home index by author index by subject

Talk; reading

Saturday stream 5 Session 09.00 - 11.10 Length 25 minutes

What happens when an unfamiliar word is encountered while reading? A metacognitive study of the strategies reported by 7- to 11-year-olds.

John R. Beech

University of Leicester jrb@le.ac.uk

Abstract

Children were asked in two studies about the strategies they used when identifying unfamiliar words while reading. In the first study poor readers was compared with two groups of average readers, one of the same reading age and the other matched by age. The majority of children in all three groups reported having a first preference to use a phonic approach to decode unfamiliar words compared to other strategies. Relatively few expressed a preference for using other strategies such as using context, asking their teacher, or looking up the word in a dictionary. This reported first preference cannot therefore explain reading delay as there was no significant difference between the three groups in the first experiment; however, there was a significant second preference given by poor readers to ask their teacher. This could be an effect of reading delay in that poor readers are placed in closer contact with teachers than the other two groups, or it could be a cause in that teachers may be supplying poor readers with solutions too readily, rather than them helping them to develop self-sustaining decoding strategies. In the second study, which examined 241 children, there was again a strong preference expressed for the phonic approach in two groups divided by age. In both studies there was no significant connection between actual skill in using phonics, as measured by reading nonwords and reported use of phonics. In the second study the older children demonstrated a contrasting sex difference in the reported first choice strategy for decoding unfamiliar words, with significantly more girls than boys preferring to use a phonic approach. This may be due on the one hand to a greater tendency for girls to give responses that they think might be expected of them and on the other for older boys to have less patience with using phonics.

Once children are capable of sustained reading of text their progress will rely on how easily individual words can be read. If there are too many unfamiliar words in the text this can be a substantial stumbling block for further progress in reading any text. Furthermore, if children do not have a means of dealing with these difficult words, it follows that sustained reading will be much more disrupted, less words will be encountered, and in turn, reading vocabulary will not be able to develop as well as it could have done. The present study explored different strategies that children could potentially use when encountering these difficult words when reading text. The potential strategies explored for identifying an unfamiliar word are: (1) letter-sound conversion ("sounding out"); (2) use of analogy; (3) use of context ("reading ahead"); (4) use of a dictionary, and (5) asking for assistance (e.g. from at a teacher or a friend). The first strategy has received the most attention in recent years. The first three strategies would be employed without necessarily incurring much cost in time, however, the last two could prove disruptive to the reading process. This list is not intended to be exhaustive, for example, reading could also involve processing the structural aspects of words (e.g. Lesch & Martin, 1998; Treiman & Zukowski, 1988).

The use of the letter-sound conversion or phonic route is explained by the dual-coding model of reading (e.g. Baron, 1977; Coltheart, 1978; Patterson & Morton, 1985, Rastle & Coltheart, 1999) which proposes that individual words are decoded by two parallel systems. One of these involves a direct visual route to access a word without any phonological involvement. The other route requires some form of phonological processing involving the conversion of letters to sounds, and then blending these together to produce a pronunciation that affords access to the word's meaning. This second system would be used for reading non-words and new unfamiliar words. Dual route theory would predict that as long as the child has developed phonic skills the letter-to-sound route would provide a good basis for dealing with unfamiliar words. This would enable the child to continue reading of text. Jorm and Share (1983) proposed that most words would be read by the direct visual route (depending on text difficulty), but the letter-to-sound route would be used when there is failure. It can, however, create problems if the word is irregular in spelling, as the generated pronunciation may be very different from the real word. An allied problem is that several pronunciations may be generated, especially as word length increases and the choices for breaking down the letter units increase. Share (1995) theoretically elaborated on Jorm and Share's work by suggesting that the letter-sound route acted as a self-teaching mechanism. Recently Share (1999) experimentally demonstrated, in support of the self-teaching hypothesis, that phonological recoding of unfamiliar letter strings provided better learning than pure visual exposure. Children with prelingual hearing impairment, because of their considerable difficulties with phonological processing, by contrast with hearing children matched in reading age, appear to read mainly by the direct visual route (Beech & Harris, 1997). Also, a longitudinal study of 5-year-olds showed that prelingually deaf children make slower progress in learning to read than their hearing peers (Harris & Beech, 1998).

