the reading comprehension abilities of dyslexic students in higher education

15
ff The Reading Comprehension Abilities of Dyslexic Students in Higher Education Fiona Simmons and Chris Singleton* Department of Psychology, University of Hull, UK The reading comprehension abilities of a group of dyslexic university students were compared with those of non-dyslexic university students. A 655-word passage, followed by literal and inferential questions, was used to measure reading comprehension. The text was designed to be syntactically complex, yet place relatively modest demands on decoding skills. Although dyslexic students performed at a similar level to the non-dyslexic students on the literal questions, their performance on the inferential questions was poorer. An index of the participants’ ability to make inferences was calculated by subtracting the inferential question score from the literal question score. The groups differed significantly on this measure, indicating that the dyslexic students were specifically impaired in constructing inferences when processing complex text. It was concluded that dyslexic students in higher education have reading comprehension difficulties that cannot be accounted for by an inability to decode individual words in the text. The possible contribution that poor lexical automaticity and an impaired working memory make to this impairment is discussed. The implications for the assessment and support of dyslexic students are considered. Copyright © 2000 John Wiley & Sons, Ltd. Keywords: adults; developmental dyslexia; higher education; reading comprehension INTRODUCTION F or various reasons, students in higher education who have been identified as dyslexic may have single word reading abilities that are in the average range for the general population. These students may have compensated for or developed strategies to circumvent the difficulties * Correspondence to: Chris Singleton, Department of Psychology, University of Hull, Hull HU6 7RX, UK. Tel.: +44 1482 465157; fax: +44 1482 465599 e-mail: [email protected] DYSLEXIA 6: 178–192 (2000) Copyright © 2000 John Wiley & Sons, Ltd.

Upload: fiona-simmons

Post on 06-Jun-2016

218 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The reading comprehension abilities of dyslexic students in higher education

ff

The Reading ComprehensionAbilities of Dyslexic Studentsin Higher EducationFiona Simmons and Chris Singleton*Department of Psychology, University of Hull, UK

The reading comprehension abilities of a group of dyslexicuniversity students were compared with those of non-dyslexicuniversity students. A 655-word passage, followed by literal andinferential questions, was used to measure readingcomprehension. The text was designed to be syntacticallycomplex, yet place relatively modest demands on decodingskills. Although dyslexic students performed at a similar level tothe non-dyslexic students on the literal questions, theirperformance on the inferential questions was poorer. An indexof the participants’ ability to make inferences was calculated bysubtracting the inferential question score from the literalquestion score. The groups differed significantly on thismeasure, indicating that the dyslexic students were specificallyimpaired in constructing inferences when processing complextext. It was concluded that dyslexic students in higher educationhave reading comprehension difficulties that cannot beaccounted for by an inability to decode individual words in thetext. The possible contribution that poor lexical automaticity andan impaired working memory make to this impairment isdiscussed. The implications for the assessment and support ofdyslexic students are considered. Copyright © 2000 John Wiley& Sons, Ltd.

Keywords: adults; developmental dyslexia; higher education; reading comprehension

INTRODUCTION

For various reasons, students in higher education who have beenidentified as dyslexic may have single word reading abilities that arein the average range for the general population. These students may

have compensated for or developed strategies to circumvent the difficulties

* Correspondence to: Chris Singleton, Department of Psychology, University of Hull, HullHU6 7RX, UK. Tel.: +44 1482 465157; fax: +44 1482 465599e-mail: [email protected]

DYSLEXIA 6: 178–192 (2000)Copyright © 2000 John Wiley & Sons, Ltd.

Page 2: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 179

they experienced earlier in life (McLoughlin, Fitzgibbon and Young, 1994;McLoughlin, 1997). At this stage in their education, the most salient featuresof their dyslexia may be apparent in other areas, e.g. taking lecture notes,writing essays, revising for examinations or comprehending large quantitiesof complex text (Gilroy and Miles, 1996; Riddock, Farmer and Sterling, 1997;Singleton, 1999).

This picture is difficult to reconcile with current theoretical views ofdyslexia. In recent years, dyslexia has generally been considered to be adisorder which primarily affects single word reading rather than compre-hension. For example, the research committee of the Orton Dyslexia Society(subsequently the International Dyslexia Association) defined dyslexia as ‘aspecific language-based disorder of constitutional origin characterized bydifficulties in single-word decoding, usually reflecting insufficient phonologicalprocessing abilities’ (Orton Dyslexia Society, 1994; our italics).

Finucci et al. (1984) compared a group of adults who were all geneticallyrelated to dyslexic children with a control group. The groups did not differin reading comprehension ability. Finucci et al. (1984) concluded that silentreading comprehension was ‘neither necessary or sufficient for specificreading difficulties to be present’ (p. 145).

