How useful are official statistics for understanding differences in educational achievement by social class, gender and ethnicity?
How do GCSE results vary by social class, gender and ethnicity?
The data below is taken from the Department for Education’s document – GCSE and Equivalent Attainment by Pupil Characteristics 2014
Firstly – GENDER – Girls outperform boys by about 10 percentage points. 61.7% of girls achieved at least 5 A*- C GCSEs (or equivalent) grades including English and mathematics compared to 51.6% of boys; this is a gap of 10.1 percentage points.
Secondly – ETHNICITY – Chinese pupils are the highest achieving group. 74.4% of Chinese pupils achieved at least 5 A*- C GCSEs (or equivalent) grades including English and mathematics. This is 17.9 percentage points above the national average (56.6%). Almost half of Chinese Pupils are achieving the English Baccalaureate (49.5%); 25.4 percentage points above the national average (24.2%).
Children from a black background are the lowest achieving group. 53.1% of pupils from a black background achieved at least 5 A*- C GCSEs (or equivalent) grades including English and mathematics; this is 3.4 percentage points below the national average (56.6%). However, things are also improving: 75.5% of black pupils are making the expected progress in English and 68.4% in mathematics; both above the national average of 71.6% for English and 65.5% for mathematics.
Thirdly – SOCIAL CLASS – Here, instead of social class we need to use Pupils eligible for Free School Meals (FSM) (meaning they come from a household with an income of less than £16000) – FSM pupils are nearly 30% points behind non FSM pupils. 33.5% of pupils eligible for FSM achieved at least 5 A*- C GCSEs (or equivalent) grades including English and mathematics compared to 60.5% of all other pupils. This is a gap of 27.0 percentage points. 36.5% of disadvantaged pupils achieved at least 5 A*- C GCSEs (or equivalent) grades including English and mathematics compared to 64.0% of all other pupils, a gap of 27.4 percentage points.
The government stats also include achievement data by ‘disadvantage’:
Disadvantaged pupils are defined as pupils known to be eligible for free school meals in the previous six years as indicated in any termly or annual school census, pupil referral unit (PRU) or alternative provision (AP) census or are children looked after by the local authority for more than 6 months.
Other statistical data included in the pupil characteristics report
The Department for Education also collects data and reports on educational achievement by English as a second language, and special educational needs. Look it up if you’re interested, I’m limiting myself here to educational attainment by ‘social class’, gender and ethnicity.
Some Strengths of Official Statistics on Educational Achievement by Pupil Characteristic
ONE – Good Validity (as far as it goes) – These data aren’t collected by the schools themselves – so they’re not a complete work of fiction, they are based on external examinations or coursework which is independently verified, so we should be getting a reasonably true representation of actual achievement levels. HOWEVER, we need to be cautious about this.
TWO – Excellent representativeness – We are getting information on practically every pupil in the country, even the ones who fail!
THREE – They allow for easy comparisons by social class, gender and ethnicity. These data allow us to see some pretty interesting trends – As in the table below – the difference between poor Chinese girls and poor white boys stands out a mile… (so you learn straight away that it’s not just poverty that’s responsible for educational underachievement)
FOUR – These are freely available to anyone with an internet connection
FIVE – They allow the government to track educational achievement and develop social policies to target the groups who are the most likely to underachieve – These data show us (once you look at it all together) for example, that the biggest problem of underachievement is with white, FSM boys.
Some Disadvantages of the Department for Education’s Stats on Educational Achievement
ONE – We need to be a little cautious about the validity of some of these results, especially when making comparisons over time. This is because until last year schools could count any one of 3000 ‘soft’ subjects as equivalent to a GCSE, which could make the results look better than they actually are. Also, with coursework subjects there is a potential problem with ‘grade inflation’ within schools, and not to mention the fact that with coursework we are least partially measuring the degree to which parents have helped their children, rather than their children’s actual personal achievement.
TWO – comparisons over time might be difficult because of recent changes to the qualifications that are allowed to be counted towards attainment measurements. In 2014 the following changes were made:
1. The number of qualifications which counted towards ‘GCSE or equivalent’ results were drastically reduced – around 3,000 unique qualifications from the performance measures between 2012/13 and 2013/14.
2. The associated point scores for non-GCSEs was adjusted so that no qualification will count as larger than one GCSE in size. For example, where a BTEC may have previously counted as four GCSEs it will now be reduced to the equivalence of a single GCSE in its contribution to performance measures.
3. The number of non-GCSE qualifications that count in performance measures was restricted to two per pupil.
All of this has had the effect of making the results look worse than they actually are:
THREE – These stats don’t actually tell us about the relationship between social class background and educational attainment. Rather than recording data using a sociological conception of social class, the government uses the limited definition of Free School Meal eligibility – which is just an indicator of material deprivation rather than social class in its fuller sense. Marxist sociologists would argue that this is ideological – the government simply isn’t interested in measuring the effects of social class on achievement – and if you don’t measure it the problem kind of disappears.
FOUR – and this is almost certainly the biggest limitation – these stats don’t actually tell us anything about ‘WHY THESE VARIATIONS EXIST’ – Of course they allow us to formulate hypotheses – but (at least if we’re being objective’) we don’t get to see why FSM children are twice as likely to do badly in school… we need to do further research to figure this out.
No doubt there are further strengths and limitations, but this is something for you to be going on with at least…