A key factor here is the establishment of phonemic awareness or the ability to recognise individual phonemes within a word. There is evidence (e.g. Bradley and Bryant, 1983) to suggest that training to a reasonable level of phonological awareness helps the development of this 'sounding-out' strategy and subsequent reading. It has been established (e.g. Perfetti & Roth, 1981; Stuart et al., 1999) that the process of learning to read also promotes further development of phonemic awareness and phonic skills. In other words, the relationship between developing phonology and reading ability is bi-directional, thus highlighting the importance of developing the grapheme-phoneme access strategy. Finally, it should be noted that there are critics of dual-route theory. For instance, Metsala, Stanovich and Brown (1998) in a meta-analysis showed there is no evidence for a predicted reduction in the regularity effect (by dual route theory) when those with reading disabilities are compared with reading-matched controls.

The second strategy, the use of analogy to decode words, is another valuable skill encouraged in the classroom that has generated much interest (e.g. Goswami & East, 2000; O'Shaughnessy & Swanson, 2000). Analogical coding involves identifying words by matching part of the orthographic structure to a similarly spelled word within the reader's visual lexicon. This strategy therefore assumes that the visual lexicon already has representations of a number of words. Obviously, the efficacy of analogical coding does not stand up to that of the previously discussed phonological strategy as it is not viable for those beginning readers who have not yet developed a visual lexicon. However, it could be a useful complementary tactic once reading development is under way. Bryant and Goswami (1987) advocated the use of analogy as a potential strategy to teach reading.

A related strategy in the sense that once learned the reader need not rely on extrinsic sources of assistance, is the use of contextual information in order to deduce the meaning of a difficult word. However, Adams & Huggins (1985) suggested that such a tactic is limited in its usefulness, proving to be helpful only if the child is close to establishing identification of the word in question. More negatively Schatz and Baldwin (1986) in three separate experiments found that context did not help students infer the meaning of unknown words.

The last two strategies are dictionary use and asking for help. The availability of a dictionary affords developing readers with a potentially valuable source of information regarding the meaning or spelling of unknown words. Several skills are necessary in order to exploit this information, including a good knowledge of the alphabet. Of course, using the dictionary for spelling purposes also requires that the user should have a reasonable level of phonemic awareness in order to decide with which letters to begin the search.

A final strategy considered here is when a reader asks someone to identify a word for them. Some children will have no qualms about doing this. However, seeking such assistance may not be so easy for others. For instance, a child who is poor at social skills and uncommunicative would find this a difficult option. The provision of a suitably oriented environment, one that encourages such enquiries, may provide some children with access to a resource that would otherwise have been denied. However, there is the possible danger of developing over-dependence. An intelligent child who is used to being supplied by an available adult may not develop other self-help strategies such as letter-sound conversion. Furthermore, as seen in the Share (1999), study showing the advantage of working out a word's pronunciation compared with simple exposure, a child who is simply supplied the word when having difficulties is not developing their decoding strategy that could have developed their sight vocabulary better. One case study, RE (Campbell & Butterworth, 1985) examined a fluent reader but who was unable to read non-words who went to university. In other words, RE had phonological dyslexia with virtually no sub-lexical route but nevertheless she had learned to read. This led to problems such as difficulty in reading chemical names, as reading these required a sub-lexical route. The relevance here is that her mother stayed at home to educate her for several years. Having her mother continuously on hand to help her with new unfamiliar words while reading could have contributed to this lack of development of a sub-lexical route.

Whatever the strategy or combination of strategies that are used to identify unfamiliar words, children need to have some means of sustaining reading. There is evidence that the more practice children have at reading, the better children become. If children have an efficient strategy for decoding words then more text could be read and this in turn would help to develop reading vocabulary. Adams and Huggins (1985) noted that vocabulary expansion is about 3000 words a year. It has been argued that such an expansion cannot be accounted for in terms of direct teaching, such as by the use of flash cards, but must instead be a by-product of reading for interest, such as in books and comics. Stanovich and Cunningham (1992) have suggested that exposure to print in the longer term uniquely confers other benefits such as in helping to expand knowledge, which in turn assists other information processing tasks. However, this reading has to be at the appropriate level as according to Carver (1994) reading for pleasure may not necessarily improve reading vocabulary if this reading is within such a restricted domain that no unknown words are encountered.