Nevertheless, there is now a growing recognition of individual variationin the strengths and limitations displayed by different adult dyslexics, aswell as changes in performance as dyslexics grow older and progressthrough education (McLoughlin, Fitzgibbon and Young, 1994; Beaton, Mc-Dougall and Singleton, 1997b; Everatt, 1997; McLoughlin, 1997; Riddock,Farmer and Sterling, 1997; Singleton, 1999). Rack (1997) considered theprofiles of abilities and difficulties of 28 adults who were assessed fordyslexia at the Dyslexia Institute. He identified three major sub-types:developmental phonological difficulties, phonological/visual motor co-ordi-nation difficulties, and visual motor co-ordination difficulties. Each groupdisplayed different characteristics. Everatt (1997), comparing adult dyslexicstudents with a control group, found that the dyslexics showed a similarlevel of performance on tests of single word reading rate, but poorerperformance on spelling and reading comprehension tests. These resultssuggest that the underlying cognitive deficits can affect different behaviourslater in life.

The problems of dyslexic undergraduate students have received particularattention of late (see, for example, Gilroy and Miles, 1996; Beaton, McDou-gall and Singleton, 1997a; Riddock, Farmer and Sterling, 1997; Singleton,1999). The increasing number of students coming forward for assessmentwas highlighted in the report of the National Working Party on Dyslexia inHigher Education (Singleton, 1999). Singleton (1995, 1996) and Beech andSingleton (1997) have expressed concern over the different assessment meth-ods and criteria used by universities to identify dyslexic students. Asdiagnosis leads to academic and financial support, it is important to beconsistent and accurate. The National Working Party report, in part, at-tempts to promote greater consistency and objectivity of assessment meth-ods used in higher education.

Previous studies of the reading comprehension abilities of adult dyslexicshave been impaired by the inappropriateness of the available assessment

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 3: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton180

tools. Researchers have often resorted to the use of children’s tests because ofthe lack of a standardized test suitable for the adult British population. Anartificial time limit is sometimes put on children’s tests when they are usedwith adults. This improves discriminability, but reduces validity. The resultswould be confounded by reading speed; candidates may perform poorly asa result of slow reading, rather than limitations in reading comprehensionability.

Reading comprehension tests designed for children also have poor ecolog-ical validity when used with adults. They consist predominately of shortexcerpts from children’s stories. The themes and ideas that are describedhave little relevance to the lives of adults (McLoughlin, 1997), and manyaspects of skilled reading cannot be assessed: the short length of the passagerestricts the reading skills that are required. When adults are required toapply their reading comprehension skills in the real world (particularly inhigher education), they are usually dealing with much longer passages.

Oakhill and Garnham (1988) have reviewed text-level reading comprehen-sion skills. To understand sentences, one must be able to syntacticallyprocess them (identify the linguistic structures), and then infer the meaningfrom the linguistic and non-linguistic context. One must then integrate thesentence meanings to produce a coherent picture of the situation theydescribe. Bransford, Barclay and Franks (1972) identified the need to makeboth integrative (combining information from different parts of the text) andconstructive (inferring information that is not explicit in the text) inferences,in order to build a mental model of the information presented. Very shortpassages limit the amount of time one has to hold information in workingmemory in order to make inferences. A longer text will, therefore, be a moresensitive instrument for assessing difficulties caused by an impaired work-ing memory capacity.

Metacognitive reading skills, which are an integral part of skilled adultreading (McLoughlin, 1997), are not well addressed by short passages. If ashort passage has to be comprehended, most of its content will be recalledafter a single reading, and the aspects that have been forgotten can beaccessed fairly efficiently by re-reading the passage. However, a longerpassage will not be memorized during the first reading. A skilled reader willselect an appropriate strategy to access the information required to answerthe questions, e.g. locating an item of factual information by skimming thetext for the relevant area rather than reading every word.

Finucci et al. (1984) used the Gates–MacGinitie Reading Test, Level F(MacGinitie, 1978) to compare the reading comprehension abilities of acontrol group with the adult genetic relatives of children with dyslexia. Nosignificant differences in reading comprehension ability were found betweenthe two groups. Everatt (1997) found that the dyslexic undergraduatesperformed significantly worse than controls on the GAPADOL Reading Com-prehension Test (MacLeod and Anderson, 1973), but differences found usingthe Neale Analysis of Reading (Neale, 1989) were not significant. Gottardo,Siegel and Stanovich (1997) found adults with specific reading disabilityscored significantly lower on the Nelson–Denny Reading Test (Brown et al.,1985) than controls. All these tests were designed for students under the ageof 18. The Neale Analysis of Reading is particularly inappropriate, as it was

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 4: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 181

designed for 6- to 12-year-old children. Further research is required to clarifythese conflicting findings. However, the underlying difficulties that typicallycharacterize dyslexia suggest that dyslexic individuals would find readingcomprehension difficult. There are a number of possible reasons for this.

(i) Lexical processes not becoming automatic. Perfetti (1985) suggested thatlexical processes do not become automatic in dyslexics. Nicolson andFawcett (1990, 1995) have also hypothesized a generic automatizingdeficit in dyslexic individuals. Consequently, as dyslexics are likely tohave to rely more on active processes to recognize the words, this willreduce the cognitive workspace available for any additional cognitiveactivity, including comprehension.

