Copyright © The British Psychological Society

Reproduction in any form (including the internet) is prohibited without prior permission from the Society

683

The
British
Psychological
British Journal of Educational Psychology (2006), 76, 683–696
q 2006 The British Psychological Society
Society

www.bpsjournals.co.uk

Profiles of children with specific reading

comprehension difficulties

Kate Cain1* and Jane Oakhill2

1
University of Essex, UK
2
University of Sussex, UK

Background. Children with fluent and accurate word reading in the presence of
poor text comprehension are impaired on a wide range of reading-related tasks.
Aims. This study investigated the consistency of skill impairment in a sample of poor
comprehenders to identify any fundamental skill weakness that (i) might be associated
with poor text comprehension, and (ii) might lead to depressed reading development.
An additional aim was to determine whether reading comprehension difficulties are
associated with more general educational difficulties.
Sample. Twenty-three poor comprehenders and 23 good comprehenders with age-
appropriate word reading accuracy were assessed when aged 8 years. Concurrent
reading and language performance and reading, educational attainment and reasoning
skills 3 years later are reported.
Methods. The following skills were assessed when aged 8 years: word reading, text
comprehension, vocabulary, syntax, cognitive ability, working memory, comprehension
subskills. Listening comprehension, SAT scores and reasoning scores at 11 years are
also reported.
Results. There was no evidence for any fundamental skill weaknesses in the
population of poor comprehenders at Time 1. However, poor vocabulary skills led to
impaired growth in word reading ability and poor general cognitive ability led to
impaired growth in comprehension. Poor comprehenders obtained lower SAT scores
than did the good comprehenders at 11 years.
Conclusions. These findings indicate that a single underlying source of poor
comprehension is unlikely. Poor comprehenders are at risk of generally poor
educational attainment, although weak verbal or cognitive skills appear to affect the
reading development of poor comprehenders in different ways.

Text comprehension draws on many different language skills. These include lower-level
lexical skills such as word reading efficiency and vocabulary knowledge, sentence-level

* Correspondence should be addressed to Kate Cain, Department of Psychology, Fylde College, Lancaster University, Lancaster
LAI 4YF, UK (e-mail: k.cain@lancs.ac.uk).

DOI:10.1348/000709905X67610
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

684 Kate Cain and Jane Oakhill

skills such as knowledge of grammatical structure and higher-level text processing skills
such as inference generation, comprehension monitoring and working memory
capacity (e.g. Carr, Brown, Vavrus, & Evans, 1990; Hannon & Daneman, 2001; Palincsar
& Brown, 1984; Perfetti & Hart, 2001; Perfetti, Marron, & Foltz, 1996). Higher-level skills
are related to text comprehension because they enable the reader to make the necessary
integrative and inferential links to construct a meaning-based representation of the text.
Efficient lower-level lexical skills facilitate reading comprehension by enabling more
resources to be devoted to higher-level processes.
The relation between these higher- and lower-level skills and text comprehension
have been demonstrated in both correlational and longitudinal studies. Concurrent
measures of vocabulary, working memory, inference making and comprehension
monitoring are related to reading comprehension skills in 7- to 10-year-olds (Cain,
Oakhill, & Bryant, 2004). Longitudinal studies have shown that vocabulary and working
memory (Seigneuric & Ehrlich, 2005) and word reading, grammatical awareness and
vocabulary (Muter, Hulme, Snowling, & Stevenson, 2004) are related to later reading
comprehension.
It is not surprising, therefore, that children with reading comprehension difficulties
show impairments on a range of language tasks. When word reading ability and written
vocabulary knowledge are controlled, poor comprehenders demonstrate deficits on
higher-level skills relative to same-age good comprehenders. Impairments have been
found on measures of working memory (Yuill, Oakhill, & Parkin, 1989), inference
making (Cain & Oakhill, 1999; Cain, Oakhill, Barnes, & Bryant, 2001), narrative
production (Cain, 2003; Cain & Oakhill, 1996) and comprehension monitoring (Ehrlich,
Remond, & Tardieu, 1999; Oakhill, Hartt, & Samols, 2005). Weaknesses in syntactic
knowledge and processing have also been reported (Nation, Clake, Marshall, & Durand,
2004; Stothard & Hulme, 1992). When vocabulary, word reading accuracy and/or
non-word reading are allowed to vary, lower-level lexical weaknesses in exception word
reading and semantic processing are also apparent (Nation & Snowling, 1998a, 1998b,
1999). Thus, there are many language factors that might be causally related to reading
comprehension level and the development of this skill.
Cornoldi, de Beni, and Pazzaglia (1996) urge caution in what we conclude from such
correlational studies for the following reasons. First, there are four different ways that a
skill can be related to comprehension ability: it may be a prerequisite of reading
comprehension, a facilitator, a consequence or simply an incidental correlate.
Comparisons between good and poor comprehenders do not distinguish between
these alternatives. Second, such comparisons do not establish whether poor
comprehenders in general present a particular deficit or whether every poor
comprehender presents that deficit. Given the number of skills associated with good
reading comprehension, it would be surprising (and discouraging from a remediation
point of view) if all the abilities associated with good reading comprehension were
deficient in all poor comprehenders, and necessary for good reading comprehension to
develop. Skills that are fundamental to reading comprehension should be impaired in
the majority of children with specific reading comprehension deficits. However, skills
for which deficits can arise as a consequence of poor reading comprehension or are
present as an incidental correlate should be preserved in some individuals with poor
reading comprehension but weak in others.
Cornoldi et al. (1996) explored these issues in a sample of 11-year-old Italian
schoolchildren who had weak reading comprehension skills in the presence of good
decoding skills and normal intelligence. Different patterns of strength and weakness
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 685

