Reading Recovery (RR) is the trade mark of a remedial reading programme developed by Marie Clay in New Zealand in the 1970s. It is a whole language programme, developed in response to falling levels of literacy engendered by whole language teaching, and has a primary emphasis on reading for meaning. It has eschewed, or marginalised, the teaching of systematic, explicit phonics as part of learning to read. There has been a substantial amount of criticism, by the international reading research community, of the RR programme and the claims made for its effectiveness, e.g. Grossen, Coulter & Ruggles, Tunmer and Chapman and others . Critics point out that RR cannot produce any independent research to back its claims.
Explicit phonics instruction
Nancy Salvato, in the US, reports  that Marie Clay, in 2001, refused to modify RR to include explicit, systematic phonics instruction in order to make it eligible for Reading First funding in the US. Instead, she stated that she would rewrite the description of RR’s components in such a way as to bring it into compliance with the Reading First legislation.
Reading Recovery in the UK
Reading Recovery has been used in the UK for many years. However, it is a very costly intervention and fell out of use as school budgets were squeezed. The RR network in the UK is based at the University of London’s Institute of Education. In 2005 the RR network obtained funding of £10m from a charitable foundation, and support from the DfES, to implement RR in UK schools under the title ‘Every Child A Reader’ (ECAR). In the first year, 2005/06, RR was rolled out in 61 schools. ECAR plans to fund 255 schools in 2006/07 and 310 in 2007/08 . In his November Budget Statement Gordon Brown (Chancellor) announced that a further £10m of government money would be put into the project. He has stated that he would like to see RR in every school. To put the current coverage of RR into perspective I should point out that there are some 24,000 KS1/2 schools in the UK. To put the costs into perspective, if (as the ECAR report says) it cost £1m to fund 61 schools in the first year, then it would seem that it would cost around £400m to roll this out across the UK.
Reading Recovery in the UK enjoys a very high profile as a result of its 2005 funding coup and further encouragement from the Chancellor. Has this funding enabled RR to maintain that high profile with Parliamentary receptions, extensive media coverage and the production of weighty reports?
Its stated aims are highly laudable; to ensure that every child leaves school able to read, and there is no doubting this sincerity. To the general public, the government and the educational establishment, from academics to teachers, RR presents no problem. The RR methodology is familiar, as it comprises the Balanced Literacy methods for so long practised in schools and promoted by the DfES. So why is the Reading Reform Foundation objecting to this apparently worthwhile project?
At the time that the then Education Minister, Ruth Kelly, announced the inception of the ‘Every Child A Reader’ programme, The Rose Review of the initial teaching of reading was underway. This review was requested by the House Of Commons (HOC) Select Committee on Education and Skills. The Select Committee’s inquiry was, in its turn, a response to growing concern that, despite the inception of the National Literacy Strategy in 1998, standards of literacy were demonstrably not rising as expected. An initial apparent improvement had levelled out at around 80% of children reaching the required standard and, despite a spate of DfES intervention programmes, had not moved.
The Rose Report was published in March 2006. Jim Rose and his review panel undertook exhaustive research into the subject (details of the evidence considered can be found in Appendix 23 of a document submitted to the HOC Select Committee on Science and Technology  ). The Report recommended that the NLS be revised in the light of the findings of the Review body. Subsequent guidance has been published in the Primary Framework for literacy and mathematics. (Supporting guidance to head teachers and chairs of governors. DfES document ). This document sets out the criteria for high quality phonic work and contains statements which are pertinent to the use of RR as an intervention.
“Teaching needs to be staged so that priority is given to the development of word-reading processes in the early stages of learning to read...”