In order to overcome the problem of reading text with more unfamiliar words the child needs to be able to have at least one workable strategy for dealing with problem words. Some children may have several. There are two studies in this paper. In the first, three groups of children were selected. The focus here was on a group delayed in reading development who may report having different strategies for approaching unfamiliar words compared with normal readers of their own age and compared with reading-age-matched children, but who are younger. The second study is a comparison of two groups of children of different ages to see if there is a shift in orientation with developing maturity and skill. One major aspect of interest is to see if the level of a child's reading skill and phonic skill determine the kind of strategy that they report using. For instance, children who are behind in reading and who are weak in phonic skills may express a preference for not using a phonic approach with unfamiliar words as they may be less encouraged by the worth of such an exercise. Similarly, as children get older and become more skilled, they may develop a preference for techniques other than phonics, which by that stage they perceive to be too slow in their efforts to gain meaning from print.

Study 1

Method
Participants
Children were selected randomly from primary schools within the county of Leicestershire, UK. There were 39 poor readers, 39 of the same equivalence in reading but much younger than this first group and 31 of the same chronological age as the poor readers but approximately average in reading ability. Table 1 shows the main characteristics of these three groups. A reading quotient is the reading age divided by chronological age multiplied by 100, with a score of 100 indicating that reading age is on a par with chronological age. Younger normal readers matched with older normal readers in reading quotient and the younger normal readers matched the older poor readers in reading age. Reading quotient was lower in the poor readers compared with the other two groups, with the maximum level for the poor readers at 88.7, whereas the minimum level in reading quotient for the younger readers and the older age equivalent groups were 91.5 and 92.6, respectively. Poor readers were 15-50 months behind in reading age. The maximum was 9 months behind for the two other groups. There were 19 males and 20 females in the poor reader group, 18 males and 21 females in the younger group and 13 males and 18 females in the age-matched group. The other tests were given individually (except where stated) and were as follows: British Abilities Scales (BAS) Reading Test (Elliot, 1983). This test has 90 words that are initially short and regularly spelled, gradually becoming longer and more irregular as the test progresses. The child reads each word aloud.

Children's Spelling Test. This test examines the skill to spell words without any benefit of context. It required the completion of 50 partially spelled words, for instance: "n__ise (din). The first 8 words consisted of pictures as clues along with the partial word. The rest had clue words next to them. The score was converted to a percentage correct. This spelling test has high internal reliability (Cronbach's alpha = 0.96). The test was given to about 6 children at a time.

Non-word reading test. Like the spelling test, this was constructed for the study. The children read aloud 64 non-words that gradually increase in difficulty and were 2-6 letters in length (e.g. bo and rakio). Testing was discontinued if there was failure on 10 successive items. The test is designed to test skill in letter-to-sound translation and blending and has high internal reliability (Cronbach's alpha = 0.96).

Raven's Progressive Coloured Matrices (Raven, Court & Raven, 1982). This is a well-established and widely used non-verbal test of intellectual ability. It has 3 sections, each with 12 problems that have to be solved. These increase in difficulty within and between each section. Each problem is presented in the form of a matrix and the child has to select one of six alternatives. The test was given to groups of approximately 6 children and the total score was converted into a percentile with respect to the child's age.

Children's strategies when encountering unfamiliar words. Each child was asked: "What do you do when you are reading a book or something and you come across a word that you can't read?" This question was paraphrased and repeated when necessary. The following categories were annotated as they spoke, so that the first preference received a "1", the next "2", and so on: (a) I try to sound it out (b) I think of a word that's like it (c) I ask my teacher (d) I ask a friend (e) I use a dictionary (f) I work it out from the other words (ie. uses context) (g) I just skip it and go on (h) other. When the children had made a first response, they were encouraged to suggest other ways, but no suggestions for responses were made.

Table 1: Means and standard deviations of poor readers relative to two control groups

All cells are based on 39 poor readers, 39 young readers and 31 age-matched readers. *p<.01 and **p<.001 = significant difference between control group and poor readers by t test.

