Robert Savage of McGill University, Canada, reports on his research in the UK, which found two ways for classroom assistants to aid early reading development and intervention.

For many years applied reading researchers have sought to develop methods to identify children who are struggling with literacy and to find ways to intervene early to prevent reading related problems. In a series of recent studies, we have attempted to develop schools' self-sufficiency in early identification and in brief, but differentiated, preventative early intervention in Year 1.

Much recent reading research has focused on the role of phonological awareness. Researchers have used the term 'phonological awareness' to describe a range of tasks requiring children to manipulate the sounds of spoken language. For example, children are asked to identify or manipulate the speech sounds associated with 'c' - 'a' - 't' in 'cat'; or to say the word 'cat' without the sound associated with 'c'; or to identify spoken word pairs that rhyme (e.g. 'cat / hat' word pairs versus 'cat / can' word pairs).

Research has shown us that, alongside decoding skills, tests of 'phonological awareness' are amongst the best predictors of literacy available. However, the usefulness of such measures alongside current school screening practice ('baseline' assessment, pupil background data) and to National Curricular outcome measures was not known. We therefore explored how useful phonological awareness and early reading measures might be alongside baseline assessment, and pupil background factors such as gender, free school dinners, and SEN status at age five in the prediction of English, maths, and science performance. The study covered 435 children in key stage 1 of the National Curriculum in England.

Importantly, we were also interested in whether school staff could carry out such tasks as reliably as researchers had done, as no research existed on this issue either. We therefore trained schools' own classroom assistants (CAs) to administer the early literacy and phonological awareness measures.

Analyses of results revealed that phonological awareness and decoding measures were excellent predictors of key stage 1 levels two years later. Furthermore, if we combined this measure with important pupil background measures such as baseline skills, SEN status and gender there was an even higher degree of prediction. We were able to correctly identify whether children would achieve 'average' attainment levels two years later 90 per cent of the time. Encouragingly, even in our most preventatively-oriented schools, staff told us that measures helped them identify children they might have missed. Finally, in our most recent studies (Savage and Carless, in press a) we have confirmed that the most important phonic and decoding unit for CAs to screen is phoneme manipulation (for example, segmenting 'cat' into 'c'-'a'-'t' or blending such units into the word 'cat').

We concluded therefore that CAs can be readily trained to administer phonological awareness measures to cohorts of five-year-old children. Such measures used in conjunction with baseline measures, and other within-school measures such as SEN status, significantly enhance prediction of key stage 1 performance. Practically, deploying CAs in this fashion helps a school to identify young children for whom additional support may be necessary. Appropriately trained and supported, classroom assistants can thus play a significant role in helping develop preventative approaches in Year 1.

Intervention studies

Having identified children who might experience difficulties we wanted to help schools implement a plan of action. There is now a very large body of research evidence suggesting that training children to adopt analytic phonologically-based word-level strategies in early reading leads to improvements in reading and spelling ability. Ehri, Nunes, Willows, Schuster, Yaghoub-Zadeh and Shanahan (2001) have recently analysed some 52 intervention studies confirming the reliability of this finding.

More specifically, the evidence from Ehri et al suggests that preventative training of very young children to attend to the phonological structure of language has both immediate and lasting effects on word identification through the primary years. The Ehri et al analysis shows that such training is particularly effective when also combined with teaching of letter-sound correspondences, where children also learn the 'linkage' or connection between these letter-sound units, and interestingly, when such work is undertaken in small groups of three to four children, rather than in one-to-one tutoring.

Formal intervention studies like those described by Ehri et al are, by their nature, high-profile, well-resourced, often using experienced staff. Such research-led scenarios are clearly very different from those in regular schools. In other words, this research relies on evidence from well-designed scientific studies using 'random controlled designs' and often marked by high levels of control. In contrast, we investigate the effectiveness of reading interventions in 'real world' settings. That is, we utilise research with minimal change in routine practice and resources, under circumstances that frequently occur in schools. Such research may possibly have greater external validity, being more generalisable to real-world settings.

Effective reading interventions

In the current English context, adult CAs commonly work with small groups of poor readers, helping children to meet differentiated 'word-level' reading and spelling targets that relate to whole-class teaching themes. CAs, while often very skilled and experienced, frequently have only basic (and sometimes no) formal training. Despite the crucial role they have in supporting 'at-risk' children in schools, and proposed developments in their role in schools, the actual contribution of CAs has been little-explored, and little formal evidence exists of the effectiveness of such practices.