“The Primary National Strategy distinguishes three ‘waves’ of teaching and intervention which adequately cover the range of provision that best supports children with significant literacy difficulties. It is important to recognise that these ‘waves’ signify types of provision and not categories of children. High quality phonic work, as defined by the review, should be a key feature of the provision in each of these ‘waves’:”
Synthetic phonics practitioners had every right to be reasonably satisfied with the conclusions of the Rose Report, and the subsequent moves towards its implementation by the DfES. However, it soon became clear that there was a startling contradiction between the actions of the DfES in, on the one hand, implementing the Rose Report and, on the other hand, actively promoting the use of intervention programmes such as RR which directly contravene the principles of ‘high quality phonics instruction’. Not only do they appear to only pay lip service to phonic work, but they also appear to promote the very searchlights strategies (use of context, initial letters and pictures to ‘predict’ words) which were heavily implicated in the failure of children to learn to read. Marilyn Jager Adams  concluded that the 3 cueing system, on which the searchlights model was based, was, in effect, an educational myth. Dr. Morag Stuart  commented “The model of reading which is presented to teachers (in the NLS) which is this black hole of four things (i.e. four ‘searchlights’) operating and disappearing into a text is completely and utterly misleading and bears no relation to any research on reading that I know of”.
Children in need of extra help in learning to read are still beginning readers. They still need intensive extra phonics instruction, ideally reinforced by the use of decodable texts, to promote fluent decoding and engender confidence. These children are still ‘in the early stages of learning to read’ when priority should still be given to word reading processes (says the Rose Report). It is incredible that children who are at this stage are being offered RR, with its eclectic mix of phonics, ‘other strategies’ and book banded texts.
How does Reading Recovery teach reading?
I have analysed the description of a typical session given in the Every Child A Reader report and an article on ‘phonics’ by Sue Bodman, RR Coordinator/National trainer in the Nov 2006 edition of the RR Newsletter, The Running Record.
Strategies for identifying words
The Rose Report makes it clear that the only strategy to be taught to beginning readers for identifying words is decoding and blending. So what do we make of this passage from the description of a RR session given in their report on Every Child A Reader?:
“They (the teacher) ..encourage the child to identify their own independent reading strategies:
What was it you did on that page to help you to work out the word...?
Can you find a page where you got stuck and then found a way of sorting things out? What was it that helped you?
How did you know that it said ‘brothers’ not ‘kids’? Why isn’t this ‘box’?”
“The teacher might also ask the child to ‘take a word to fluency’, practising a common word over and over, tracing it in sand, or salt, or multicoloured pens...........The teacher might set a target for the child to use an independent reading strategy.”
Sue Bodman writes of ‘providing a context for working with a range of word reading strategies”.
This very clearly implies that the RR teacher is teaching more than one ‘strategy’ for identifying words. Would any teacher of ‘high quality phonics’ recognise these procedures as those that they would use when teaching struggling children to work out what a word is?
RR is claiming that it complies with the requirements of the Rose Report. The ECAR report states that the scheme is: “Using the opportunity presented by international changes to RR’s methodology in its revised core texts to take a fresh look at the role of synthetic phonics in the scheme, ensuring consistency with the developments underway (sic) as a consequence of the government’s Rose Review of the early teaching of reading”.
And here is Sue Bodman in The Running Record: “Phonics has always been an important part of Reading Recovery’s success with the most vulnerable literacy learners.”
But Marie Clay, in 2001, “refused to modify Reading Recovery to include explicit, systematic phonics instruction in order to make it eligible for Reading First funding in the US”. Has there been a change of heart? Or, as with Clay’s ‘rewriting of the description of the programme’s components to bring it in line with the Reading First Legislation’, is RR UK adapting their terminology in order to comply with the requirement for ‘high quality phonics’?
The ECAR report session description:
“...Then there will be work with magnetic letters on a whiteboard to develop phonemic and phonological skills. The child will compose and write their own sentence or story, analysing the sounds in words in order to write them.”
Sue Bodman: “‘Making and breaking’ is part of the Reading Recovery lesson that involves children in the manipulation of magnetic letters to notice connections between words, clusters of letters or parts of words”. Discovery learning? Note that the child is not taught.
Ms Bodman’s description of the phonic work is worrying: “We work with the few words or letters that the child knows, demonstrating how they look similar to or different from other letters. ‘Learning to look at print’ (Clay 1993) is a specific teaching procedure in Reading Recovery, where the child is taught to attend to print features in helpful ways, ‘since the beginning reader can only sample some features in print which he recognises’ (Clay 1993 p. 23).”