were evident across a range of measures important for text comprehension and not all
poor comprehenders experienced impairments in all of the skills. There may be
subtypes of poor comprehender with different fundamental weaknesses, however
Cornoldi et al.’s sample was too small to detect such groups.
One potential fundamental weakness is verbal ability, which is strongly correlated
with reading comprehension (Carroll, 1993). Yet, as Nation, Clark, and Snowling (2002)
point out, there have been few investigations of underlying verbal and non-verbal ability
in children with reading comprehension deficits. Stothard and Hulme (1996) found
depressed Verbal IQ (VIQ) relative to Performance IQ (PIQ; pro-rated from the WISC) in
a population of 7- to 8-year-old poor comprehenders. In contrast, Nation et al. found
depressed levels of both verbal and non-verbal ability in poor comprehenders aged 8 to
9 years old, although the non-verbal deficit was only marginally significant for poor
comprehenders with normal range general ability scores. Surprisingly, the poor
comprehenders with high and low general ability scores obtained comparable word
reading and reading comprehension scores, even though the two groups differed
significantly on verbal ability. These results indicate that it is important to examine
further the relation between verbal ability and reading comprehension. If verbal ability
is causally related to reading comprehension, one would expect the low-ability group to
show a more severe comprehension deficit than their peers.

Purpose of the study

There are three central issues that arise from the work of Cornoldi, Nation and their
colleagues addressed in this paper as follows.

Variations in language and cognitive skills in children with poor reading

comprehension
Cornoldi and colleagues found varying patterns of skill strength and weakness in their
poor comprehenders between the ages of 11 and 13 years. In the UK, research has
focused on comprehension impairments in younger children aged between 8 and 11
years and, to date, these investigations have focused on single component skills of
reading comprehension (e.g. Cain et al., 2001; Oakhill et al., 2005; Stothard & Hulme,
1996). We know little about the consistency of impairments in this age range. As noted
earlier, differences between the selection procedures used in different studies might
lead to the presence of deficits in some studies and not others. We present profiles of
poor comprehenders across a range of language and cognitive skills to identify whether
there are consistent deficits associated with poor reading comprehension in this age
group, which would inform both the remediation of poor comprehension and our
theoretical understanding of reading comprehension.

Relations between verbal and vocabulary ability and literacy development in children
with poor reading comprehension
Nation et al. did not find an association between general cognitive skills and the severity
of the comprehension deficit presented by poor comprehenders. We investigated an
additional hypothesis that cognitive and verbal skills act as facilitators of reading
development, by looking at progress in reading over time. If general cognitive ability is
associated with poor reading comprehension, poor comprehenders should experience
non-verbal as well as verbal processing deficits. We explored the association between
verbal, quantitative and spatial reasoning skills and comprehension ability 3 years after
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

686 Kate Cain and Jane Oakhill

initial selection when our sample was aged 11 years, to investigate the specificity of the
relation between reasoning skills and comprehension level.

Persistence and educational implications of a comprehension impairment

Cornoldi et al. found that 7 of the original 12 poor comprehenders demonstrated
comprehension deficits 2 years later, whilst the other 5 participants were no longer at
risk. This statistic is very encouraging, suggesting that around 40% of those with a
comprehension problem at 11 years do not have a persistent severe deficit. We looked at
the persistence of a comprehension deficit between 8 and 11 years. We also compared
the good and poor comprehenders’ achievement performance in three key curriculum
areas, English, mathematics and science when they were aged 11 years, to determine
whether early problems with reading comprehension are associated with later
scholastic difficulties, and the specificity of any problems.

Method
Participants
The participants were taken from a group of 102 children (63 girls, 39 boys), first assessed
at 7 to 8 years of age in a longitudinal study. The children were drawn from 17 classrooms
in six schools serving middle- and lower-middle-class catchment areas. Word reading
accuracy and reading comprehension were assessed with the Neale analysis of reading
ability – revised British edition (Neale, 1989). Two groups of children, good and poor
comprehenders, were selected according to the following criteria. Children with Neale
word reading accuracy scores that were more than 6 months below or more than 12
months above their chronological age were excluded. The good comprehenders had
Neale reading comprehension scores that were at or above that predicted by their word
reading accuracy scores. The poor comprehender group were the remaining children
who had a discrepancy of 12 months or greater between their chronological age and their
reading comprehension age, and also between their word reading accuracy and
comprehension age. The participants therefore all had age-appropriate word reading skill
and either good or poor comprehension. At Time 1, each group comprised 23 children
(good comprehenders: 13 girls, 10 boys; poor comprehenders: 15 girls, 8 boys).

Additional assessments: Time 1

A number of reading and reading-related skills were measured using standardized
assessments and our own experimental measures. A brief description of each test is
given below (further details can be obtained from the first author).

Vocabulary knowledge
Two assessments of vocabulary knowledge were taken. Sight vocabulary was assessed
using the Gates-MacGinitie Test (MacGinitie & MacGinitie, 1989), which is group
administered. Children were given the Level 2 test (Form K), in which they had to select
one of out four words to go with an accompanying picture. The total number correct was
calculated. Receptive vocabulary was assessed using the British Picture Vocabulary Scales
(BPVS; Dunn, Dunn, Whetton, & Pintillie, 1982). Standardized scores were calculated.

