4%; on average, re-sit candidates improved their grade by 0.72 grade points.
Why is the chart interesting?
The majority of subjects fall within the 95% prediction interval but this interval is relatively wide for a given entry size. Further research would be required in order to explain this variability and to better understand why a candidate chooses to re-sit a particular subject. One particularly interesting group of subjects is the 'community languages'. The figure shows that many of these subjects have very high re-sit rates given their entry size, in particularly languages like Bengali and Chinese. This suggests that there may be family or community factors that influence the decision to re-sit these subjects.
The data presented here also provide a baseline against which future re-sit rates can be compared, for example, to explore the effect of recent A Level reforms on re-sit rates and outcomes.
Full details on the data presented here, as well as re-sit rates by subject, can be found in:
Gill, T. & Crawford, C. (2016) 'The re-sitting patterns of a cohort of A level students'. Statistical Report Series No. 110
, Cambridge Assessment. http://www.cambridgeassessment.org.uk/our-research/all-published-resources/statistical-reports/