Table 1 shows the means for these tests of literacy and for the Raven's Matrices. The t-test comparisons between the poor readers and the two other groups confirmed the controlled differences between them. Reading quotient is significantly worse only for the poor readers compared with the two control groups, and reading age is significantly worse for the poor readers and the younger readers relative to the chronological age-matched controls.

Despite these controls the poor readers were significantly better in spelling than the young readers, but as expected, poor readers were significantly worse than the chronological age control group. There was no significant differences in non-word reading, a test of letter-sound translation skills, between the poor readers and the younger readers, but the poor readers were significantly worse than the same chronological age group. This concurs with earlier work by Beech and Harding (1984) and Treiman and Hirsh-Pasek (1985) who found a similar lack of difference in non-word reading when poor readers or those with dyslexia were compared with younger readers of the same reading age. However, according to a review by Rack, Snowling and Olson (1992) more studies have found significant differences. Nevertheless, the poor readers and younger readers are also matched in letter-sound decoding skills in the present study.

Results and Discussion

Table 2: Numbers of children expressing first and second preferences for decoding an unfamiliar word while reading for the three groups in Study 1.

The substantial majority of children in all three groups expressed a first preference for sounding out unfamiliar words in text. Table 2 gives the frequencies of responses to the children's preferences when they could not read a word. The top half of Table 2 shows their first response and other half of the table, their next preferred strategy when asked what else they would do. Thus 69%, 77% and 80% of the poor, young and age-matched readers sounded out the word as their first choice, demonstrating a majority preference for a phonic-like approach in these circumstances. There was no difference across the three groups classifying them by either using a sounding out strategy or using some other strategy, c2(2) = 1.17. Thus this lack of difference between the groups' first preferences rules out this measure as a possible cause for reading delay. When asked to make a second choice, the greatest preference for all three groups was to ask the teacher. There was a much stronger preference to ask the teacher for the second choice among the poor readers and this effect was significant, X²(2) = 8.69, p < 0.05, when classifying the three groups by either asking the teacher or another strategy. This second preference may be a factor in reading delay, but it could equally be an effect of the situation of the poor readers. Because the poor readers were more frequently put in small groups with a teacher during reading, or were simply given more attention by their teachers, this could be the reason why they were more likely to ask their teacher. Alternatively, it may paradoxically be a contributory cause of their poor reading if they ask their teacher rather than work out the word for themselves. The test of non-word reading indicated equivalent levels of decoding skills in the poor and younger readers. Nevertheless, despite the poor readers and younger readers being weaker in phonic skills, they expressed the same degree of preference for the phonic strategy as for the age-matched readers. Their preference for phonics with unfamiliar new words irrespective of their underlying skill could be because they were being explicitly taught to use this strategy by their teachers.

In the next study two groups of children split by age were examined to find out if there was a developmental progression in the type of strategy used when encountering difficult words. In matching the two groups care was taken that the two samples were equivalent in relative non-verbal intelligence as measured by Raven's Matrices and reasonably balanced in gender.

Study 2

Participants
Data from the children in the first study and data from other children who were selected randomly from 15 primary schools within Leicestershire were amalgamated for the purposes of the second study. Two groups were formed which either had age ranges between 7 years 0 months to 9 years 5 months or between 9 years 6 months to 11 years 4 months. Initially 270 children were tested, however, in order to balance group performance on Raven's Matrices between the two age groups this sample was reduced to 241. The younger group had 124 children consisting of 62 males and 62 females and the older group of 117 children had 57 males and 60 females. The performance of the two groups on the major variables is shown in Table 3.

Table 3: Means and standard deviations of younger and older readers

Cells are based on 124 younger readers and 117 older readers. **p<.001 = significant difference between the two groups by t test.

Procedure and design

All subjects were tested on the same variables described in Experiment 1. Two groups of children divided by age, as previously described, were compared.

Results and discussion

Table 4: Numbers of children expressing first and second preferences for means of decoding an unfamiliar word while reading divided by age in Study 2.

Table 4 shows the first and second preferences for the strategy used when encountering an unfamiliar word. As in the previous study the majority preference for the first response was to sound out the word and the majority second preference was to ask the teacher. This was the case for both age groups with 68% and 71% of the younger and older groups, respectively, choosing to sound out as their first choice.

