In the 1940s the Belgian philosopher Albert Michotte identified our tendency to believe we could see causality. His book, The Perception of Causality, published in French in 1945 showed how certain very simple visual sequences carry the appearance of causal connectedness. Click this link for an example. This paper is a good recent update on how illusions of causality bias our judgement.
Human beings are natural pattern seekers. We see shapes in clouds, faces in wallpaper and meaning where there is just random noise. In particular, we believe we can see causes when all we can actually see are effects. In teaching, we look at what happens in a classroom and think we know why it happened. We may, occasionally be right, but usually we’ll miss the hopeless complexities of real life preferring to stick with a convenient narrative: Miss Crumb is an ineffective teacher; Gavin is a feckless, work shy toe rag; Parvinder always tries her best. This is much easier than actually doing the hard work of trying to find out what else might have caused the effects we’ve observed.
One of the most prevalent examples of the perception of causality in education surrounds the observation that girls outperform boys. There’s no end of data to support this assumption and it really does seem to be the case that, on average, girls do better at school and are 75% more likely to go to university than boys. The OECD have found that boys are 50% more likely than girls to fall short of basic standards in reading, maths and science. And of course, everyone knows why. As this Economist article explains, there are 3 clear reasons why girls are doing better than boys: girls read more, spend longer doing homework and boys are ‘too cool for school’. Simple.
Consequently many, if not most, schools have a gender policy. As Chris Curtis points out here, teachers are under pressure to create a ‘boy friendly’ curriculum to ensure that boys are more engaged in lesson content, thereby magically closing the attainment gap.
If only reality really were this simple. My suspicion is that what works well for girls will also work well for boys. Chris makes the following observations:
A boy who knows what they want from school succeeds.A girl who knows what they want from school succeeds.A student who knows what they want from schools succeeds.
To believe otherwise is to believe that we know the causes for the effects we observe. But it’s so tempting, isn’t it? In Chapter 1 of my book, What if everything you knew about education was wrong? I tell the story of ‘Mr Garvery’.
During a training session, Mr Garvery and his colleagues were presented with data showing a difference in the mean scores for average GCSE points – with girls achieving a higher mean than boys. The obvious conclusion drawn was that this difference mattered and something needed to be done. Urgently. No ‘proper’ statistics were used to quantify the significance of this difference. So Mr Garvery went back to the raw data and performed a factor analysis of the impact of the following variables:
- Free school meals (FSM)
- Originating primary school
- Key Stage 2 English/maths/science results
- Key Stage 3 English/maths/science results
- Reading age
- Pupil attendance
- Teacher attendance
- Special educational needs (SEN)
- English as a foreign language (EAL)
All had an impact ‘on average’, but the most significant factors were:
- Teacher attendance
- Pupils’ attendance
- Key Stage 2 English results
Of those factors measured, gender and free school meals were the least significant. When Mr Garvery shared this finding with his head teacher, his was told to stop causing trouble and come up with a suitable gender policy.
Why has the story of boys’ underachievement become such a widely accepted and compelling narrative? The problem is that we see graphs with girls’ performance clearly ahead of boys’, so the cause must be due to gender. The way information is presented makes it appear that gender is the biggest factor underlying students’ achievement, but the data makes it clear that attendance and prior achievement correlate much more closely.
Mr Garvery had conclusive evidence that gender difference was among the least important factors impacting pupils’ performance. He continued his data exploration and surveyed all Key Stage 4 pupils for gender, as well as whether they lived in houses with an odd or even house number, whether they owned a games console and whether they were left-handed or right-handed.
When I tell school leaders that whether you’re left or right-handed might have more bearing on your educational attainment than your gender they sometimes start nodding and you can see a left-handed policy being born. After all, this findings still seems plausible. When I reveal that console ownership was even more statistically significant they look worried: are they going to have to blow the Pupil Premium budget on X Boxes? But when I tell them the most statistically significant factor correlated with outcomes was whether students live in an odd or even house, the relief is palpable. No one can believe that the number of the house you live in can in any way be causally linked with attainment. It’s a silly idea and we dismiss it immediately.
The ‘pattern’ of boys’ underachievement is compelling because of the way we think about gender: girls are quiet, hard-working and sensible; boys are immature, unruly and easily bored. But as any teacher and every parent could tell you, these are stereotypes – a shorthand that saves us from having to think about reality. As ever though, reality is always more complex. Recent research into achievement and gender differences has found that school behaviour is much more likely to be a decisive factor for achievement than gender. Hard-working pupils achieve good grades while badly behaved pupils perform more poorly and get worse grades. The distribution of boisterous pupils among the two genders is much the same – about 40 per cent are girls.
I’m not suggesting gender has nothing to do with attainment – it probably does have some bearing – but maybe a lot less than we’re inclined to believe. And the extent to which gender might be causal is more likely due to cultural rather than biological causes, as this article makes clear. Our best bet is probably to insist on high expectations for all students and not let boys get away with being ‘just boys’.