(ii) Working memory deficiencies. Reading comprehension requires the use ofworking memory (Oakhill and Garnham, 1988). Whilst decoding text,one must integrate knowledge about the real world with informationderived from text already processed to create and progressively modifya mental model. Baddeley et al. (1986) found that adults’ workingmemory capacity was a significant predictor of reading comprehensionability. Studies have found that dyslexic adults have impaired workingmemories (e.g. Beech, 1997; Hanley, 1997; Nicolson and Fawcett, 1997;Palmer, 2000). It would, therefore, be logical to suggest that this impairstheir performance on reading comprehension tasks.

(iii) Other difficulties. Oakhill and Cain (1997) found that listening compre-hension, inference-making skill and knowledge about text structure allpredicted specific comprehension deficits in children. Baddeley et al.(1986) reported that adults’ comprehension ability was predicted bytheir performance on vocabulary tests, lexical decision tasks with non-homophonous words and a letter–name matching task. It is possiblethat dyslexics are impaired on any one of these skills, either because ofan innate physiological defect, or because of some secondary environ-mental factor (e.g. lack of exposure to text as a result of poor readingabilities or lack of enjoyment in reading). Gottardo, Siegel and Stanovich(1997) found some evidence that supports this hypothesis. Adults witha specific reading difficulty performed significantly worse on tests ofvocabulary and listening comprehension.

It is quite likely that more than one of these problems affects comprehen-sion skills. For example, lexical processes may not be automatic and, there-fore, reading speed is slow. This would place further strain on an alreadylimited working memory. Whether dyslexia is viewed primarily as a phono-logical difficulty (Snowling et al., 1997), as an impairment in the workingmemory resulting from an impoverished phonological store (McLoughlin,Fitzgibbon and Young, 1994), or as a result of both of these factors (Palmer,2000), one would still predict reading comprehension difficulties.

In the light of the above criticisms, it was decided that standardized testsdesigned for children would not be used in this study. Instead, a new taskwas constructed to discriminate between university students according totheir reading comprehension skills. This consisted of a passage of text,followed by inferential and literal multiple-choice questions. The literalquestions only required information that was stated explicitly in the text.

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 5: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton182

The inferential questions required the participant either to integrate morethan one piece of information or use their knowledge about the world tointerpret an ambiguous statement. The inferential questions were not de-signed to put additional strain on decoding skills. Therefore, relative profi-ciency on the literal questions, and difficulty with the inferential questions,would indicate a specific difficulty in inference formation. The structure ofthe task in this study was based on an approach used by Oakhill (1984)designed to compare the inference-making skills of children with good andspecifically poor reading comprehension abilities.

The study was designed to assess whether dyslexic students in highereducation experience reading comprehension difficulties with textual mate-rial appropriate for their age and intellectual level. It allowed considerationof the contribution that poor decoding ability made to comprehensiondifficulties. It was hypothesized that a cognitive deficit impairs the readingcomprehension abilities of dyslexic students at higher education level. Thishypothesis would be supported if the dyslexic participants score signifi-cantly worse on inferential questions, but similarly on literal questions,indicating that the comprehension deficit is not a direct result of poordecoding skills.

METHOD

Participants

The participants comprised ten dyslexic adults (five male, five female; meanage 27.6 years, standard deviation (S.D.) 10.22 years; mean single wordreading standard score 93.00, S.D. 12.77), and ten non-dyslexic adults (twomale, eight female; mean age 21.4 years, S.D. 3.53; mean single word readingstandard score 110.10, S.D. 8.23). Nineteen of the participants were currentundergraduate students of the University of Hull. The remaining individual(in the dyslexic group) had graduated from the University in the last 3 years.The dyslexic participants’ mean single word score, although within theaverage range, was significantly lower than the controls (F(1, 18)=10.96,p=0.004).

Materials

Reading comprehension was measured using a specially designed task,which consisted of a passage of text followed by ten multiple-choice ques-tions (five literal questions and five inferential questions). The two types ofquestions were presented alternately.

The text, which was 655 words in length, described the modernization ofa factory. A passage of this length would require exercise of metacognitiveskills in order to answer the questions efficiently. It was recognized that thetext could be made difficult to read by the inclusion of low frequency andlong words. However, this strategy was avoided, and instead, the textcomplexity was increased by the use of long sentences with syntacticallydifficult structures. These factors have been found to reduce reading speed

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 6: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 183

because they make text more difficult to process (Fodor, Bever and Garrett,1974). The readability of the text was calculated using Flesch’s reading easescore and found to be 59.2 (fairly difficult) on a scale of 100 (very difficult)to 0 (very easy) (Gilliland, 1972). Passages at this level are regarded as beingsuitable for individuals at college level, or who have had several years ofsecondary school education. The WRAT-3 Single Word Reading Test (Wilkin-son, 1993) was used to measure decoding skills.

Procedure

Dyslexia assessment and support staff of the University of Hull contactedparticipants who had been diagnosed with dyslexia and asked if they wouldparticipate. The diagnostic procedure at Hull University conforms to theguidelines recommended in the report of the National Working Party onDyslexia in Higher Education (Singleton, 1999). Volunteers were recruitedfrom the undergraduate population to act as controls; these were askedwhether they had ever been diagnosed with dyslexia before the study began,and any who indicated they had a diagnosis were excluded from the controlgroup.