Table 5: Numbers of children who sound out unfamiliar words for first preference vs. the rest divided by age and sex in Study 2.

A further analysis was undertaken dividing the children within each group into those who sounded out for their first preference versus the rest, further subdivided by gender and Table 5 reveals an interesting sex difference interacting with age group. The younger children had an even split between the sexes in preferences, whereas the older children were unevenly divided. The great majority of the girls (85%) had preferred sounding out as their first choice, but proportionally fewer of the males (57%) expressed a first preference for sounding out. This interaction between the sexes for the older children was significant, X²(1) = 11.4 p < .01. Differences between these two groups (sounding out as first choice vs. the rest) on reading quotient, age, sex, Ravens, spelling and non-word reading revealed some differences between the groups. Within the younger group, the "sounding out" group were significantly worse in spelling (24.6 vs. 32.5), t(120) = 2.16, p < .05. In the older group, the "sounding out" group were significantly older (10.62 vs 10.29 years), t(116) = 2.95, p < .01; but there were no differences in spelling performance (49.5% vs 47.2%, t < 1). There does not seem to be any sound a priori reason why there should have been this significant age difference. This weakness in spelling in the younger group suggests that early readers who have a stronger preference for a phonic strategy may over-extend this strategy in their spelling. This would imply a weakness for spelling irregularly spelled words, as they would be concentrating more on a sound-to-letter conversion rather than the pattern or letter sequence. The marked sex difference in preferences for phonics in the older group indicates a difference in approach between the sexes as reading develops. One possibility is that once males become more skilled in reading they may become impatient with phonic strategies in their quest to understand the meaning of a text. Within the older groups, divided between sounding out versus the rest, non-word reading ability was very similar (75.4% vs. 74.0%, t < 1) indicating no difference in letter-to-sound translation skills.

General Discussion

One major consistent finding across both studies has been how most children report that as a first preference they identify an unfamiliar or difficult word in text by breaking it down and sounding out its constituent sounds. This does not alter with reading age nor with differences in reading skill. This is interesting as one might have expected that as phonic training can improve reading performance (e.g. Adams, 1990) that there might be less inclination to use phonics in those who are less skilled in reading and in those who are behind in reading. Furthermore, by testing non-word reading performance we can examine any potential interaction with the skill of phonics. One might expect that those who are weak in alphabetic coding would be less willing to use this strategy. But again, no such relationship was forthcoming. One significant relationship was that rather than helping performance, a reported preference for using phonics as a first attempt in reading was associated with slightly weaker spelling performance in the younger cohort in the second study.

Another finding was that in the older group significantly more girls than boys said that the alphabetic approach was a first choice for decoding unfamiliar words. This might suggest that as soon as they are able, boys prefer not to use what (to them) might be perceived as a slow process to get to the word's meaning. There was no significant difference in non-word reading between the groups, so this is a sex difference in their reported preference to use phonics rather than a difference in skill in using phonics. Girls find it socially more acceptable to work for examinations than boys in Britain (Warrington, Younger & Williams, 2000). This suggests that girls are influenced by what is more socially acceptable. In the same way, girls give more socially acceptable responses and perhaps thought that the expected answer in the present study would be that they should respond that they sound out unfamiliar words while reading. There are differences between the sexes in that boys tend to prefer to read non-fiction whereas girls enjoy fiction. Dawson (2000) found that Year 7 Australian girls had a more negative attitude to the physical sciences and that this was the same attitude that had been found 17 years earlier, despite subsequent improvements to science education. It is possible that reading in different domains (fiction vs. non-fiction) may need different requirements for letter-sound translation.

A corollary to the strong preference for the use of phonics by all groups is that other approaches do not get chosen very much. For example, some investigators have suggested a role for the use of context in helping the decoding of unfamiliar words in text. Stanovich (1981) has suggested that as early word recognition is such a slow process, some will rely on the use of context and try to guess the word rather than use an alphabetic approach. Goodman (1970, 1985) has suggested that fluent readers make more use of context; however others (e.g. Perfetti & Roth, 1981) have shown that contextual guessing is likely to slow down the more fluent reader.