Memory
Two assessments of working memory were completed. One was a listening span task in
which children were required to provide a single word completion for a sentence
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 687

spoken by the experimenter (processing component) and remember that word for later
recall (storage component). The other assessment was Yuill et al.’s (1989) digit reading
task. Children read out loud groups of three digits (processing) and remembered the
final digit for later recall (storage). Children completed three trials of two, three and four
groups of sentences or digits for each task.

Knowledge of syntax
The Test for Reception of Grammar (TROG; Bishop, 1983) was used to measure
children’s knowledge of syntax. Blocks L, N, O, Q, R, S and T were administered, because
pilot work demonstrated that children in this age group were not at ceiling on these
blocks. The total number of items correct was calculated.

General intellectual ability

Estimates of verbal and non-verbal intelligence were made using the WISC-III (UK
edition; Wechsler, 1992). Children completed two subtests from the verbal scale –
vocabulary and similarities – and two from the performance scale – block design and
object assembly. The mean standard scores are reported.

Specific comprehension subskills

Three comprehension subskills were assessed: comprehension monitoring; inference
and integration skill; and knowledge about story structure. In the comprehension
monitoring task, children read stories, some of which contained pairs of inconsistent
statements. Their task was to underline ‘any bits that did not make sense together’. The
sum of correctly identified inconsistencies and correctly identified consistent stories
was calculated. Integration and inference making skill was assessed using Oakhill’s
(1982) integration task, which assesses the ability to make constructive inferences
between two sentences in three-line texts. The score used was the total number of
integrated statements and literal statements that were accepted. Two tasks assessed
knowledge of story structure. The first was a story anagram task (Stein & Glenn, 1982) in
which children put together stories that had been cut up into their six constituent
sentences and randomized. Concordance scores (between target and actual sentence
order) were computed. Understanding of the purpose of story titles (the sorts of
information about characters, locations and events that titles contain) was assessed by a
structured interview (Cain, 1996).

Outcome assessments
The outcome measures were taken when the children were aged 10 to 11 years.
Seventeen of the original good comprehenders and 19 of the poor comprehenders were
available for re-retesting on the Neale analysis of reading ability - revised British edition.
We had access to the raw scores obtained by 16 good comprehenders and 17 poor
comprehenders on their Key Stage 2 Standard Assessment Tests (SATs) in English,
mathematics and science. These tests are taken by all English schoolchildren during the
May in the year that they turn 11 years and assess how well children have progressed in
the core curriculum subjects. We also assessed the listening comprehension skills of 16
children in each group. Children were read the six stories of the Neale Analysis of
Reading Ability, Form 2. After each story they provided written answers to the set of
questions. The schools provided us with the standardized scores on the three
assessments of reasoning that make up the Cognitive Abilities Tests – second edition
(Thorndike, Hagen, & France, 1986): verbal (17 poor comprehenders, 15 good
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

688 Kate Cain and Jane Oakhill

comprehenders); and quantitative and non-verbal reasoning (17 poor comprehenders,

16 good comprehenders). These assessments were administered in the autumn term of
the year that the children turned 11 years.

Results
Reliability of our experimental measures was assessed using Cronbach’s alpha over
items. In most cases, the reliability coefficient was acceptable (.60 to .80). However, the
measure of inference and integration produced an alpha level of .48 (over all 24 items).
For all significant group comparisons, Cohen’s d was calculated as a measure of effect
size. A value of .2 is regarded as a small effect, .5 is medium and .8 and above is large
(Cohen, 1988).

Skills that differentiate good and poor comprehenders at Time 1

A series of t tests was performed on the Time 1 measures comparing good and poor
comprehenders. The mean scores and results of the t tests are shown in Table 1. The
two groups differed on many comprehension-related skills that have differentiated
similarly selected groups in previous work; namely, verbal working memory, the ability
to structure stories, knowledge about the purpose of story titles, inference and
integration and comprehension monitoring (Cain, 1996, 2003; Cain et al., 2004; Cain,
Oakhill, & Lemmon, 2004; Oakhill et al., 2005; Oakhill, 1982). The two groups did not
differ on the digit working memory measure (in contrast to Yuill et al., 1989). There was
little evidence of more generalized language and verbal impairments. The groups did not
obtain significantly different scores on the measure of syntactic knowledge, TROG (in
contrast to Stothard & Hulme, 1992), but the good comprehenders obtained
significantly higher scores on the measure of receptive vocabulary, BPVS. The two
groups obtained comparable scores on the performance ability measure of the WISC- III
although the good comprehenders demonstrated marginally higher verbal ability
scores.1

Variations in language and cognitive skills in children with poor reading

comprehension
Table 2 shows the distribution of scores obtained by the poor comprehenders at Time 1.
The scores used were the standardized scores for the BPVS, PIQ and VIQ assessments,
for which the hypothesized mean is 100 with a standard deviation of 15. For the
remaining assessments, z scores for the entire sample in our study (N ¼ 102) were used
to indicate skill strength or weakness. A composite memory score was calculated from
the mean z scores of the two memory measures and a composite comprehension score
was calculated from the mean of the four comprehension assessments. The two
vocabulary assessments were explored separately because the Gates-MacGinitie
measure involves word reading skill, whereas the BPVS is a measure of receptive
vocabulary. The group is divided into those who exhibited a deficit (score below the
mean) or an advantage (above the mean) on each measure.
The majority of the poor comprehenders performed below the sample mean on all
measures with the exception of the two vocabulary measures (Gates-MacGinitie, BPVS)