The task was administered individually, in a quiet room. Each participantwas first asked to read the text through carefully. They were then asked toread and complete the multiple-choice questions (by ticking the appropriateboxes), having access to the text throughout. The time they took to read thetext and the time they took to answer the questions were measured. Theparticipants were informed that there was no limit on the time they couldtake to read the text or answer the questions. After they had completed thetask, they were asked to estimate the number of questions they believed theyhad got correct.

RESULTS

Participants’ scores are shown in Table 1. Both groups found the inferentialquestions more difficult. This difference is confirmed by a one-way analysisof variance (ANOVA) comparing question type (F(1, 18)=37.33, pB0.001).On literal questions, there was little difference between the mean scores ofthe two groups, the mean score for the dyslexic group being slightly higher.A one-way analysis of covariance (ANCOVA), in which age and sex werecontrolled, confirmed that the difference between the two groups was notsignificant (F(1, 3, 16)=0.421, p=0.74). However, the dyslexic groupachieved lower scores on the inferential questions. A one-way ANCOVA, inwhich age and sex were controlled, indicated that the difference between thetwo groups was significant (F(1, 3, 16)=6.44, p=0.02).

The difference between participants’ literal and inferential scores was usedas an index of their specific difficulty in making inferences when processingcomplex text. The larger the difference, the greater their difficulty. Thedifference exceeds 1 in 70% of the dyslexic sample, compared with 30% ofthe non-dyslexic sample. The mean difference for the dyslexic group was1.80 (S.D.=0.92), compared with 0.90 (S.D.=0.86) for the non-dyslexic

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 7: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton184

Table 1. The scores of the non-dyslexic and dyslexic participants

Non-dyslexic (n=10) Dyslexic (n=10)

Inferential Difference LiteralLiteral Inferential Differencescore scorescore score

1 4 2 23 25 4 1 4 2 2

1 3 2 14 31 2 03 242 4 4 05 30 4 13 33

1 2 5 4 13−1 5 35 24

3 1 4 1 341 4 2 23 2

0.90 3.90 2.10 1.292.90Mean 3.80S.D. 0.79 1.10 0.88 0.86 1.80 0.92

group. The dyslexic participants’ specific difficulty in formulating inferencesis confirmed by the results of a comparison of the difference indices of the twogroups using a two-way mixed analysis of covariance (MANCOVA) in whichage and sex were controlled (F(1, 3, 16)=5.03, p=0.038).

Table 2 shows the mean times for reading the text and working on thequestions for both groups. A one-way ANCOVA (controlling for age and sex)indicated that, at best, the difference between the reading times of the dyslexicand non-dyslexic participants was only marginally significant (F(1, 18)=4.12,p=0.059). Dyslexic participants took longer to answer the questions. Aone-way ANCOVA (controlling for age and sex) indicated that this differencewas significant when one extreme outlier was excluded from the analysis (F(1,18)=4.86, p=0.044). The outlier was excluded because this participant tooklonger than 10 min to complete the task, which was more than 2 S.D.s abovethe mean. After this exclusion, the mean working time for the dyslexic groupwas 293.30 s (S.D. 117.68 s), for the non-dyslexic group the mean was 191.0s (S.D. 62.13 s). Excluding the same individual from the reading time analysisdid not reduce the probability level below the critical value.

Table 2. The mean times and estimates of the non-dyslexic anddyslexic participants

Non-dyslexic Dyslexic

Mean S.D. Mean S.D.

Time (s)59.88188.20 274.10Reading 130.18

245.40 182.06Working 293.30 117.67

Estimating accuracyActual score 2.006.001.706.70

6.802.04 1.757.20Estimate score2.490.803.410.50Difference

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 8: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 185

An index of participants’ estimating accuracy was calculated by subtract-ing their actual score from their estimate (see Table 2). Both dyslexic andnon-dyslexic participants slightly over-estimated their scores. A one-wayANOVA indicated that the accuracy levels of the two groups did not differsignificantly (F(1, 18)=0.051, p=0.83).

No statistical relationship was found between literal question score andsingle word reading score. The literal questions can be answered, even bythe dyslexic participants, with relatively low single word reading scores. APearson correlation analysis confirmed the lack of a significant relationshipbetween the two scores (r=0.17, p=0.47). However, there was a moderatepositive relationship between single word reading score and inferentialquestion score. A Pearson correlation analysis confirmed the significance ofthis relationship (r=0.49, p=0.03).

DISCUSSION

Dyslexic individuals performed at a comparable level with non-dyslexicindividuals on the literal questions in this reading comprehension task, butsignificantly poorer on the inferential questions. This suggests that theycould decode the words within the passage, but were specifically impairedin making inferences. As there was no time limit on the task, this effect wasnot an artefact produced by artificial time constraints.

These findings confirm those of Everatt (1997) and Gottardo, Siegel andStanovich (1997), who found differences between adults with dyslexia (orspecific reading difficulty) and their controls on reading comprehensiontests. The contradictory results found by Finucci et al. (1984) may have beena result of their use of an experimental group consisting of the geneticrelatives of dyslexics, rather than individuals who had been diagnosed ashaving dyslexia. Alternatively, they may have used less challenging testingmaterials which were insensitive to reading comprehension differences.

