Scientific rationale for LADS Plus
One might pose the question: since adults with dyslexia typically have problems with reading, writing and spelling, why nor simply measure those skills in order to identify the individuals who have dyslexia? The problems with this approach are (a) the differences in literacy skills between dyslexic and non-dyslexic persons are largely quantitative rather than qualitative, and (b) the development of compensatory strategies (especially by individuals with higher intelligence) often masks such differences. If one were to take a group of dyslexic and a group of non-dyslexic adults, statistically significant differences between these groups in reading, writing and spelling would almost certainly be found. But if one took an individual adult with dyslexia, he or she may have literacy skills in the average range (although below that which might be reasonably expected from their intelligence and education), while an individual adult who does not have dyslexia may have below average literacy skills. A screening test has to be capable of identifying which individuals do, and do not, have dyslexia to a reasonable degree of accuracy, and for that task measures that are more reliable than literacy skills are required.
The requirements of an effective screening system
The term ‘screening’ may be used for any procedure that facilitates rapid and easy sorting into categories where there is the expectation that some categorisation errors may occur. A certain level of error is traded for increased speed and practicality (see Singleton, 1997a). Today, the term screening is most commonly encountered in medicine where the purpose is to identify individuals at risk of some disorder or disease. In education the purpose of screening is usually to identify children or adults who have special educational needs (e.g. because of dyslexia). These individuals will require tuition or support over and above what is provided for other individuals, or may be entitled to special arrangements (reasonable adjustments, or accommodations) when being assessed or undergoing some selection process.
No screening system is 100% accurate since it is inherent within the technique of screening that a certain level of error or misclassification is inevitable. The accuracy of any screening system is indicated by the frequency of misclassifications that it makes. False positives and false negatives are the two types of screening misclassifications. A false positive is a ‘false alarm’, i.e. an instance in which a person has been classified as ‘at risk’ (in this case, possibly dyslexic) when, in fact, this is not the case. A false negative is a ‘miss’, i.e. an instance in which a person has been classified as ‘not at risk’ (in this case, not dyslexic) when, in fact, the opposite is the case (they are dyslexic). The value of any screening test depends on having low frequencies of false positives and false negatives. A high frequency of false positives generally results in the diverting of resources to dealing with cases that do not actually require attention. A high frequency of false negatives results in critical cases being overlooked, possibly with serious consequences. However, reducing misclassifications to very low levels almost always entails increasing the complexity and sophistication of the measurement, which will tend to reduce the practicality of screening and increase administrative effort and costs. In general, levels of less than 25% for each of these is advocated for effective screening (see Jansky, 1977; Potton, 1983; Singleton, 1997a).
Key indicators of dyslexia in adults
The compensation strategies that adults with dyslexia often develop in order to cope with literacy tasks can mask their dyslexia and make it difficult to determine whether a person genuinely does have dyslexia. For this reason, the use of literacy tasks alone are unreliable in the identification of dyslexia (see Singleton, 1999, for further discussion of this). Furthermore, any task that relies strongly on acquired knowledge (e.g. vocabulary) would not be suitable as results would be likely to be masked by educational factors. However, the cognitive difficulties that underpin dyslexia – particularly in phonological processing, lexical access and working memory – are much more fundamental and difficult to compensate. They are also less likely to be masked by educational factors. These underlying cognitive features of dyslexia have already been outlined in section 1.3. There is extensive evidence that, in adulthood, persons with dyslexia (even dyslexic university students) still exhibit limitations in tasks involving these cognitive domains (see Beaton, McDougall and Singleton, 1997; Gottardo et al, 1997; Hanley, 1997; Snowling, 2000; Snowling et al, 1997). In designing LADS Plus it was decided to build a system that had a strong scientific research basis and therefore to centre this on in phonological processing, lexical access and working memory. Three tests were devised, called Word Recognition, Word Construction and Working Memory, all of which rely to a greater or lesser extent on the core cognitive abilities, as shown in Table 1.
It has already been pointed out that there are alternative conceptualisations of dyslexia, which implicate dysfunctions in the visual system and/or in the cerebellar system for automatisation of skills (see section 1.3). There is insufficient space here to debate these complex issues (for a review, see Miles and Miles, 1999). However, it is undeniably the case that the vast majority of dyslexics display difficulties in phonological processing, lexical access and working memory. By contrast, at best only a much smaller proportion of dyslexics display visual processing difficulties (see Rack, 1997; Evans, 2001), and the extent of difficulties in cerebellar processing has yet to be firmly established. Hence, in developing LADS Plus it was considered prudent to focus on the core cognitive deficits. In addition, in view of the evidence on the role of speed of processing in dyslexia (see Wolf and O’Brien, 2001), it was decided to make the tasks in LADS Plus speeded ones to make them as sensitive as possible to dyslexic difficulties.
In Table 1, the number of crosses shows the extent to which the three dyslexia-sensitive tests in LADS Plus — Word Recognition, Word Construction and Working Memory — are believed to draw upon the core cognitive abilities, based on a consideration of the tasks involved. The conclusions shown in Table 1 are supported by intercorrelations computed between the three tests in a sample of 134 adults, with a 50:50 split between known dyslexics and non-dyslexics. The correlation between Word Recognition and Word Construction was found to be high (r=0.83), indicating that both these tests are assessing similar abilities; in fact, both are strongly phonological tasks. However, the correlations of these tests and the Working Memory test were much lower, although still statistically significant. This suggests that the two phonological tests also place demands on working memory, but a little more so in the case of Word Construction (r=0.34) than in the case of Word Recognition (r=0.30).4
Table 1. The cognitive basis of the three tests in LADS Plus.