1
This marginal difference should be treated with caution because of the number of comparisons made.
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 689

Table 1. Time 1 descriptive statistics for good comprehenders, poor comprehenders and the total
sample

Poor Good
Total sample comprehenders comprehenders Effect size
Measure (N ¼ 102) (N ¼ 23) (N ¼ 23) t(44) Cohen’s d

Chronological age 7, 07 (3.28) 7, 07 (2.88) 7, 07 (3.44) 0.51

Neale reading 7, 10 (6.27) 7, 10 (4.58) 7, 10 (5.30) 0.29
accuracy
Neale reading 7, 04 (11.19) 6, 05 (9.22) 8, 04 (6.64) 9.51*** 2.74
comprehension
BPVS 102.99 (9.50) 98.57 (11.69) 106.30 (6.36) 2.79** .82
Gates-MacGinitie 34.30 (4.63) 34.26 (4.22) 35.74 (4.75) 0.36
VIQ 10.42 (2.18) 9.56 (1.80) 10.67 (1.99) 1.98(*) .58
PIQ 10.45 (2.47) 9.85 (3.03) 11.15 (2.32) 1.64
TROG 21.61 (2.67) 20.96 (2.27) 22.17 (2.94) 1.55
WM verbal 11.30 (3.05) 10.30 (2.73) 12.57 (3.51) 2.43* .72
WM digit 10.68 (3.05) 10.35 (2.82) 10.83 (3.27) 0.53
Story anagram .80 (.16) .74 (.15)a .84 (.17) 2.06* .62
Story titles 2.93 (1.15) 2.23 (1.08) 3.65 (.77) 5.11*** 1.51
Monitoring 14.54 (3.10) 12.74 (3.32) 16.17 (2.25) 4.11*** 1.21
Inference and 14.84 (3.75) 13.60 (3.91) 16.00 (3.38) 2.22* .66
integration
a
One poor comprehender absent.
*p , :05, **p , :01, ***p , :001, (*)p ¼ :064.
Key for labels used in tables:
Chronological age, Neale word reading accuracy, Neale comprehension scores all given as years,
months (months).
BPVS ¼ British Picture Vocabulary Scales, standardized scores.
Gates-MacGinitie ¼ Gates-MacGinitie Sight Vocabulary Test, maximum score ¼ 45.
VIQ ¼ mean standard scores from verbal scale of WISC-III.
PIQ ¼ mean standard scores from performance scale of WISC-III.
TROG ¼ Bishop’s Test of Reception of Grammar, maximum score ¼ 28.
WM verbal ¼ verbal working memory task and WM digit ¼ digit working memory task, maximum
score ¼ 27.
Story anagram ¼ story structure task, concordance scores.
Story titles ¼ story titles task, maximum score ¼ 4.
Monitoring ¼ a comprehension monitoring task, maximum score ¼ 24.
Inference and integration ¼ inference making task, maximum score ¼ 16.

and PIQ, where a majority obtained above mean scores. For the working memory
measures, 9 of the 23 children obtained below mean scores on both measures, and 12 of
the 23 poor comprehenders obtained below mean scores on three or four of the
comprehension subskills. Despite obtaining significantly lower mean BPVS scores than
good comprehenders in the group comparisons, the majority of poor comprehenders
did not present a receptive vocabulary deficit in relation to age-appropriate
performance. In contrast, although the poor comprehenders’ mean VIQ score was
only marginally lower than that obtained by the good comprehenders, the majority of
this group obtained VIQ scores that were below the mean for their age.
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

690 Kate Cain and Jane Oakhill

Table 2. Number of poor comprehenders with scores above or below the sample mean on Time 1
measures

Deficit of at Score between Score between Advantage of at

Measure least 1 SD 0 and 21 SD 0 and þ 1 SD least 1 SD

BPVS 3 7 13 0
Gates-MacGinitie 2 9 8 4
VIQ 4 11 6 2
PIQ 7 3 8 5
TROG 3 11 8 1
Memory composite 2 13 8 0
Comprehension composite 3 16 4 0

Relations between verbal and vocabulary ability and literacy development in children
with poor reading comprehension
There was no evidence for the specific verbal ability deficit found by Stothard and
Hulme (1996). Poor comprehenders obtained mean PIQ standard scores that were
comparable to their VIQ scores (scores in Table 1): tð22Þ , 1. Variations within the
sample of 23 poor comprehenders were evident. Seven children obtained VIQ scores
below the mean and PIQ scores above the mean, whereas only two presented the
opposite pattern. Eight children were poor on both measures and six obtained scores
higher than the mean on both.
To determine whether variations in cognitive ability or vocabulary skills were
associated with different profiles of literacy performance, we classified the poor
comprehenders on the basis of their general cognitive ability (mean of the standard
scores from the four subtests of WISC-III) and their receptive vocabulary knowledge
(BPVS standardized scores). Thirteen children obtained an average score from the IQ
subtests above the population mean of 10 (M ¼ 11:08, SD ¼ 0:88) and 10 children
obtained below the population mean (M ¼ 7:93, SD ¼ 1:02). These were the high and
low cognitive ability groups, respectively. Thirteen poor comprehenders obtained a
BPVS standardized score above the population mean of 100 (M ¼ 107:16, SD ¼ 4:16)
and 10 obtained a score below mean scores (M ¼ 87:40, SD ¼ 1:41). These were the
high and low vocabulary groups, respectively.
The high- and low-ability (cognitive and vocabulary) groups did not differ in their
reading comprehension and word reading levels at Time 1 (all ts , 1:4, all ps . :15),
but general cognitive and vocabulary skills at age 8 years affected the progress made in
reading between 8 and 11 years. Poor comprehenders with low cognitive ability at Time
1 made significantly less progress in reading comprehension than those with high
cognitive ability. The means (and standard deviations) for the low cognitive (N ¼ 9) and
high cognitive (N ¼ 10) groups were 17.0 (10.2) and 27.6 months (5.9), respectively,
tð17Þ ¼ 2:81, p , :015, d ¼ 1:27. These groups made comparable progress in word
reading development: 46.9 (14.7) and 46.9 months (15.7), respectively, tð17Þ , 1. In
contrast, the comparisons between the high (N ¼ 10) and low (N ¼ 9) vocabulary
groups revealed a significant difference in the progress made in word reading accuracy.
In order, the means (and standard deviations) were 53.7 (10.2) and 39.3 (15.9),
tð17Þ ¼ 2:37, p ¼ :030, d ¼ 1:08. These two groups did not differ significantly in the
progress made in reading comprehension development, obtaining means of 24.5 (8.9)
and 20.4 (10.5) months, respectively, tð17Þ , 1. The comparisons are based on a small
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 691