An impaired working memory could contribute to difficulties in readingcomprehension. Daneman (1987) asserted that function and structure arerelated in working memory. The system itself is viewed as having a limitedcapacity, with the storage space being limited by the amount of resourcesprocessing requires. It is argued that individual differences in structuralstorage capacity do not explain the variation in reading comprehensionability, because variation is small in the general population, but processingefficiency differences do. In support of this argument, Daneman (1987)highlights the finding that storage capacity differences measured usingvarious digit span tasks have not been successful in predicting readingcomprehension differences (e.g. Perfetti and Goldman, 1976). Her own study(Daneman and Tardif, 1987) found that a language-specific working memorymeasure predicted reading comprehension scores, but spatial and numericalmeasures of working memory did not. These findings contrast with thepredominant componential view that decoding and comprehension are inde-pendent processes (Gough and Tunmer, 1986). Instead, Daneman (1987) seesreading comprehension and decoding abilities as connected. Inefficient de-coders will have less working memory space available for comprehension

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 9: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton186

processing than efficient decoders, even if they have the same structuralstorage capacity. In Daneman’s view, Perfetti’s (1985) theory of lack of lexicalautomacity and Baddeley et al.’s (1986) working memory model are notcompeting hypotheses, but are intrinsically linked.

The significant relationship between single word reading score and infer-ential question score suggests that individuals with poorer decoding skillshave impaired reading comprehension. The relationship is unlikely to bedirect, as the dyslexic participants’ success on the literal questions indicatedthat they could decode the words within the passage. If the dyslexic individ-uals devote more resources to this activity, Daneman’s (1987) model wouldsuggest that their comprehension will suffer because the total capacity ofworking memory is finite. Previous research has indicated that even whendyslexic individuals can read words, the process is less automatic. Inspeeded trials in which reading had to be automatic, Yap and van der Leij(1994) demonstrated that dyslexic children performed more poorly than theirreading age matched controls. If the structural capacity of dyslexics’ workingmemory is already limited, as several studies have suggested (e.g. Hanley,1997; Nicolson and Fawcett, 1997), the effect of these automaticity problemson reading comprehension will be compounded.

It is important to note that the reading comprehension difficulties ofdyslexic participants may not be apparent when they are tackling shortertexts. Longer passages will require the information that is necessary for theformation of integrative inferences to be held for longer in working memory.Although it is the norm for lengthy texts to be used for study in highereducation, standardized tests of reading comprehension often demand com-prehension of relatively short passages. Hence, the materials used in thepresent study arguably confer greater ecological validity than that of studieswhich have relied on less appropriate measures (e.g. Finucci et al., 1984;Everatt, 1997; Gottardo, Siegel and Stanovich, 1997).

Although an impaired working memory and poor decoding automaticityappear to be plausible explanations for the reading comprehension deficitsfound in these dyslexic adults with dyslexia, it is also possible that dyslexicssuffer from an additional problem, such as a limited vocabulary or a generalimpairment in comprehending verbal information as a result of limitedexposure to complex texts. These possibilities can be investigated in furtherstudies by examining the relationship between vocabulary and listeningcomprehension, and by comparing the listening and reading comprehensionskills of dyslexic adults.

The dyslexic participants took longer to complete the task. This may havebeen because they made more frequent references to the text to support theirimpaired working memory. However, other explanations are possible. Theirmetacognitive reading skills may be less well developed, resulting in inap-propriate strategy choice. Alternatively, they may not have read the text ascarefully as the non-dyslexic individuals (their reading times were notsignificantly longer than the control group) and, therefore, may have neededto make more references back to the text in order to compensate. Furtherresearch could address these issues more directly. The multiple-choice ques-tions could be read to the candidate (thus, controlling for the effect ofreading speed on working time), and the number of times they referred tothe text noted.

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 10: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 187

The current study has limitations, and it needs to be replicated withadditional factors controlled in order to confirm the results. All participantswere either students currently in higher education or recent graduates, andwere, therefore, believed to be of good average or above average intelli-gence. However, pairwise matching of age, sex and intelligence would addfurther weight to the results. In this study, the use of literate dyslexicindividuals (all with single word reading scores within the normal range forthe general population), the avoidance of long and obscure words in the textand the lack of significant differences on the literal questions would allsuggest that the reading comprehension difficulties of dyslexic participantswere not the direct result of an inability to decode single words. To be surethat such problems did not exist, one could ask participants to read thepassage aloud after testing was completed. Additional measures of thestructural capacity of working memory (e.g. digit span) could also beadministered. A negative relationship between inferential question score anddigit span would support the hypothesis that dyslexic individuals’ readingcomprehension difficulties are compounded by a limited structural workingmemory capacity.

It is clear from these results that despite having single word reading scoresin the average range for the general population, the dyslexic students in thisstudy would still have problems with the academic demands of a degreecourse. The effects of dyslexia are not eliminated when decoding skills areboosted by intervention and practice. The extra processing effort dyslexicindividuals devote to decoding draws resources away from other tasks. Theytook longer to complete the experimental task. This suggests that recommen-dations for extra time in examinations and assignments are justified (see thereport of the National Working Party on Dyslexia in Higher Education, Single-ton, 1999).