With high quality phonics, a child does not only sample some features in print, and does not work solely with ‘...letters that the child knows’. Instead the child is taught systematically and explicitly how a letter, or letter combination, represents a sound, and consolidates and builds on that knowledge from session to session. Children are not given text in which they can only recognise a few sounds (sample some of the print features) but texts which contain graphemes they have been taught, on which they can practice their knowledge in a meaningful way to sound out and blend all the words.
The Rose Report makes it clear that the phonic skills should be taught progressively:
“In practice, this means teaching relatively short, discrete daily sessions, designed to progress from simple elements to the more complex aspects of phonic knowledge, skills and understanding.”
I can find no mention of progression in the ECAR report, or Ms Bodman’s article.
While the Rose Review does not limit text for early reading to decodable text, it has this to say:
“There is some force in the view that, as they learn to master the alphabetic code, children should be given reading material that is well within their reach [bold added] in the form of 'decodable books', that is to say, early reading books specially designed to incorporate regular text, which children can decode using the phonic skills they have secured. The view is that this enables them to benefit from 'quick wins' in practising phonic skills and gaining confidence from reading a whole, albeit short, book. [bold added] Using such books as part of the phonic programme does not preclude other reading. Indeed it can be shown that such books help children develop confidence and an appetite for reading more widely.”  (para. 82)
The children who RR are working with are still mastering the alphabetic code and are precisely those who will benefit most from decodable texts. RR uses its whole language/ ‘look and say’ based ‘levelled’ books for their intervention. Children with Reading Ages of below 6 are presented with complex words, such as ‘amazing’, ‘grandpa’, and ‘concert’ , These require a level of phoneme/grapheme knowledge far in excess of what these children will know. The only way that these books can be ‘read’ by these children is if the teacher tells them these words, or expects them to learn them as ‘whole words’. This practice is not only incompatible with high quality phonics teaching, but is known to be damaging, as whole word learning activates a different area of the brain from all through the word decoding. This area is known from MRI brain imaging to be activated by ‘dyslexic’ readers. It also causes the children to be confused as to which strategy they should use for unfamiliar words. They should only be using one strategy, decoding.
If the children are being ‘told’ the words they are unable to read, what sort of reading experience are they getting? How frustrating must it be for them to be unable to read some of the words on the page and how discouraging to be presented with word that they have no chance of decoding successfully, even if they can recognise a few of the letters. And how can the constant halts at undecodable (for them) words and interjections from the teacher help with understanding the text?
There is still an insistence that the most important feature of learning to read is making meaning. This directly contravenes the view of the Rose Report that learning to recognise words is initially of more importance than understanding text.
According to Ms Bodman it is important that ‘analysing’ words does not slow down the reading and cause loss of meaning. High quality phonic work involves careful decoding and blending of each word first, then re-reading of the sentence for comprehension. This is an important difference in approach. Careful decoding and re-reading helps to embed the word in long term memory. Rushing through the decoding process in order to preserve ‘meaning’, will do nothing to help the child read the word more easily next time it is encountered.
Even if some attention is given to decoding, the old, deeply flawed, strategies remain firmly in place. “Children will need to move through the teaching sequence individual to each learner that encompasses attending to initial and final letters, inflection breaks, syllable breaks and prefix or suffix breaks.” (Syllables, prefixes and suffixes? Hey, these are Year 1 struggling readers!) Another essential element of high quality phonics is that a child learns to decode and blend all through the word and does not ‘dot about’ in the word looking for bits it recognises. This is not only ineffective and muddling for the child, but it also prevents them developing efficient left to right tracking through each word, which is an essential skill for fluent reading.
We at the Reading Reform Foundation are deeply disturbed by the fact that the DfES, and the Government, in the person of Gordon Brown, are actively promoting this intervention, which seems to be continuing to use the mix of methods which characterised the failed NLS. They are promoting it to the extent that it is seen as desirable that it should be available in all primary schools. The DfES has even gone so far as to issue a booklet, entitled ‘Learning from Every Child A Reader, which advises the use of RR as an intervention. The Reading Reform Foundation would like to ask if the person responsible for issuing this booklet was aware that Reading Recovery teaching of reading is not in accordance with that set out in the new Primary Framework. What is even more worrying is the declared intention of RR to ‘embed’ their practices in the whole school approach to literacy, through the person of the trained RR teacher in school.