It should be emphasised that the present study is examining what happens when reading breaks down and there is a problem caused by an unfamiliar word. Levy (1999) has reported an experiment that appears to lead to a different conclusion to that arrived at by Share (1999) described earlier. Levy has shown how many repetitions of words presented on a computer screen can subsequently help word naming time at a later point. The effect occurred for both good and poor readers irrespective of spelling regularity indicating that alphabetic coding was not an essential component of the repetition task. In the classroom (and home) frequent reading practice is needed in order to help reading performance by helping automatic word retrieval. Given this context, the use of alphabetic codes may be only necessary to maintain the flow of information when awkward words are encountered. An efficient alphabetic identification of the word without being too disruptive should help the speedy resumption of the reading of the passage.

In conclusion, both experiments show that most children report that they prefer to identify an unfamiliar or difficult word in text by breaking it down and sounding out its constituent sounds. This preference does not alter with reading age nor with differences in reading skill relative to age. As phonic or segmentation training has been shown to improve subsequent reading performance in many studies, one might have expected there to be less reported inclination to use phonics in younger children who are less skilled in reading and in those older children who are behind in reading. Similarly, one might expect the reported preference to use phonics to be related to skill in the use of phonics, as measured by non-word reading. However, none of these expectations was realised. Most children irrespective of underlying skills say that they prefer to use phonics in decoding unfamiliar words as a first approach. This reported first preference cannot therefore explain reading delay as there was no significant difference between the three groups in the first study; however, there was a significant second preference given by poor readers to ask their teacher. On present results this could be an effect of reading delay in that poor readers are placed in closer contact with teachers than the other two groups, or it could be a cause in that teachers may be supplying poor readers with solutions too readily, rather than encourage them to develop self-sustaining decoding strategies. The other notable finding has been that a substantial proportion of older boys may be impatient with segmenting difficult words and instead prefer to move on in the text in order to continue with what they are reading. We should end on a cautionary note: the strategies for working out unfamiliar words in text reported by these children may not be an accurate reflection of the actual cognitive processes they go through, especially considering the occasional mismatch on tests such as nonword reading and the reported use of sounding out. A more detailed analysis will be required to get closer to this problem.

References

Adams, M.J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press.

Adams, M.J., & Huggins, A.W.F. (1985). The growth of children's sight vocabulary: A quick test with educational and theoretical implications. Reading Research Quarterly, 20, 262-281.

Baron, J. (1977). Mechanisms for pronouncing printed words: Use and acquisition. In D. LaBerge & S.J. Samuels (Eds.), Basic processes in reading: Perception and comprehension (pp. 175-216). Hillsdale, NJ: Erlbaum.

Beech, J.R. & Harding, L.M. (1984). Phonemic processing and the poor reader from a developmental lag viewpoint. Reading Research Quarterly, 19, 357-366.

Beech, J. R. & Harris, M. (1997). The prelingually deaf reader: A case of reliance on direct lexical access? Journal of Research in Reading, 20, 105-121.

Bradley, L., & Bryant, P.E. (1983). Categorizing sounds and learning to read: A causal connection. Nature, 301, 419-421.

Bryant, P., & Goswami, U. (1987). Phonological awareness and learning to read. In J.R. Beech and A. Colley (Eds.), Cognitive approaches to reading (pp. 213-243). Chichester, England: Wiley.

Campbell, R., & Butterworth, B. (1985 ). Phonological dyslexia and dysgraphia in a highly literate subject: A developmental case and associated deficits of phonemic processing and awareness. Quarterly Journal of Experimental Psychology, 37A, 435-475.

Carver, R.P. (1994). Percentage of known vocabulary words in text as a function of the relative difficulty of the text: Implications for instruction. Journal of Reading Behavior, 26, 413-437.

Coltheart, M. (1978). Lexical access in simple reading tasks. In G. Underwood (Ed.), Strategies of information processing. London: Academic Press.

Dawson, C. (2000). Upper primary boys' and girls' interests in science: Have they changed since 1980? International Journal of Science Education, 22, 557-570.

Elliot, C.D. (1983). British ability scales. Windsor, England: NFER-Nelson.