sample size, so caution must be taken when interpreting these results. However, where
comparisons were significant, substantial effects were apparent.

Persistence of a comprehension impairment

The children classified as good and poor comprehenders at Time 1 differed significantly
in their reading comprehension level three years later (Ms ¼ 124:8 and 98.4 months,
SDs ¼ 13:7 and 12.4), tð34Þ ¼ 6:08, p , :001, d ¼ 2:02. The two groups also differed in
their scores on the listening comprehension measure (Ms ¼ 24:5 and 20.3, SDs ¼ 4:7
and 4.4), tð30Þ ¼ 2:62, p , :05, d ¼ :93. Surprisingly, there was not a significant
difference in word reading level (Ms ¼ 145:4 and 140.2 months, SDs ¼ 10:6 and 15.0),
tð34Þ ¼ 1:20, p ¼ :24. Poor comprehenders, in general, maintain their reading
comprehension deficit, but poor comprehension at 8 years does not necessarily lead
to depressed word reading development.
At 11 years of age, a large proportion of our sample obtained word reading accuracy
scores that were more than 12 months above their chronological age. (Possible reasons
for this are explored in the Discussion). Children whose reading comprehension scores
were below their chronological age at outcome were classified as poor
comprehenders.2 One child classified as a poor comprehender at Time 1 obtained an
age-appropriate comprehension score at outcome. In the original good comprehender
population, four children obtained reading comprehension scores below both their
chronological age and the sample mean at outcome. Thus, although the impairments
and advantages remained fairly consistent, there were exceptions.

Educational implications of a comprehension impairment

The good comprehenders obtained significantly higher scores than the poor comprehenders
on all SAT measures. The mean raw scores for the two groups in order are as follows: English,
Ms ¼ 66:3 and 60.3, SDs ¼ 7:4 and 8.0, tð31Þ ¼ 2:23, p , :05, d ¼ :78; maths, Ms ¼ 78:7
and 63.7, SDs ¼ 10:2 and 17.5, tð31Þ ¼ 3:57, p , :01, d ¼ 1:05; science, Ms ¼ 61:6 and
51.9, SDs ¼ 7:1 and 8.4, tð31Þ ¼ 3:57, p , :01, d ¼ 1:25. Cohen’s d indicates medium-large
effects, but it should be noted that the raw scores correspond to an average attainment of
Level 4 (age appropriate) for both groups for all subject areas.
The standardized scores obtained on the three different assessments of reasoning
ability revealed age-appropriate performance in verbal, non-verbal and quantitative skills
for both good and poor comprehenders: CAT verbal, Ms ¼ 109:7 and 100.8, SDs ¼ 7:1
and 7.4; quantitative, Ms ¼ 106:4 and 102.2, SDs ¼ 7:8 and 9.5; non-verbal, Ms ¼ 108:7
and 103.7, SDs ¼ 9:9 and 9.3. However, there was a significant difference between the
good and poor comprehenders on the measure of verbal reasoning, tð30Þ ¼ 3:43,
p , :01, d ¼ 1:23. Comparisons on the two other measures were not significant, both
ts , 1:5, both ps . :14.
Correlational analyses revealed that verbal reasoning was significantly related to SAT
performance: maths, r ¼ :54, p , :01; science, r ¼ :60, p , :001; English, r ¼ :63,
p , :001. It was also significantly correlated with both concurrent comprehension,
r ¼ :64, p , :001, and comprehension three years earlier, r ¼ :67, p , :001 (N ¼ 29).
One factor that might underpin the relations between verbal reasoning and these other

2
An alternative method by which we calculated the z scores for comprehension scores of the entire sample and compared
children who obtained comprehension scores above and below the mean gave the same results.
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

692 Kate Cain and Jane Oakhill

measures is verbal working memory, a skill that is associated with good comprehension
(e.g. Cain, Oakhill, & Bryant, 2004). There was little evidence for that. When working
memory scores were partialled out, there were significant correlations between Time 3
verbal reasoning and comprehension (r ¼ :54), SATs English (r ¼ :55) and science
(r ¼ :43, all ps , :025), although the correlation between verbal reasoning and SAT
maths no longer reached significance, r ¼ :27, p ¼ :16.