From the Every Child A Reader report: “A particularly effective innovation within Every Child A Reader project has been the development of a year long course for experienced RR teachers, aimed at providing them with the skills they need to influence literacy practice at whole school level. This year, 13 teachers from all the London Boroughs receiving funding took part in the course, jointly tutored by staff at the Institute of Education and members of the Primary National Strategy literacy team.”
Why are members of the Primary Strategy literacy team involved in training for RR, when its methods do not conform to the Rose Report recommendations?
The Reading Reform Foundation is also concerned that the promotion of RR, and similar programmes such as Catch Up, shows no confidence in the expected success of the implementation of high quality phonic work in all primary schools. It seems that there is an expectation of failure, before this has been fully implemented. It may have been appropriate to intensify intervention in the short term, for children who have already been failed by the NLS, but all the signs are that RR means to manoeuvre itself into an unassailable position (with help from the DfES and Gordon Brown) and expand wherever possible. Has the DfES/government not taken note of the success of the Dunbartonshire project to eliminate reading failure with the use of synthetic phonics for initial teaching of reading and for remediation? Dunbarton reported only 6% of children at below the expected level last year . Are there elements within the DfES and the government which do not believe that England can achieve the same success with synthetic phonics?
What is the evidence of the ‘success’ of Reading Recovery in the UK?
In November 2006 Every Child A Reader published an ‘evaluation’ of the performance of the first year of the ECAR project . It showed that children in the project, on the whole, made excellent gains compared with a control group.
However, closer examination of the evaluation throws up a number of questions.
The first, and maybe the most pertinent, is the question of the control group. As Diane McGuinness noted, in a brief analysis , “... there are control groups and control groups. The RR children received extensive one-to-one tuition during their second year at school, and the so-called control group, children without RR, got no consistent treatment. In fact, the description of what they received is so vague as to be non-existent”.
It would be extremely surprising if a child receiving daily one-to-one support with a highly trained teacher failed to make any progress at all. Comparing this input with very little intervention, or none at all (93 of the ‘control group’ of 147 did not receive any intervention) is bound to show improvement in the intervention group, unless it is a very bad intervention indeed. Surely the results would have been more meaningful if the control group had also received a daily one-to-one intervention, delivered by a highly trained teacher or teaching assistant. As it is, it the evaluation serves mainly to show that any intensive intervention is better than little or no intervention.
Professor McGuinness also points out that, at pre-testing, the children scored an average of 100 standardised score (normal) and then are described as virtual non readers for the age at time of testing. She points out that both cannot be true. So, exactly how ‘poor’ were the children selected for the programme?
What is not mentioned in the RR Evaluation is the numbers of children who were ‘referred on’ for further intervention at the completion of the programme. This figure is reported by ECAR as 86 children (23% of the 373 children in the ECAR programme nationally in 2005/06). There must have been some of these children among the 145 London Borough children surveyed in the evaluation.
There is no standardised spelling test used in the evaluation. Instead, the children are tested on a measure of ‘Writing Words’ (p.8), with no further clarification. It is stated that at the end of the programme the RR children could write ‘more than 45 words’ at the end of the programme (p.14). It further says that ‘Children able to write correctly around 45 frequently occurring words have become fluent writers for their age’ (p.14). This leads one to believe that the children were tested on writing the 45 YR/1 High Frequency Words. Not only does mastery of these words not indicate an ability to write fluently, as it would enable the children to write only text which contained these words, but the children are very likely to have been intensively coached in these words in their normal class teaching in order to reach one of the objectives in the Primary Framework. A standardised spelling test would have been a more useful indicator of the children’s general spelling achievement. It has been argued that performance in spelling gives a truer picture of children’s reading ability as they must be able to read words that they can spell . It is noticeable that the Sounds-Write data show that even the poorest children in YR (16.7% who scored below the baseline in spelling) had a mean spelling age (SA) 2.5 months ahead of chronological age (CA) at the end of Year 1.
