It was an exciting new venture when I was asked to take part in the 1980 follow-up of the 1970 British Births Cohort Study. This study related to all those children born in England, Wales and Scotland during the week 5-11 April 1970. There was a follow-up study in 1975, when the children were aged five, and a further follow-up in 1980, when they were aged 10.
The original number of children studied was 14 906. In 1980 the number was 12 905. The main reasons for this drop in numbers appear to have been the difficulty in tracing some families, their removal overseas and, sadly, in some cases, death.
At birth and in subsequent follow-up studies medical, social and, in 1980, educational data were obtained. My own involvement was in the follow-up in 1980; I was invited to explore what could be found out from the cohort study about the existence and nature of dyslexia (note 19.1).
The organisers of the study had already agreed that certain educational tests should be given to children when they reached age 10. They included a test of single word reading, a test of reading comprehension, which afterwards became the Edinburgh Reading Test (1985), a spelling test in the form of a dictation and tests of non-word reading and non-word spelling. Non-words - sometimes called 'nonsense words' or 'pseudowords' - are combinations of letters which do not spell actual words but which are pronounceable. They are included in many research studies on the grounds that they test the person's knowledge of letter-sound correspondences and their ability to use analogies with familiar words. In the present study examples of the eight non-words to be read included 'ronumental' and 'teague'. The four non-words which had to be spelled were 'prunty', 'slimber', 'grondel' and 'blomp'.
There was also a mathematics test, a test of whether the child could understand the significance of pictures and four items from the British Ability Scales (BAS) (Elliott et al, 1983) (note 19.2). These were: Similarities, Matrices, Recall of Digits and Word Definitions.
The scores on these tests were standardised so as to give a mean (average) of 100 and a standard deviation of 15. This made it possible to know whether a particular child was above or below average on a given test and to ensure comparability between tests.
I took the view that a necessary condition for a diagnosis of dyslexia was poor reading or spelling in relation to intelligence. In choosing a measure of intelligence it was necessary to bear in mind the existence of the so-called ACID profile (see Chapter 9) and the general danger that on some of the traditional tests of intelligence dyslexic children might obtain scores that did not do justice to their intellectual level.
Fortunately I had had the opportunity to see in advance some work by my colleague Dr Michael Thomson on the performance of dyslexics on the BAS (note 19.3). Michael found that on some items the dyslexics in his group scored above the average as determined by the published norms and on some items below it. He found that on both the Similarities and Matrices tests his subjects scored above average. Since in the British Births Cohort study scores were available on both these tests, we decided that they should be used in conjunction as a measure of intelligence. We therefore standardised the combined scores on the Similarities and Matrices tests so that, like the other tests (see above), it should have a mean (average) of 100 and a standard deviation of 15.
With some hesitation, we decided to exclude from our main analyses of the data those children with scores on the Similarities and Matrices tests of less than 90. This was not because we believed that children of low intelligence could not be dyslexic but because we thought that in their case there might be complicating factors which were not typical of dyslexics within the average ability range.
The test of single-word reading (word recognition) seemed preferable to the Edinburgh Reading test; the latter involved comprehension and there was reason to believe that dyslexics were unimpaired at this. Such a test would not therefore reveal the full extent of their reading disability. We also decided to pay particular attention to the spelling scores, on the grounds that a child who was spelling adequately at age 10 could not be severely dyslexic. To qualify as a dyslexic, therefore, a child had either to be a severe underachiever at spelling regardless of their score on the word recognition test or a moderate underachiever at spelling who was also a severe underachiever at word recognition.
On the basis of scores on the Similarities and Matrices tests on the one hand and scores on the word recognition test and spelling tests on the other we divided the cohort into three groups - normal achievers, moderate underachievers and severe underachievers (note 19.4).
As possible indicators of dyslexia it was agreed to include items from the Bangor Dyslexia Test, the first edition of which was due to be published in 1982. A decision had therefore to be made as to which of the 10 items to choose. The constraints were that there was time for only about three items and that their administration should be straightforward, since they were to be given to the children by their class teachers. We had therefore to choose items which could most easily be administered in a standardised way.
The items which we eventually chose were Left-Right, Months Forwards and Months Reversed. The Recall of Digits item from the BAS had already been included (see above), and it made sense that we should add this test to the other three, since there was already good evidence that poor readers also performed poorly on this item (Miles and Ellis, 1981, p. 230). In earlier times it had been assumed that ability to recall digits was an indication of intelligence, but by now it was plain that many highly intelligent dyslexics often found this task difficult.
We decided to refer to these four items as the 'supplementary' items, since they provided a supplement to the rest of the educational data. We did not at this stage wish to beg any questions by calling them the 'dyslexia' items, since we believed that they could be regarded as indicators of dyslexia only in the context of severe underachievement. In a current paper, for ease of reference, we have used the expression 'possible indicators (of dyslexia)', abbreviated to PIs (Haslum and Miles, in press). Specified responses were indicators of dyslexia only in the appropriate context.
It is not at all easy, of course, in a survey on this scale to avoid mishaps, and we do not know how much our results were contaminated by 'noise' resulting from teachers' failures to follow the instructions to the letter (note 19.5). However, the size of the population studied made it unlikely that such mishaps, if they occurred, would seriously affect our results.