Discussion
In this paper, we addressed three issues related to poor reading comprehension:
(i) Variations in language and cognitive skills in children with poor reading
comprehension – are there consistent language and cognitive skill deficits associated
with poor reading comprehension skills at 8 years of age? (ii) Relations between verbal
and vocabulary ability and literacy development in children with poor reading
comprehension – does poor verbal ability or vocabulary knowledge result in a more
severe comprehension impairment at 8 years or affect the progress made in reading
between 8 and 11 years? (iii) Persistence and educational implications of a
comprehension impairment – how does poor comprehension when aged 8 impact
upon later reading performance and educational attainment? We discuss the findings
relating to each, in turn, and consider the educational and theoretical implications of
this work.
Children with reading comprehension problems presented deficits on a range of
literacy assessments associated with the meaning-based aspects of reading. Of interest to
both theoretical models of reading development and educational practitioners was the
substantial heterogeneity within this population. The effect sizes for the significant
group comparisons indicated medium-large effects, but we did not find consistent
deficits associated with poor reading comprehension. This work extends previous
studies by demonstrating profile differences across a range of literacy tasks. Although
poor comprehenders in general may suffer from weak monitoring skills or inferior
inference making ability, some poor comprehenders demonstrate average or even good
performance on these tasks relative to peers. The low reliability of the inference
measure means that that result should be treated with caution. However, this measure
has previously demonstrated reliable group differences between good and poor
comprehenders (e.g. Oakhill, 1982).
There were no significant differences between good and poor comprehenders’
knowledge of the meanings of single written words or grammatical structures. However,
the performance of some poor comprehenders on these measures was greater than one
standard deviation below the group mean. Although this analysis was not designed to
address causality, we can speculate that some children in our sample might have
experienced problems in their processing of text because of weak word- or sentence-
level skills, whilst others may have been limited by weak higher order processing skills.
Unfortunately, the sample size does not provide sufficient power to test such
hypotheses; further work with a much larger sample is required.
The results of our profiling concur with those obtained by Cornoldi et al. (1996) in
that no clear fundamental weakness was apparent. Instead, this work indicates that
group comparisons may obscure crucial weaknesses in the individual. For the
practitioner, these findings highlight the need to tailor intervention programmes to the
specific weaknesses presented by each child. For the theorist, they indicate that reading
comprehension level can be determined by many different language and cognitive
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 693

factors. Our understanding of comprehension development may be better advanced by

investigation of the interaction between different language and cognitive abilities,
rather than a focus on lower- or higher-level reading-related skills.
There was no evidence that poor comprehenders with low cognitive or vocabulary
skills were more severely impaired on concurrent measures of language and literacy
than their more able peers. Thus, cognitive and verbal skills do not appear to be
prerequisites for the attainment of age-appropriate reading comprehension in the early
stages of reading. However, initial levels of cognitive ability were related to growth in
their reading comprehension and initial levels of receptive vocabulary were related to
growth in word reading indicating that cognitive ability and vocabulary skills act as
facilitators of comprehension and word reading, respectively. Although the sample size
was small, the effect sizes of these differences were substantial. Thus, this finding
warrants further investigation with a larger sample.
The relation between cognitive ability and growth in reading comprehension might
have arisen if the cognitive ability scores reflected general information processing and
reasoning capacity. These skills could, in turn, underpin growth in text processing skill.
There was no evidence that the good and poor comprehenders differed in reasoning
skill in general, when aged 11; a significant and sizable group difference was only found
on the measure of verbal reasoning. Thus, it may be that poor comprehenders have an
impairment with processing complex information in the verbal domain, but not a
general reasoning impairment. One source of these difficulties in text processing and
verbal reasoning might be the well-documented working memory deficits experienced
by poor comprehenders. However, the significant relation between comprehension skill
and verbal reasoning was apparent even when verbal working memory skills were
partialled out. Processing capacity is clearly not the sole determinant of this relation.
A detailed analysis of the processing skills and abilities that determine verbal reasoning
should inform the relation between these measures.
The relation found between vocabulary and word reading development is very
interesting. Vocabulary is associated with good text comprehension, but there is
evidence for a relation with word reading as well. For example, Hagtvet (2003) found
strong correlations between vocabulary knowledge and word decoding skill in poor
readers. One possibility for the relation between semantic knowledge and the
development of accurate word reading is the mechanism specified in some
connectionist models of word reading. Harm and Seidenberg (2004) suggest that
whereas the meanings of unfamiliar words are accessed only after decoding, a stronger
direct link from the written to the semantic representation exists for familiar words.
Another possibility is that the initial levels of receptive vocabulary skill reflected home
literacy environment (e.g. Foy & Mann, 2003), which might go on to influence
out-of-school reading and practise in word reading and therefore growth in this skill.
Alternatively, phonological memory may underpin the relation – it is related to
acquisition of receptive vocabulary (Gathercole, Hitch, Service, & Martin, 1997) and
also word reading development (Passenger, Stuart, & Terrell, 2000). Further longitudinal
work is required to disentangle these relations.
The majority of the good and poor comprehender groups maintained their status
across the three years. However, the poor comprehenders did not fall behind in their
word reading skills. Children with poor comprehension may engage in less out-of-school
reading because they derive less enjoyment from the activity that children with good text
comprehension skills, or they may choose to read less demanding literature and therefore
encounter fewer new words in print each year. If so, one would predict that their word
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