4 Note that these correlations were computed on the basis of the full, original (unadaptive) versions of the tests, and are different from the data on the adaptive forms presented in Table 11.
The role of reasoning and intelligence
The decision to include tests of reasoning in LADS Plus was made for two reasons. First, to improve screening accuracy (as explained in Section 2.1) and s, second, to give Administrators additional information that would be helpful in interpreting results. The reasoning tests are not dyslexia-sensitive measures, but give a fairly good estimate of intelligence. Since very bright dyslexic adults tend to develop more proficient compensatory strategies this can make them more difficult to detect in a screening test of this nature. Conversely, individuals who are at the lower end of the ability range may struggle with some of the requirements of the dyslexiaIsensitive tests in LADS Plus, not because they have dyslexia, but because their vocabulary may be limited and because their overall speed of information processing may be relatively slow. In order to avoid excessive numbers of false negatives and false positives in such cases, the information provided by the reasoning measure enables Administrators to take this into account when considering results and making decisions on the most appropriate course of action.
Note that this does not mean that interpretation of LADS Plus results must necessarily conform to a discrepancy model of dyslexia.5 Such models, particularly when applied to identification of dyslexia in school-age children, have been criticised in recent years (see BPS, 1999). In 1999 a working group of the British Psychological Society’s Division of Educational and Child Psychology produced a report entitled: Dyslexia, Literacy and Psychological Assessment (BPS, 1999). The principal aim was to provide guidance for educational psychologists in the assessment of schoolchildren who are suspected of having dyslexia. A working definition that did not implicate any particular theoretical model of dyslexia was proposed:
‘Dyslexia is evident when accurate and fluent word reading and/or spelling develops very incompletely or with great difficulty. This focuses on literacy learning at the ‘word level’ and implies that the problem is severe and persistent despite appropriate learning opportunities. It provides the basis for a staged process of assessment through teaching.’ (BPS, 1999, p. 8)
It will be readily appreciated that this definition raise many critical issues (for discussion see Cook, 2001). For example: what does ‘very incompletely’ mean? What learning opportunities can be deemed to be ‘appropriate’? How much ‘additional effort’ is required? Taken at face value, it would not permit a diagnosis of dyslexia to be made until additional effort and/or instruction has been put in and that was not found to have brought about significant improvements. But we know that dyslexic children’s literacy skills can be improved by specialist skilled tuition (see Miles and Miles, 1999). Does this mean that these children are no longer dyslexic? Or were never dyslexic in the first place?
Outside of the UK school system, very few researchers endorse the BPS working group approach. Snowling, one of the foremost international authorities in the field states: ‘For clinical utility, the discrepancy approach needs to be supplemented by positive dyslexic markers that will allow practitioners to identify children who show early or residual signs of dyslexia that require intervention, and do not depend solely on the extent of the child’s reading problem.’ (Snowling 2000, p.25.) Many educational psychologists defend the use of the discrepancy approach (e.g. Ashton, 1996).
In the field of adult dyslexia, the BPS working group approach has sometimes been adopted in Further Education, especially where support staff have not had access to psychologists who could carry out diagnostic assessments. In such cases, dyslexia has often been regarded as being synonymous with ‘poor literacy skills’. On this basis, however, it is not possible to distinguish between:
a) Adults with a specific learning difficulty (i.e. dyslexia).
b) Adults with low general abilities.
c) Adults who are not in categories (a) or (b) but who have poor literacy skills as a result of inadequate educational experience (poor teaching, prolonged absence from school) or limited experience of English (e.g. immigrants).
Furthermore, the BPS working group’s approach will not work when it comes to assessing adults because the psychologist cannot rely on having adequate knowledge of an adult client’s educational history. Poor literacy alone is not an adequate criterion for identifying adult dyslexia. Many adults have compensated for their difficulties and consequently can score reasonably well on tests of single-word reading and spelling. In fact, it is difficult – if not impossible – to identify dyslexia reliably in bright adults without taking intelligence into account to some degree. Indeed, the fact that such individuals are typically perceived to be under performing in relation to expectations based on general ability, either in education or employment, indicates that some sort of discrepancy is critical to appreciating the nature of the problem. The use of statistically significant discrepancies when making diagnostic decisions enables psychologists to bring a measure of objectivity to what would otherwise be largely a subjective process. Hence, most authorities in the field advocate using a combination of discrepancy measure and positive dyslexia markers (e.g. in working memory or phonological processing) when identifying dyslexia, especially in adults (see Bartlett and Moody, 2000; Kirk, McLoughlin and Reid, 2001; McLoughlin, 1997; McLoughlin, Fitzgibbon and Young, 1994; McLoughlin, Leather and Stringer, 2002; Rack, 1997; Reid and Kirk, 2001; Singleton, 1999; Turner, 1997).
Hence LADS Plus is built on a model of dyslexia identification that takes into account both key dyslexia indicators and performance that is below expectations based on estimated intelligence.
5 A discrepancy model of dyslexia is based on the assumption that dyslexia can be identified where there is a statistically significant discrepancy between intelligence and attainment in literacy (strictly speaking, between actual literacy levels and literacy levels as predicted by age and intelligence).