The Psychological Assessment of Reading

John Beech, Chris Singleton

• Publisher: Routledge
• Year: 2021
• ISBN:
  9781000446029
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

A useful guide to best practice including reviews of the latest and most helpful tests available. In Part One, contributors discuss the theory of reading assessment including issues such as screening, legal aspects, memory and visual problems, computer based assessment and the dyslexias. Part Two contains the review section where experts give comprehensive reviews of named tests.

1 The psychological assessment of reading

DOI: 10.4324/9781003209225-1

Theoretical issues and professional solutions

John R. Beech and Chris Singleton

Why do we need to assess literacy? Often because we, as professionals (either psychologists or specialist teachers), are called upon to do so. Usually it is because a child is experiencing (or is perceived by parents and/or teachers to be experiencing) some difficulties in learning to read, write or spell. We have to establish: (1) whether there really is a difficulty (not just an imagined problem); (2) the extent of the difficulty, should it exist; (3) the most likely cause, or causes, of the difficulty, and finally, (4) we are usually asked to recommend ways of putting things right. Causes are important, because they affect the recommendations one would make regarding appropriate help or support for the child. Many children with poor literacy skills are referred for assessment (whether privately or within the state system) on suspicion of dyslexia. Some are indeed found to have dyslexia, while others are often found to have been deprived of appropriate teaching (most commonly phonics tuition) or to lack the practice which is essential for basic literacy skills to become fluent. If the cause seems to be lack of appropriate teaching, simply recommending certain educational input may not be a sufficient remedy if the child’s teacher does not have the appropriate skills or the child’s school does not have the appropriate resources.

We could say initially that we need to assess children in order to monitor their level of achievement in literacy and we want to assess adults and adolescents because we need to see whether their achievement is sufficient to meet the requirements of whatever they are doing at work or in an educational setting. We assume here that assessment is in relation to the average literacy achievements of others of equivalent age. We might also implicitly assume that we are making the assessment in order to do something about it, if necessary.

Going deeper into this, we begin to get more controversial. Yes, we can monitor achievement in relation to others, but sometimes an assessor might say: ‘I’m afraid that this level of achievement is to be expected given overall cognitive ability.’ For example: ‘This child’s intelligence is predicting a level of performance in reading that is actually close to her actual level of reading.’ This does not mean that we expect reading skill to be necessarily anchored firmly to expectation. For instance, a child on a par with expectation could probably still enhance reading skills. There may even be reciprocity so that improving reading skills gradually feeds back to improve intelligence (see Stanovich, 1986).

Returning to the adult or the adolescent, our problem is not just about potential cognitive processes in relation to literacy, but also about what we can do about it to compensate within an examination setting. Some might feel that it is a little late to tackle the problem in the examination room by allowing extra time, when the roots of the problem might have been tackled earlier. Another example might be that the adult is referred by the employer as the job requires an upgrade. Is the individual capable of handling more paperwork, perhaps taking down telephone messages, or entering information into a computer database? If not, could training be provided to improve the necessary skills?

These are some of the problems that permeate how we make an assessment in reading – and in writing as well. Assessment has grown in importance, due partly to the development of better techniques for assessment and to a change in political climate that requires more accountability. Not just politicians but parents are developing an increasing awareness of the issues involved. At the same time there has been a lot of publicity about the decline in standards in literacy However, although there may be an impression of a decline, there is now little objective evidence for such a decline, simply because in the UK formal monitoring of standards was disbanded several years ago.

DYSLEXIA AND ASSOCIATED CONTROVERSIES

Many people still believe that it is not possible to make a reasonable assessment of reading until the child is about 8 years of age. The reason given is that children differ in their rate of progress in their early years and as they have only started to learn to read at about the age of 5 years, it is difficult to tell reliably whether children are seriously behind in their reading development. Nevertheless, testing is reliable enough for it to be possible to use the WORD reading and spelling tests in conjunction with the Wechsler Intelligence Scale for Children (WISC-III), to work out if there is a significant disparity between observed and expected WORD scores down to the age of 6 years. The likelihood of finding a discrepancy and its reliability in the early years is probably lower compared with later. However, a bright child who can hardly write anything coherently and who cannot read and understand simple sentences would probably be a good candidate for testing at the age of 6 or 7 years.