Our first task was to specify how children whom we believed to be dyslexic should be picked out. If we simply showed that some children had reading and/or spelling problems, or even unexpected reading or spelling problems in view of their intelligence, this would have provided no justification for claiming that they were dyslexic in the required sense. To adopt the terminology used in Chapter 10, if it were to be of any use, the word 'dyslexia' needed to contain 'bite' and, in the terminology of Chapter 18, provide a strong taxonomy. If the dyslexia concept were to achieve this, it would lead to findings which we could not otherwise have predicted or discovered. We had therefore to decide what definition of dyslexia would generate interesting questions and make possible their investigation, given the basic components of the syndrome as described in Chapter 18.
It was also important that we should not make a stipulation which departed widely from the commonly accepted meaning of the word 'dyslexia', in so far as there was one. In this connection we needed to follow the advice of the philosopher Sidgwick (1922, p. 264) quoted in Chapter 7. Sidgwick made the general point that it was in order to make our ordinary concepts more precise as long as we did not depart too far from ordinary usage - 'We may, so to speak, clip the ragged edge of common usage, but we must not make excision of any considerable portion.'
In the present context what was needed was to operationalise the concept of dyslexia, that is provide a set of operations in terms of which its presence or absence could be determined - in such a way that we were not departing too far from the commonly accepted meaning of the word. As will be seen from Chapter 20 we in fact ended up by making proposals that the boundaries of the concept of dyslexia needed to be revised, but at least our starting point had to be the commonly accepted meaning of the word. It would then be necessary to test hypotheses as to ways in which dyslexics and non-dyslexics might be different. In this way, were scientifically important differences to be found, progress would most certainly be being made.
It was also important, in this connection, to conform to Popper's principle of falsifiability (Popper, 1963, p. 36): it was logically incumbent on us to specify what outcomes would indicate that we were wrong. Then, if these outcomes failed to occur, we could say that our attempts to knock down a particular hypothesis had been to that extent unsuccessful. Those who claimed that the hypothesis was wrong were left with facts which somehow or other they would need to explain away.
On the basis of my clinical experience I had reached the conclusion that typical dyslexics were those who were retarded at reading and spelling in relation to their intelligence and who additionally showed positive indicators on the Bangor Dyslexia Test. There would not be the same opportunity for detailed examination in the British Births Cohort Study, since there were only four 'dyslexia' items in place of the 10 which were available in the Bangor Dyslexia Test, and there would not be any opportunities for me to observe the children at first hand or give further tests. However, I had reasonable confidence on clinical grounds that the four supplementary items were relevant to dyslexia in the sense of the word which I needed. It was therefore a question of finding out if dyslexia-positive responses to these items were related to other findings.
Although at the time I thought of dyslexia as an 'all or none' phenomenon - either you were dyslexic or you were not - I decided that it was necessary to have what we called a 'buffer' group - those who were possibly dyslexic but about whom there was some doubt. In addition to subdividing the children into normal achievers, moderate underachievers and severe underachievers, therefore, my colleagues and I also subdivided the children into three groups based on their performance on the supplementary items. For this purpose we used the method of scoring which had been devised for the Bangor Dyslexia Test, where a typically dyslexic response was scored as 'plus', a marginally dyslexic response as 'zero' and a response free from any indicators of dyslexia as 'minus'. On the Bangor Dyslexia Test a 'zero' was scored as half a 'plus', so that a diagnosis of dyslexia could be made either on a relatively small number of major indicators or a rather larger number of minor ones.
For purposes of the British Births Cohort Study we decided that those with two or more 'pluses' or three or more 'zeros' on the supplementary items, if they were also severe underachievers, should constitute our dyslexic group; those with two 'zeros' or a 'plus' and a 'zero' should belong to the buffer group, while the remainder, comprising those with a single 'plus', a single 'zero' or no positive indicators at all, should be put into a third group. For ease of reference we designated those with two or more pluses or three or more 'zeros' as 'underachievers A', those in the buffer group as 'underachievers B' and the remainder as 'underachievers C'. We later extended this A, B and C classification to the moderate underachievers and the normal achievers, since we found some interesting results among those who came out as normal achievers but who, nevertheless, were showing positive indicators on the supplementary items.
A dyslexic was therefore defined as a severe underachiever at word recognition or spelling who had two or more 'pluses' or three or more 'zeros' on the supplementary items. We did not wish to claim that the boundaries marked by the A, B and C categories were firm ones; it was rather that, if the supplementary items were of value, among the severe underachievers there would be a greater proportion of typical dyslexics in the A group and fewest in the C group. Rather than doing a head count of individual dyslexics, our statistics would show up trends on the basis of which conclusions about dyslexia could be drawn.
Although the number of children tested was 12 905, it was inevitable that in a survey of this magnitude some of the results had to be recorded as 'missing data'. Full details to account for small discrepancies in the number of children tested are given in the original papers. The main thing which we had to be careful about was that there should not be sufficient missing cases to bias our results.
It will be noted, both from this Chapter and from the discussion of the Bangor Dyslexia Test in Chapters 7 and 8, that we specified various different ways of being dyslexic. A child could be dyslexic either if ... or if ... Dyslexia, in other words, is a disjunctive concept: some limbs of the disjunction need to be present but not all of them and not necessarily the same ones in different individuals. This notion of a disjunctive concept is one to which I shall return in Chapter 23.
Was this article helpful?
This is a comprehensive guide covering the basics of dyslexia to a wide range of diagnostic procedures and tips to help you manage with your symptoms. These tips and tricks have been used on people with dyslexia of every varying degree and with great success. People just like yourself that suffer with adult dyslexia now feel more comfortable and relaxed in social and work situations.