Several studies have shown that metacognitive strategies are beneficial foryounger readers who have comprehension difficulties (e.g. Ekwall andEkwall, 1989; Tregaskes and Daines, 1989). The emphasis of these ap-proaches is on generating questions while reading, re-reading where neces-sary to achieve satisfactory comprehension, explictly relating newinformation to existing knowledge, and converting new information tomental images. The findings of the present study suggest that that trainingin metacognitive strategies for reading comprehension would also be appro-priate for dyslexic students at university level. First, dyslexics and non-dyslexics in the study did not differ significantly in their ability to estimatetheir reading comprehension scores. This suggests that dyslexic studentshave some metacognitive awareness of their limitations in reading compre-hension. Second, although the dyslexic participants took longer to completethe task, the groups were not found to differ significantly on initial timetaken to read to the text. Hence, it is plausible that the dyslexic studentswere either applying inappropriate strategies for assimilating the text, and/or having to re-read the text more often (because of lower retention ofinformation contained in the text).

An example of the metacognitive approach is the PQ4R (p6 review,q6 uestion, r6 ead, r6 eflect, r6 ecite, r6 eview) technique described by Cheek andCheek (1983) and advocated by McLoughlin, Fitzgibbon and Young (1994)

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 11: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton188

for use by adults with dyslexia. This method, which was developed from themore familiar SQ3R (survey, question, read, review, revise) technique, isdesigned to promote active reflection on the text being read, as well asencouraging re-reading any difficult sections. The adoption of a metacogni-tive technique such as this would, therefore, be expected to help to combatthe reading comprehension difficulties encountered by dyslexic students,especially when dealing with lengthy and conceptually complex texts.

APPENDIX A. THE READING COMPREHENSION TASK

The Craigforth factory

The Craigforth works was the largest manufacturer of copper telecommuni-cations cables in Europe. It was built at the turn of the century by a Victorianbusinessman called Michael Phillips. In the middle of this century, it wasbought by a large multinational corporation. The corporation’s head office issituated in Houston, Texas. Nearly all of the employees at the factory belongto a union; the ARA represents the cable makers, whilst the BSB representsthose in clerical and technical roles. The relationship between the manage-ment and the unions is fairly hostile.

Recently, there have been significant redundancies, over the past tenyears, the number of people working in the factory has been reduced fromover two thousand to less than three hundred. Both external conditions andinefficient working practices have lead to this situation. The copper telecom-munication cables manufactured at the site are less in demand, as mostexchanges now use optical fibres. The price of copper has increased dramat-ically. To continue business, the plant must renew its lucrative contract withthe telephone company, however, it faces stiff competition from two otherfirms. There have also been attempts to try to expand the export market,which at present is very limited.

The factory’s work-force is old; over half the employees are aged overfifty. There is a high incidence of absenteeism, long term ill health problemsaccount for a significant proportion of this. The length of service of many ofthe staff means they have difficulty accepting change. They have done thesame things in the same ways for decades. The modernization initiatives thatwere designed to address some of the factory’s inefficient working practiceshave had to be carefully negotiated with the unions. Many of the changeswere proposed by a young team of crisis management consultants sent fromHead Office. As they were outsiders, their ideas were viewed with suspicion.

There was also a drive to create a more flexible labour force. The cablemaking process consists of three stages; core (where the copper is stretchedand coated in plastic), twin (where the wires are twisted together) andsheath (where the wires are coated in rubber). Before the modernizationinitiatives were implemented, the operatives were only trained in the skillsnecessary for one stage, a programme is now in place to equip the em-ployees with the skills necessary for all three stages.

The staffing structure has been changed. The redundancies included asubstantial number of staff in white collar roles. Many middle management

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 12: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 189

jobs were viewed as unnecessary, consequently, a flatter management struc-ture was introduced. The whole of the payroll department was scrapped.The old system of a weekly cash payment has been replaced by monthlysalary, which is transferred straight into the employee’s bank accounts. Thisnew arrangement is administered by head office. A significant number ofother clerical tasks can now be done by computers.

On the shop floor, a team-based approach to work has been implemented,supervisors and charge hands have been replaced by team leaders. Much ofthe work previously done by middle managers has been transferred to theteam leaders. Some of the old supervisors who were recruited into these newroles have had difficulty coping with the additional work. They are receivingadditional training in computing and people management skills.

To encourage the new teams to work well together, they have been sentaway with the senior management team on various outdoor activity week-ends. They completed a number of tasks designed to improve group work-ing and problem solving skills. Some managers thought that the weekendswere a waste of time, they felt that in such desperate times, the companyshould not be paying for ‘holidays’. Union representatives also had reserva-tions about the whole team working approach, they were concerned that itwould simply mean fewer people having to do more work. However, therewas little risk of industrial action, the employees were frightened by therecent redundancies. (655 words.)