These tests are psychometric tests that cannot be undertaken by the teacher but only by an appropriately qualified psychologist. Most teachers at present would not make an assessment concerning whether the child is reading significantly below potential, only that reading performance was below the child’s peers. (If they desired they could use an open test of intelligence, a test that is unrestricted in use, and calculate centile or standard score discrepancy.) This is currently an important distinction in mode of operation between psychologist and teacher. It is also part of the controversy surrounding intelligence testing referred to earlier.

The basis of the reading discrepancy approach is to produce regression equations on large samples of reading and Intelligence Quotient (IQ) scores. This enables a prediction from the regression line of what the expected reading level would be for a given level of intelligence, allowing for chronological age. The important assumption is that there is a good linear relationship, or correlation, between the two variables. One area of debate has been to argue that the correlation with reading is less for fullscale IQ than for verbal IQ. Full-scale IQ involves the assessment of verbal and non-verbal processes or abilities in equal measure. It would therefore be better to base these regression equations on verbal IQ alone. One slight problem with this suggestion is that the mental arithmetic component of verbal IQ can be affected by poor verbal memory in the WISC-III. Mental arithmetic requires an involvement in verbal working memory, which can be problematic in dyslexia.

Another problem is a practical one. In a city like Leicester, for example, it would be difficult for the educational authority, responsible for more than 300 schools, to make a mass assessment of reading potential based solely on verbal intelligence as a substantial part of the community is of Indian, Pakistani, African and Afro-Caribbean origin. A proportion of these children are using English as a second language and testing their verbal intelligence on an English verbal intelligence test may be underestimating their intelligence. For example, Beech and Keys (in press) found that controlling for nonverbal IQ, there was a marked impairment in receptive oral vocabulary (a verbal IQ component) in a group of bilingual British Asian children of lower socioeconomic status aged 7-8 years compared with monolingual controls. This underestimate in verbal IQ would in turn reduce the proportion of such children who would be considered to be under-using their potential, because a lower predicted IQ in turn means a lower predicted potential in reading. An added complication is that the current data on the importance of phonology for early reading apply to an alphabetic language such as English in which letters approximately correspond to phonemes. This may not be so important in some other languages.

One alternative would be to produce special tests for each foreign language group, but this is also going to have the problem of being discriminatory. Clearly the issue has to be handled sensitively, but if one is too sensitive, many children may be forgoing appropriate skill learning.

If an education authority were to adopt the discrepancy approach there is the further problem that an application of an intelligence test, such as the WISC, has to be done individually by a qualified psychologist; this option is automatically ruled out for mass screening at present on the grounds of prohibitive cost.

The reading discrepancy approach can give rise to two different types of reading difficulty. First, there is the child with a specific reading difficulty, whose difficulties are confined to reading (and possibly other areas of literacy) but whose other skills and attainments are not significantly lower than expected. ‘This is often referred to as ‘dyslexia’.¹ In the context of reading, the term now generally refers to someone who from their full scale intelligence score is predicted to have a reading performance at a certain level, but who is statistically significantly behind this level of performance. Some authorities would argue that a ‘discrepant’ measure like this should be regarded as the defining characteristic of a specific reading difficulty, but not necessarily of dyslexia. There are many reasons why a child may be attaining in reading far lower than intelligence levels would predict. One reason might be that the child has not received appropriate teaching. If that were the sole reason for the child’s difficulties one would not want to class this as a case of dyslexia, which usually implies a constitutional cause of some kind. Other authorities have sought to qualify a discrepancy approach by requiring several additional criteria, or impediment& However, to restrict the concept of ‘dyslexia’ to children who show no other difficulties (such as hearing problems) contradicts the fundamental concept of dyslexia being a constitutional condition. The same criticism may be levelled at those who in the past have adopted a view of dyslexia that has resulted in the condition being observed mainly among children from middle-class homes (see Critchley, 1970). This was an acknowledgement that there may be circumstances at home that may be responsible for a lack of literacy development. This is not necessarily to do with class, of course; it is thus difficult to make an assessment of this kind of influence. Similarly, a child may be emotionally affected in some way, perhaps manifesting substantial behavioural problems.

Another criterion used by some authors is that there has to be a minimum level of intelligence. For instance, Vellutino (1979) advocated an IQ more than 90 (the average, adjusting for age, is 100); Critchley refers to ‘an adequate IQ’. However, given the use of the regression equation, one does come across children who have a lower IQ than 90, who still have a significant deficit in reading in relation to expectation. Realistically they should be referred to as having dyslexia. From the point of view of carrying out a research study on children with dyslexia one would normally use a cut-off point of something like 90. However, when testing an individual child for dyslexia, there should be no such limitation.