Finally, we also looked at the relative effectiveness of different methods for teaching early word decoding, contrasting rhyme-based intervention programmes (for example, using familiar words such as 'hat', 'cat' or 'fat' to decode unfamiliar words such as 'pat' or 'chat') and phoneme-based interventions (for example, blending sounds associated with 'c'-'a'-'t' in 'cat').

In our first intervention study, Savage and Carless (in press b) sought to explore whether CAs can also improve performance using these methods. The poorest reading 20 per cent of all six-year-old children in nine schools were screened and selected for this study. CAs were briefly trained to administer phonic programmes to these children as small-group interventions over nine weeks. Rime and phoneme-based programmes were contrasted with a control group receiving the National Literacy Strategy only.

After the intervention, we found that all children in the two CA-taught intervention groups were better 'decoders' of words and unfamiliar made-up words. They also had better phonological awareness and letter-sound knowledge than classroom-taught controls. The phoneme-based group had better letter-sound knowledge than the other intervention groups.

In order to confirm this finding we ran the same intervention in nine different schools the following year (Savage, Carless and Stuart, 2003). Essentially, we again found reliable effects for decoding and letter-sound knowledge for all intervention groups. We therefore concluded that briefly-trained CAs can deliver effective early preventative programmes for literacy difficulties. We were unable to find any very clear advantages for phoneme versus rime-based approaches. One reason for this may be that, overall, despite the different teaching approaches used, all children who benefited from the interventions seemed to be using phoneme-based decoding skills (Savage, 2004).

Classroom practice

Overall, we feel that there are a number of practical implications of our research. Firstly, we would argue that while successful reading involves many more skills than we assessed, a simple assessment of phoneme segmentation and blending, letter-sound knowledge and basic decoding skills is useful early in Year 1. This information can, especially when placed alongside pupil background data, help to identify potentially at-risk poor readers. Children who are unable to segment and blend many simple short words (like segment dog into 'd'-'o'-'g' or to blend 'c'-'a'-'t'- into 'cat'), who know few letter-sounds, and who cannot decode any simple made-up 'words' such as 'zif' and 'jub' by November or December of Year 1, despite full opportunities to develop reading in whole-class settings, might be at risk of reading problems if not supported preventatively for the rest of the school year.

Secondly, we have shown in two studies that CAs can meaningfully contribute to the development of literacy skills for these sorts of children. As school support professionals (educational psychologists and literacy support team staff respectively), our model of training was to train CAs to use existing reading packages that were augmented by phonological activities. It is generally assumed that working with CAs is a good way to further aid inclusive practices. We would argue that professional support for the development of CAs can, alongside excellent classroom teacher practices, be a highly effective model for facilitating a schools' capacity to identify and prevent reading difficulties, and thus, can aid inclusion.

References

L.C. Ehri, S.R. Nunes, D.M. Willows, B.V. Schuster, Z. Yaghoub-Zadeh and T. Shanahan (2001) Phonemic awareness instruction helps children learn to read: evidence from the national reading panel's meta-analysis. Reading Research Quarterly, vol. 36, no. 3, pp. 250-287.
R.S. Savage (2001) A re-evaluation of the evidence for orthographic analogies: a reply to Goswami (1999) Journal of Research in Reading, vol. 24, pp. 1-18.
R.S. Savage (2004) Predicting growth of non-word reading and letter-sound knowledge following rime and phoneme-based teaching. Journal of Research in Reading, vol. 27, no. 3, pp. 195-211.
R. Savage, S. Carless and M. Stuart (2003) The effects of rime- and phoneme-based teaching delivered by learning support assistants. Journal of Research in Reading, vol. 26, no. 3, pp. 211-233.
R. Savage and S. Carless (2004) Predicting curriculum and test performance at age years from pupil background, baseline skills and phonological awareness at age 5. British Journal of Educational Psychology, vol. 74, pp. 155-171.
R.S. Savage and S. Carless (in press a) Phoneme manipulation but not onset-rime manipulation is a unique predictor of early reading. Journal of Child Psychology and Psychiatry and Allied Disciplines.
R.S. Savage and S. Carless (in press b) Learning Support Assistants can deliver effective reading interventions for 'at-risk' children. Educational Research, vol. 47, no. 1.

Subscribe to Literacy Today