Multiple-choice questions (with correct answers marked * and requiredinferences)

1. Why is there little risk of industrial action?a) Pay and conditions are excellentb) The unions agree totally with the modernization programmec) The employees are worried they may lose their jobs*d) The management have an excellent relationship with the unionse) The employees are frightened of the managers responseInference: The employees found the recent redundancies frightening because itsuggests that their jobs are not secure.

2. What initiatives did the company implement to improve efficiency?a) Introduced team working *b) Bought cheaper copperc) Asked employees to work longer hoursd) Banned union membershipe) Reduced wages

3. Who have been directly affected by the new team-working approach?a) Employees in clerical rolesb) Employees in middle management rolesc) Members of the BSBd) Employees in the payroll departmente) Members of the ARA*

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 13: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton190

Inference: Integrate the fact that team working was introduced on the shopfloor andtherefore affected cable makers who are members of the ARA.

4. What happens at the sheath stage?a) The copper is stretchedb) The copper is coated in plasticc) The copper is coated in rubber*d) The wires are twisted togethere) The copper is melted

5. What reservations did some of the management have about the modern-ization initiatives?a) They believed the weekends away were a waste of money*b) They felt the new teamleaders would not be effectivec) They were unhappy about the scraping of the payroll departmentd) They thought team working would mean less people doing more worke) They felt it would be impossible to convince older workersInference: infer that the phrase, ‘should not be paying for ‘holidays’’ means that theweekends are a waste of money.

6. What conditions have contributed to the large number of redundancies?a) A fall in copper pricesb) A reduced market for copper cables*c) Industrial unrestd) A reduced market for optical fibrese) Outdated machinery

7. Who has received additional training?a) The managersb) Employees in technical rolesc) The clerical workersd) The operatives*e) The union representativesInference: infer that the operatives working on the shopfloor are receiving training toallow them to work on more than one stage.

8. How has the staffing structure changed?a) The number of people in technical roles was reducedb) The number of people in middle management roles was reduced*c) The number of people in clerical roles was increasedd) The number of people in technical roles was increasede) People from Head Office were moved into senior management roles

9. What factors made the changes difficult to introduce?a) They were suggested by consultants from Head Office*b) They were expensivec) The workforce was inexperiencedd) They did not receive full support from Head Officee) The management were incompetentInference: infer that because the staff were suspicious of the head office consultantsthat the changes were difficult to implement.

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 14: The reading comprehension abilities of dyslexic students in higher education

Reading Comprehension of Dyslexic Students 191

10. Which of the following statements describe the Craigforth factory?a) It is the largest manufacturer of copper telecommunications cables in

Europeb) It has recently escaped significant redundanciesc) It exports a large proportion of its outputd) It is heavily unionized*e) It has good union management relations

References

Baddeley, A., Logie, R., Nimmo-Smith, I. and Brereton, N. (1985) Components of fluentreading. Journal of Memory and Language, 24(1), 119–131.

Beaton, A., McDougall, S. and Singleton, C.H. (Eds) (1997a) Dyslexia in Literate Adults(special issue of the Journal of Research in Reading, 20(1)). Oxford: Blackwell.

Beaton, A., McDougall, S. and Singleton, C.H. (1997b) Humpty Dumpty grows up?Diagnosing dyslexia in adulthood. Journal of Research in Reading, 20(1), 13–21.

Beech, J.R. (1997) Assessment of memory and reading. In J.R. Beech and C.H. Singleton(Eds), The Psychological Assessment of Reading, pp. 143–159. London: Routledge.

Beech, J.R. and Singleton, C.H. (1997) The psychological assessment of reading: theoreticalissues and professional solutions. In J.R. Beech and C.H. Singleton (Eds), The PsychologicalAssessment of Reading, pp. 1–26. London: Routledge.

Bransford, J.D., Barclay, J.R. and Franks J.J. (1972) Sentence memory: a constructive versusinterpretative approach. Cognitive Psychology, 3(2), 193–209.

Brown, J.I., Bereton, Bennett, J.M. and Hanna, G. (1985) The Nelson–Denny Reading Test.Chicago, IL: Riverside Publishing Company.

Cheek, E.H. and Cheek, M.C. (1983) Reading Instruction Through Content Teaching. Columbus,OH: Merill.

Daneman, M. (1987) Reading and working memory. In J.R. Beech and A.M. Colley (Eds),Cognitive Approaches to Reading. Chichester: Wiley, pp. 491–508.

Daneman, M. and Tardif, T. (1987) Working memory and reading skill re-examined. In M.Coltheart (Ed.), Attention and Performance 12: The Psychology of Reading, pp. 491–508. Hove:Erlbaum Associates.

Ekwall, E. and Ekwall, C. (1989) Using metacognitive techniques for the improvement ofreading comprehension. Journal of Reading Education, 14(3), 6–12.

Everatt, J. (1997) The abilities and disabilities associated with adult developmental dyslexia.Journal of Research in Reading, 20(1), 13–21.

Finucci, J.M., Whitehouse, C.C., Issacs, S.D. and Childs, B. (1984) Derivation and validationof a quantitative definition of specific reading disability for adults. Developmental Medicineand Child Neurology, 26(1), 143–153.

