Age-adjusting exam scores is more unfair than what it replaces

16th May 2013 at 10:05



Summer-born children ‘more likely to have been born in summer’ claims new report


Fans of educational comedy weren’t disappointed this week, as the Institute for Fiscal Studies released their research showing that children born in August, and therefore young for their cohort, were at a disadvantage compared to their more venerable peers born a few months later. This select group was deemed less likely to get a degree, more likely to be identified for special educational needs, less likely to get five good GCSEs and so on.


They came to this conclusion after looking at the data from 48,500 children, so the evidence base is broad enough to demand attention. Of course, this proposition isn’t new, and may works of literature support the general analysis, from grown-ups like Bell and Daniels (2010) to Malcolm Gladwell’s pop psychology stocking-filler Outliers in 2008, the Freakonomics of success theory.


Why does it matter? Because sober people with degrees believe that this means we should adjust children’s attainment scores according to their birth month, in order to achieve a true reflection of their potential, and avoid damning them to a lifetime of misidentification. So if two children sit the same exam, then August’s child will see their scores massaged up a little compared to their wintry table mate.


"Our findings point to a simple solution to the pitfalls of testing children born at the start and end of the academic year at very different ages," says Ellen Greaves of the IFS. "Age-adjusting the cut-offs...would ensure that no child is prevented from going on to further or higher education simply because of the month in which they were born."


The birth-month factor isn’t generating a twelve-fold caste system: 12.5% of August born children are identified as having special needs compared to 7.1% of September babies (which might say more about our ridiculously trigger-happy identification system of SEN rather than an actual issue); the former only 1% less likely to gain a degree than the latter; 54% achieving 5 A*-Cs as opposed to 60% and so on. And by the time they reach adulthood, employment prospects and earning power are unaffected, so at least it appears to even out after time.


My main objection to this proposal is based on fairness. While it appears an easy adjustment to simply create an algorithm that redresses the educational balance, justice is not so easily obtained. Perhaps some children are disadvantaged by their birth month; maybe they do feel intimidated by all the fractionally taller children with handfuls of extra weeks to their life span. Who knows? Even if there are, there will also be many for whom being marginally younger will have no significant impediment. Would they also have their scores adjusted upwards? There will also be children for whom being older presents no advantage. Relatively speaking, their achievements will be adjusted downwards.


In other words, no blanket application of a statistical manipulation can provide a fair account of a whole cohort. The perceived injustice – if any exists – has simply been moved around. Winners and losers will accrue in an arbitrary manner, and years later, another research paper will suggest some other bone-headed solution to the problem that life is not perfectly equal or fair.


What further adjustments can we expect? Higher grades for children who sit next to the window and therefore are exposed to more distractions? Lower grades for children with parents who read stories to them at night? How about this: grades given to children based on what they have actually achieved, or how they have actually performed in real life, as opposed to the wish-fulfilment of well-meaning educationalists who should really consider how much harder their proposals would make things for teachers and children before floating ideas like this.