The second type of reading problem is often called ‘poor reading’, although in the past it has been referred to as ‘reading backwardness’. This is usually where the child’s intelligence is considerably below the norm. Using the same regression equation one would still be able to predict a certain level of reading performance. Because the child is at such a low level of IQ performance, the child is predicted to be low in reading as well. Thus this category of reader should be easily noticed in the classroom as reading performance would be poor on a standard reading test. The child with specific reading difficulties would not necessarily be noticed in the classroom, because the reading problem might be mild. However, in relation to potential it could be severe.

Do we treat the two types of children differently and should we treat them differently? The motivation of the teacher could be affected by this knowledge. One might think that learning is going to have to be at a slower pace for the poor reader. But this is not necessarily going to be the case. Yule et al. (1974) in a study of all the 9 and 10 year olds on the Isle of Wight found that children with specific reading difficulties actually made poorer progress. The main contrasts between the groups were that three-quarters of those with specific reading difficulties were boys and there were no organic disorders compared with 11% of the poor readers. There are interesting examples of children with low IQs who become very good readers, in the sense that their reading out loud becomes very accomplished.

Perhaps the major difference may be at the reading comprehension level. Brighter children with poor basic word reading skills often manage to compensate when reading text by making intelligent inferences about the gist of the text. They also use context more effectively, have a better knowledge of syntax and have a ‘world knowledge’ enabling them to make guesses that can compensate for specific word knowledge. By contrast, those who are less bright may have a better idea of word identification, but find it difficult to get to the appropriate meaning of the text. Another contrast is that those with specific reading difficulties tend to have problems more concentrated in terms of phonological difficulties (Jorm et al., 1986).

Some have argued against using the intelligence test altogether, especially on the grounds that it is culturally biased. As far as assessing reading comprehension is concerned, some have advocated the use of listening comprehension as a simpler method of measuring the disparity between the potential to comprehend and actual level of comprehension by means of reading print. This is discussed further in Chpater 9 on reading comprehension. Listening comprehension is obviously going to be useful to assess reading comprehension potential, but for tests of accuracy of reading print we are still left with full scale IQ as all that is available at present.

To be even handed, there are researchers (e.g. Siegel, 1992) who have argued that the discrepancy definition of reading problems is unnecessary as poor readers (whose reading is consistent with their IQ scores) are not differentiated from discrepant readers as both have problems in phonology and verbal memory and other aspects. (This is in contrast to Jorm et al. cited earlier (1986).) Another argument against the intelligence test is that it is increasingly being recognised that dyslexia is a constitutional condition that is generally inherited (but which may be due to birth difficulties). There is no reason why it should not occur among children of below average IQ. It could also (and probably will) co-occur with other disabilities (e.g. hearing impairment). The conventional definition of dyslexia rules out such children. Perhaps one day a definition of ‘genetic dyslexia’ will be devised.

The practical problem is that even if there were no difference in the difficulties of the two different types of problem readers, it does not preclude the importance of finding children who are underachieving. There could be large numbers of children who are discrepant in their reading and who are not being identified as having problems as such because their absolute level of reading and spelling performance is superficially indicating that nothing is seriously amiss. The situation at the moment is that such children are only coming to light if their parents make a fuss. Many LEAs do not explicitly tie specific learning difficulties to a statistically significant discrepancy between reading quotient and IQ, but to some arbitrary threshold (e.g. reading age two years behind chronological age). Often the child lags behind, but not enough to t...

Table of contents

• Cover Page
• Half-Title Page
• Title Page
• Copyright Page
• Table of Contents
• List of illustrations
• List of contributors
• Series editors’ preface
• 1 The psychological assessment of reading: theoretical issues and professional solutions
• 2 Assessment by classroom teachers
• 3 Assessment by educational psychologists
• 4 Screening early literacy
• 5 Assessment of visual problems in reading
• 6 Assessment of phonological skills and their role in the development of reading and spelling skills
• 7 Assessment of memory and reading
• 8 Assessment of affective and motivational aspects of reading
• 9 Assessment of comprehension in reading
• 10 Assessment of spelling and related aspects of written expression
• 11 Assessment of adult reading skills
• 12 Assessment of acquired disorders of reading
• 13 Computer-based assessment of reading
• 14 The legal aspects of the assessment of literacy and writing difficulties
• Test reviews
• Name index
• Subject index