Fodor, J.A., Bever, T.G. and Garrett, M.I. (1974) The Psychology of Language: An Introduction toPsycholinguistics and Generative Grammar. New York: McGraw-Hill.

Gilliland, J. (1972) Readability. London: University of London Press.

Gilroy, D.E. and Miles, T.R. (1996) Dyslexia at College, 2nd Edn. London: Routledge.

Gough, P.B. and Tunmer, W.E. (1986) Decoding, reading and reading disability. Remedialand Special Education, 7(1), 6–10.

Gottardo, A., Siegel L.S. and Stanovich, K.E. (1997) The assessment of adults with readingdifficulties: What can we learn from experimental tasks? Journal of Research in Reading, 20(1),42–54.

Hanley, J.R. (1997) Reading and spelling impairments in undergraduate students withdevelopmental dyslexia. Journal of Research in Reading, 20(1), 22–30.

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)

Page 15: The reading comprehension abilities of dyslexic students in higher education

F. Simmons and C. Singleton192

MacGinitie, W.H. (1978) Gates–MacGinitie Reading Test, Level F. Windsor, Berkshire:NFER-Nelson.MacLeod, J. and Anderson, J. (1973) GAPADOL Reading Comprehension. London: HeinemannEducational Books.McLoughlin, D., Fitzgibbon, G. and Young, V. (1994) Adult Dyslexia: Assessment Counsellingand Training. London: Whurr.McLoughlin, D. (1997) Assessment of adult reading skills. J.R. Beech and C.H. Singleton(Eds), The Psychological Assessment of Reading, pp. 224–237. London: Routledge.Neale, M.D. (1989) Neale Analysis of Reading Ability, 3rd Edn. Windsor, Berkshire:NFER-Nelson.Nicolson, R.I. and Fawcett, A.J. (1990) Automaticity: a new framework for dyslexia research.Cognition, 30, 159–182.Nicolson, R.I. and Fawcett, A.J. (1995) Dyslexia is more than a phonological disability.Dyslexia, 1, 19–36.Nicolson, R.I. and Fawcett, A.J. (1997) Development of objective procedures for screeningand assessment of dyslexic students in higher education. Journal of Research in Reading, 20(1),77–83Orton Dyslexia Society (1994) A new definition of dyslexia. Bulletin of the Orton DyslexiaSociety (now the International Dyslexia Association), Fall, 1994.Oakhill, J.V. (1984) Inferential and memory skills in children’s comprehension of stories.British Journal of Educational Psychology, 54, 31–39.Oakhill J.V. and Cain K. (1997) Assessment of comprehension in reading. In J.R. Beech andC.H. Singleton (Eds), The Psychological Assessment of Reading, pp. 176–203. London:Routledge.Oakhill, J. and Garnham, A. (1988) Becoming a Skilled Reader. Oxford: Blackwell.Palmer, S. (2000) Developmental deficit in phonological recoding in working memory indyslexic teenagers. Journal of Research in Reading, 23(1), 28–40.Perfetti, C.A. (1985) Reading Ability. Oxford: Oxford University Press.Perfetti, C.D. and Goldman, S.R. (1976) Discourse memory an reading comprehension skill.Journal of Verbal Learning and Verbal Behaviour, 14, 33–42.Rack, J. (1997) Issues in the assessment of developmental dyslexia in adults: Theoretical andapplied perspectives. Journal of Research in Reading, 20(1), 66–76.Riddock, B., Farmer, M. and Sterling C. (1997) Students and Dyslexia: Growing up with aSpecific Learning Difficulty. London: Whurr.Singleton, C.H. (1995) Issues in the diagnosis and assessment of dyslexia in HigherEducation. In J. Waterfield (Ed.), Dyslexia in Higher Education: Learning Along the Continuum(Proceedings of the International Conference on Dyslexia in Higher Education, November,1994), pp. 11–18. Plymouth: University of Plymouth.Singleton C.H. (1996) Dyslexia in Higher Education-Issues for policy and practice. In C.Stephens (Ed.), Dyslexic Students in Higher Education, pp. 9–18. Huddersfield: Skill andUniversity of Huddersfield.Singleton, C.H. (Chair) (1999) Dyslexia in Higher Education: Policy, Provision and Practice (TheReport of the National Working Party on Dyslexia in Higher Education). Hull: University ofHull.Snowling, M., Nation, K., Moxham, P., Gallagher A. and Frith, U. (1997) Phonologicalprocessing skills of dyslexic students in higher education: a preliminary report. Journal ofResearch in Reading, 20, 31–41.Tregaskes, M. and Daines, D. (1989) Effects of metacognition on reading comprehension.Reading Research and Instruction, 29(1), 52–60.Wilkinson, G.S. (1993) WRAT-3: Wide Range Achievement Test. Delaware: Jastak Associates.Yap, R. and van der Leij, A. (1994) Automaticity deficits in word reading. In A. Fawcett andR. Nicolson (Eds), Dyslexia in Children: Multidisciplinary perspectives, pp. 77–106. HemelHempstead: Harvester Wheatsheaf.

Copyright © 2000 John Wiley & Sons, Ltd. Dyslexia 6: 178–192 (2000)