694 Kate Cain and Jane Oakhill

reading skills would not develop in line with those of their peers. The absence of word
reading deficits in our poor comprehender sample when aged 11 was surprising and
warrants consideration. This finding might be a consequence of the criteria used to select
the good and poor comprehenders for this analysis. Children who obtained word reading
ages that were more than 6 months below their chronological age were excluded, to
ensure that the poor comprehenders were not weak readers, in general. The majority of
our sample was reading above their chronological age at 11 years, as measured on our
standardized assessment. These children may have had the fundamental skills necessary
to make good progress in reading such that their word reading skills were not at risk. We
would predict limited growth for children who had weaker word reading skills at outset.
Children with poor comprehension at 7 to 8 years had a poorer educational outcome
than their peers, in contrast to Nation et al.’s (2002) finding that only the low-ability poor
comprehenders were poorer on maths (assessed concurrently). It should be noted that
the poor comprehenders were not underperforming for their age, obtaining an average
score that corresponds to Level 4, yet their scores were significantly below those of their
skilled peers and the effect sizes were substantial.
One explanation for the link between reading comprehension and performance on the
curriculum-based maths and science tests might lie in the processing skills that underpin
performance on these measures. Reading comprehension and working memory capacity
are correlated (Cain et al., 2004), and working memory capacity is related to performance
on the national curriculum assessments taken at 7 years (Gathercole & Pickering, 2000).
However, although working memory skills may have affected performance on these tasks,
some of our poor comprehenders did not experience a working memory deficit. Thus,
other factors must have accounted for their poorer performance.
Relative to good comprehenders, the poor comprehenders obtained lower scores on
measures of English, mathematics and science when aged 11 years, yet their reasoning
skills were not impaired relative to their age. The poor comprehenders obtained verbal
reasoning scores below those obtained by the good comprehenders and the correlations
between verbal reasoning and performance on the SATs indicates a surprisingly strong
association. These relations were based on a very small sample size and should be
treated cautiously. However, we recommend that further analysis of the teaching and
assessment of maths and science is carried out to assess how reliant they are on verbal
reasoning skills. As it stands, these findings suggest that children with poor verbal
reasoning skills may be impaired across the wider curriculum.
This study raises many interesting questions about reading comprehension
impairments for both the theorist and practitioner. We did not identify any concurrent
skill deficits that were consistently associated with poor comprehension: within the
population of poor comprehenders, there was considerable difference in the range of
skill strengths and weaknesses. However, poor comprehenders with weak vocabulary
skills and poor cognitive ability were vulnerable to impaired growth in word reading and
reading comprehension, respectively. Furthermore, poor comprehenders performed
more poorly than their peers in areas of the educational curriculum other than English.
We conclude that the determination of reading comprehension skill is considerably
more complex than the result of a simple relation with cognitive level, verbal ability or
reasoning skills, although these factors clearly play a role. When comprehension
problems are identified, careful analysis of other language and cognitive skills must
inform remediation. Intervention studies will not only aid the education of the poor
comprehender in the classroom, but they will also inform our understanding of the skills
that contribute to comprehension problems.
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Profiles of poor comprehenders 695

References
Bishop, D. (1983). Test for the reception of grammar. Manchester: Chapel Press.
Cain, K. (1996). Story knowledge and comprehension skill. In C. Cornoldi & J. V. Oakhill (Eds.),
Reading comprehension difficulties: Processes and remediation (pp. 167–192). Mahwah, NJ:
Erlbaum.
Cain, K. (2003). Text comprehension and its relation to coherence and cohesion in children’s
fictional narratives. British Journal of Developmental Psychology, 21, 335–351.
Cain, K., & Oakhill, J. (1996). The nature of the relationship between comprehension skill and the
ability to tell a story. British Journal of Developmental Psychology, 14, 187–201.
Cain, K., Oakhill, J., & Bryant, P. E. (2004). Children’s reading comprehension ability: Concurrent
prediction by working memory, verbal ability, and component skills. Journal of Educational
Psychology, 96, 31–42.
Cain, K., Oakhill, J., & Lemmon, K. (2004). Individual differences in the inference of word
meanings from context: The influence of reading comprehension, vocabulary knowledge, and
memory capacity. Journal of Educational Psychology, 96, 671–681.
Cain, K., & Oakhill, J. V. (1999). Inference making ability and its relation to comprehension failure.
Reading and Writing, 11, 489–503.
Cain, K., Oakhill, J. V., Barnes, M. A., & Bryant, P. E. (2001). Comprehension skill, inference making
ability and their relation to knowledge. Memory and Cognition, 29, 850–859.
Carr, T. H., Brown, T. L., Vavrus, L. G., & Evans, M. A. (1990). Cognitive skill maps and cognitive
skill profiles: Componential analysis of individual differences in children’s reading efficiency.
In T. H. Carr & B. A. Levy (Eds.), Reading and its development: Component skills approaches
(pp. 1–55). London: Academic Press.
Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. New York:
Cambridge University Press.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed). Hillsdale, NJ:
LEA.
Cornoldi, C., De Beni, R., & Pazzaglia, F. (1996). Profiles of reading comprehension difficulties: An
analysis of single cases. In C. Cornoldi & J. Oakhill (Eds.), Reading comprehension difficulties:
Processes and interventions (pp. 113–136). Mahwah, NJ: Erlbaum.
Dunn, L. M., Dunn, L. M., Whetton, C., & Pintillie, D. (1982). British Picture Vocabulary Scale.
Windsor: NFER-Nelson.
Ehrlich, M. F., Remond, M., & Tardieu, H. (1999). Processing of anaphoric devices in young skilled
and less skilled comprehenders: Differences in metacognitive monitoring. Reading and
Writing, 11, 29–63.
Foy, J. G., & Mann, V. (2003). Home literacy environment and phonological awareness in preschool
children: Differential effects for rhyme and phoneme awareness. Applied Psycholinguistics,
24, 59–88.
Gathercole, S. E., Hitch, G. J., Service, E., & Martin, A. J. (1997). Phonological short-term memory
and new word learning in children. Developmental Psychology, 33, 966–979.
Gathercole, S. E., & Pickering, S. J. (2000). Working memory deficits in children with low
achievements in the national curriculum at 7 years of age. British Journal of Educational
Psychology, 70, 177–194.
Hagtvet, B. E. (2003). Listening comprehension and reading comprehension in poor decoders:
Evidence for the importance of syntactic and semantic skills as well as phonological skills.
Reading and Writing, 16, 505–539.
Hannon, B., & Daneman, M. (2001). A new tool for measuring and understanding individual
differences in the component processes of reading comprehension. Journal of Educational
Psychology, 93, 103–128.
Harm, M. W., & Seidenberg, M. S. (2004). Computing the meanings of words in reading:
Cooperative division of labor between visual and phonological processes. Psychological
Review, 111, 662–720.
 Copyright © The British Psychological Society
Reproduction in any form (including the internet) is prohibited without prior permission from the Society

