This is not a new fact, but I was recently reminded that 20% of students in English schools are classified as have a special educational need. I am sure that I heard this when I lived in the U.K. but I didn’t think about it much at the time. Reflecting on it now, it seems extraordinary.

Imagine we used some kind of test to identify students with special educational needs. Now, I know that we would not *actually do this* but humour me for a minute as I explain. We would expect the results to follow a normal distribution around a mean value. Normal distributions are mathematically well-described :

If we selected the lowest 20% of scores then some of these students *would actually be within one standard deviation of the mean*.

For comparison, the mean adult male height in Australia is about 176 cm (5’9″) and we could expect, based on other data, that the standard deviation is around 7 cm (3″). So a man who is 169 cm (5’6″) would be about one standard deviation below the mean. Would you classify such a man as being especially short? Indeed, is a guy who is 183 cm (6′) especially tall?

For interest, if we decided to only classify those students who are *two* standard deviations below the mean as having a special educational need (equivalent to an adult male height of 162 cm or 5’4″) then this would represent just over 2% of the population.

There is a problem with this comparison. Students are identified as having a special educational need in a number of different categories. So the 20% could be composed of students who are all in the bottom 16% in that particular category. I wonder how much of a factor this is. It seems likely that one special educational need will correlate to another.

One of the invidious aspects of seemingly benign education theories like learning styles is that they imply labelling students. You have to wonder how a teacher’s expectations change when a child is labelled as a ‘kinaesthetic learner’. Clearly, a label of ‘special educational needs’ in some way *has* to alter teachers’ perceptions of a student. What will this mean for the student?

Well, there is evidence that it could affect teacher assessments.

And what about early reading? The following is quite possible: Students are taught with less effective methods, a significant minority of these children therefore do not learn to read, they then gain the label of having a special educational need which explains this and so nobody seeks to scrutinise the teaching approach. I am sure that this doesn’t happen but it seems to me that it *could*.

“We would expect the results to follow a normal distribution around a mean value. ”

This is a very dangerous assertion. Many real world distributions are not Normal. Intelligence is only Normally distributed because the raw data has been transformed.

Could you maybe expand on this a little?

I’ll have a go. Give me a day or so, I just fell off a ladder today!!

The Normal distribution is what you get when the attribute being measured is the sum of a large number of independently acting influences on the thing being measured. This is best seen when samples of a large enough size (at least 30) are taken from a population whose distribution is unknown, and the means or averages of the sample values in each sample then have a close to Normal distribution. This is a loose description of the Central Limit Theorem.

Yes, I agree with andrewsabisky (below) that there are naturally occurring quantities with a Normal distribution, but there are many more without. More importantly is the figure of 20%. There are institutional interests in classifying kids as SEN as it has an effect on the finances.

There is also an inflation in diagnosis, particularly in ADD and Autism, which has helped to push up the percentage.

There are in fact good reasons to think that intelligence is normally distributed, but it is true that the normal distribution of intelligence test scores produced by a typical battery is an artifact of test construction, though historically this was not the case and tests that were not supposed to yield a normal distribution still did so.

However, I think this artifact probably reflects reality quite well for the following reasons:

1) the central limit theorem

2) when we can measure intelligence on a ratio scale, the scores are more or less normally distributed. Digit Span Forwards and Backwards are both normally distributed, as is reaction time. Note that -/+ 2 SD outside the mean normality is no longer quite the case – there is are more very low IQ scores than you would expect in a perfectly normal distribution (due to rare genetic disorders of large effect), and there are also slightly more very high IQ people than you would expect also.

Greg,

On a different note, special education is not only for those on the left side of the curve but also on the right side. Western societies have ONLY focussed on children who fall under the norm, neglecting children with special needs on the other side. All of this with the thought “They’ll solve it themselves”. Of course low- achieving children need special ed, but also “cognitively excellent” children tend to end up with learning problems and behavioural problems because schools and [politics neglects them: high potentials that achieve lower than “average” students.

paul

This is a good point and the needs of these students must be addressed. However, I think the definition of ‘special educational needs’ in the UK excludes these students because they have a separate category of ‘gifted and talented’. I might be wrong about this.

And what about the “twice exceptional” kids that are both gifted and have a learning difficulty. Yes these special people do exist. Do you count them twice?

Cognition and learning one category of SEN; what about pupils with visual and/or hearing impairment; high-functioning asd, medical conditions such as muscular dystrophy that don’t impact on intelligence?

These are good points but I wonder what proportion of the school population would be in these categories. 20% is a *very* large number of students.

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/362704/SFR26-2014_SEN_06102014.pdf This may be of interest.

And it must be borne in mind that mental health is a category of SEN under the new C of P .