Goodman, K.S. (1970). Reading: A psycholinguistic guessing game. In H. Singer and R.B. Ruddell (Eds.), Theoretical models and processes of reading. (1st. edition, pp. 259-272).

Goodman, K.S. (1985). A linguistic study of cues and miscues in reading. In H. Singer and R.B. Ruddell (eds.), Theoretical models and processes of reading. (3rd edition, pp129-134).

Goswami, U & East, M. (2000). Rhyme and analogy in beginning reading: Conceptual and methodological issues. Applied Psycholinguistics, 21, 63-93.

Harris, M. & Beech, J.R. (1998). Implicit phonological awareness and early reading development in prelingually deaf children. Journal of Deaf Studies and Deaf Education, 3, 205-216.

Jorm, A.F., & Share, D. L. (1983). Phonological recoding and reading acquisition. Applied Psycholinguistics, 4, 103-147.

Lesch, M.F. & Martin, R.C. (1998). The representation of sublexical orthographic-phonologic correspondences: Evidence from phonological dyslexia. Quarterly Journal of Experimental Psychology, 51A, 905-938.

Levy, B.A. (1999). Whole words, segments, and meaning: Approaches to reading education. In R.M. Klein and P. McMullen (Eds.) Converging methods of understanding reading and dyslexia. Cambridge, Mass.: MIT Press.

Metsala, J.L., Stanovich, K.E., & Brown, G.D.A. (1998). Regularity effects and the phonological deficit model of reading disabilities: A meta-analytic review. Journal of Educational Psychology, 90, 279-293.

O'Shaughnessy, T.E. & Swanson, H.L. (2000). A comparison of two reading interventions for children with reading disabilities. Journal of Learning Disabilities, 33, 257-277.

Patterson, K.E. & Morton, J. (1985). From orthography to phonology: an attempt at an old interpretation. In K.E. Patterson, J.C. Marshall and M. Coltheart (Eds.), Surface dyslexia: neuropsychological and cognitive studies of phonological reading. London: Erlbaum.

Perfetti, C.A., & Roth, S.F. (1981). Some of the interactive processes in reading and their role in reading skill. In A. Lesgold and C.A. Perfetti (Eds.) Interactive processes in reading. Hillsdale, NJ.: Erlbaum.

Rack, J.P., Snowling, M.J., & Olson, R.K. (1992). The nonword reading deficit in developmental dyslexia: A review. Reading Research Quarterly, 27, 28-53.

Rastle, K. & Coltheart, M. (1999). Serial and strategic effects in reading aloud. Journal of Experimental Psychology: Human Perception and Performance, 25, 482-503.

Raven, J.C., Court, J.H., & Raven, J. (1982). Manual for Raven's progressive matrices and vocabulary scales. London: Lewis.

Schatz, E.K. & Baldwin, R.S. (1986). Context cues are unreliable predictors of word meanings. Reading Research Quarterly, 21, 439-453.

Share, D.L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition, 55, 151-218.

Share, D.L. (1999). Phonological recoding and orthographic learning: A direct test of the self-teaching hypothesis. Journal of Experimental Child Psychology, 72, 95-129.

Stanovich, K.E. (1981). Attentional and automatic context effects in reading. In A.M. Lesgold and C.A. Perfetti, (Eds.), Interactive processes in reading (pp. 341-267). Hillsdale, NJ: Erlbaum.

Stanovich, K.E., & Cunningham, A.E. (1992). Studying the consequences of literacy within a literate society: The cognitive correlates of print exposure. Memory & Cognition, 20, 51-68.

Stuart, M., Masterson, J., Dixon, M., & Quinlan (1999). Inferring sublexical correspondences from sight vocabulary: Evidence from 6- and 7-year-olds. Quarterly Journal of Experimental Psychology, 52(A), 353-366.

Treiman, R., & Hirsh-Pasek, K. (1985). Are there qualitative differences in reading behavior between dyslexics and normal readers? Memory & Cognition, 13, 357-364.

Treiman, R. & Zukowski, A. (1988). Units in reading and spelling. Journal of Memory and Language, 27, 466-477.

Warrington, M., Younger, M. & Williams, J. (2000). Student attitudes, image and the gender gap. British Educational Research Journal, 26, 393-407.