696 Kate Cain and Jane Oakhill

MacGinitie, W. H., & MacGinitie, R. K. (1989). Gates-MacGinitie reading tests. Chicago: Riverside
Publishing.
Muter, V., Hulme, C., Snowling, M. J., & Stevenson, J. (2004). Phonemes, rimes and grammatical
skills as foundations of early reading development: Evidence from a longitudinal study.
Developmental Psychology, 40, 665–681.
Nation, K., Clarke, P., & Snowling, M. J. (2002). General cognitive ability in children with reading
comprehension difficulties. British Journal of Educational Psychology, 72, 549–560.
Nation, K., Clarke, P., Marshall, C., & Durand, M. (2004). Hidden language impairments in children:
Parallels between poor reading comprehension and specific language impairment? Journal of
Speech, Language, and Hearing Research, 47, 199–211.
Nation, K., & Snowling, M. J. (1998a). Individual differences in contextual facilitation: Evidence
from dyslexia and poor reading comprehension. Child Development, 69, 996–1011.
Nation, K., & Snowling, M. J. (1998b). Semantic processing and the development of word-
recognition skills: Evidence from children with reading comprehension difficulties. Journal of
Memory and Language, 39, 85–101.
Nation, K., & Snowling, M. J. (1999). Developmental differences in sensitivity to semantic relations
among good and poor comprehenders: Evidence from semantic priming. Cognition, 70,
81–83.
Neale, M. D. (1989). The Neale Analysis of Reading Ability – Revised. Windsor: NFER-Nelson.
Oakhill, J., Hartt, J., & Samols, D. (2005). Levels of comprehension monitoring and working
memory in good and poor comprehenders. Reading and Writing: An Interdisciplinary
Journal, 18, 657–686.
Oakhill, J. V. (1982). Constructive processes in skilled and less- skilled comprehenders’ memory
for sentences. British Journal of Psychology, 73, 13–20.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension-fostering and
comprehension-monitoring activities. Cognition and Instruction, 1, 117–175.
Passenger, T., Stuart, M., & Terrell, C. (2000). Phonological processing and early literacy. Journal
of Research in Reading, 23, 55–66.
Perfetti, C. A., & Hart, L. (2001). The lexical bases of comprehension skill. In D. Gorfien (Ed.), On
the consequences of meaning selection (pp. 67–86). Washington DC: American Psychological
Association.
Perfetti, C. A., Marron, M. A., & Foltz, P. W. (1996). Sources of comprehension failure: Theoretical
perspectives and case studies. In C. Cornoldi & J. V. Oakhill (Eds.), Reading comprehension
difficulties: Processes and remediation (pp. 137–165). Mahwah, NJ: Erlbaum.
Seigneuric, A., & Ehrlich, M. F. (2005). Contribution of working memory capacity to children’s
reading comprehension: A longitudinal investigation. Reading and Writing: An Interdisci-
plinary Journal, 18, 617–656.
Stein, N. L., & Glenn, C. G. (1982). Children’s concept of time: The development of a story
schema. In W. J. Friedman (Ed.), The developmental psychology of time (pp. 255–282).
New York: Academic Press.
Stothard, S. E., & Hulme, C. (1992). Reading comprehension difficulties in children: The role of
language comprehension and working memory skills. Reading and Writing, 4, 245–256.
Stothard, S. E., & Hulme, C. (1996). A comparison of phonological skills in children with reading
comprehension difficulties and children with word reading difficulties. Journal of Child
Psychology and Child Psychiatry, 36, 399–408.
Thorndike, R. L., Hagen, E., & France, N. (1986). Cognitive Abilities Test second edition:
Administration manual. Windsor: NFER-Nelson.
Wechsler, D. (1992). Wechsler Intelligence Scale for Children – III (UK Edition). London:
Psychological.
Yuill, N. M., Oakhill, J. V., & Parkin, A. J. (1989). Working memory, comprehension skill and the
resolution of text anomaly. British Journal of Psychology, 80, 351–361.

Received 4 May 2004; revised version received 25 May 2005