I have earlier touched on the use of RR levelled books as a measure of reading success. This is an example of words found in a Band 5 text (RR levels 12-14): video, camcorder, anniversary, policewoman. The L15 text shown in the RR evaluation contains ‘competition’ and the L15 text example in the ECAR report contains the word ‘policewoman’. These are surprisingly complex words for a 6 yr old previous ‘non reader’ to decode independently, even after a daily 5 minute ‘phonics’ session. It may well be that the impressive gains in ‘Book Levels’ made by the RR children are due to the fact that the children were given levelled books to read as part of their RR sessions and, with several re-readings, have memorised many of the words as ‘wholes’ and also whole portions of the text.
How does this compare with children receiving high quality phonics instruction?
In the preamble to the evaluation a comparison is made with ‘early phonic training’ studies (Hatcher, 2006,’Reading Intervention’ and Johnston & Watson, 2005, Clackmannanshire). It is claimed that the children in these studies had ‘less challenging problems’ and even that some were deliberately excluded from studies. This seems an extraordinary claim, in particular in regard to the Clackmannanshire study, which was deliberately sited in an area of social disadvantage and placed the most disadvantaged children in the synthetic phonics group. It also did not exclude any children . The Clackmannanshire study shows only 2.7% and 4.3% of children significantly ‘behind’ on reading and spelling respectively at the end of Primary 2, and this without any intervention. (The actual numbers were 7 out of 251 and 11 out of 253.) Similarly, Dr Jonathan Solity  reports that in a study comparing extremely disadvantaged schools in Essex, where ERR was implemented, with schools using the NLS (Solity & Deavers et al 2000), the typically lowest 25% in the experimental groups had reading ages (RAs) only 6 -9 months behind their CAs, but 12 months ahead of comparison low achievers, and that the incidence of children perceived to have special educational needs (SEN), a usual consequence of children failing to learn to read, fell from approximately 20-25% to 2-5%. Sounds~Write data  shows only 11% of children significantly behind in reading (RA 6 months or more below CA) at the end of Year 1.
The mean of the final WRAPS test (a diagnostic assessment of word recognition and phonic skills) given to the 89 RR children is 75 months (6y 3m), still below their mean CA of 6y 5m (mean 5y 7m, Sept 05 plus 10 months to July 06). The entire cohort of children in the schools receiving RR only achieved a mean WRAPS score of 77.5 months (6y 5.5m) at the end of the 10 month period. This is very slightly ahead (0.5m) of CA (6y 5m). In contrast, the Clackmannanshire children, who started from a lower base of Word Reading Age 57m (CA 60m), had a mean word Reading Age of 76m (CA 65m) after only 4 months of synthetic phonics instruction. Sounds~Write children had a mean Word Reading age of 83months (6y 11m; mean CA 6y 3m) at end Y1 with only 11% at more than 6months below CA (no data for YR). Although this demonstrates that high quality phonics programmes can achieve excellent results, with fewer struggling readers, it also says something about the initial reading programmes used in the RR schools: that 24% of their children are achieving so poorly after 10 months of reading instruction in YR. That these programmes were not high quality phonics programmes is evident from the results of the ‘Teachers Beliefs about Literacy teaching and learning’ questionnaire (Appendix to the Evaluation). It is concluded from the responses to the questionnaire that “...class teachers in this London sample ...are within a middle band who could be said to adopt a balanced approach towards teaching and learning in literacy”. A ‘balanced approach’ to Literacy teaching is the approach promoted by the old NLS and is not compatible with high quality phonics teaching.
Cost of RR
It is well known that Reading Recovery is a very expensive intervention. It is estimated that it costs about £2,500 per child. ECAR produced a long report justifying the cost of the intervention with reference to the long term costs of reading failure. Long term costs are a legitimate concern. The RRF contends, however, that the cost of intervention is reduced as a direct result of the use of high quality phonics. Firstly, data from schools using high quality phonics show that far fewer children are underachieving at the end of the first year of instruction and so less intervention is needed. Secondly, the Rose Report recommends that intervention for those children who need it should involve at all stages “High quality phonic work, as defined by the review”, and so money should not be spent on interventions that do not conform to this criterion. Thirdly, the intensive, expensive training given to RR teachers is unnecessary; it is possible to train teachers and competent teaching assistants in the principles and practice of high quality phonics instruction in a matter of days. Once stripped of the mystique which surrounds whole language/mixed methods of reading instruction, the process of providing extra high quality phonics teaching for children who need extra help is straightforward and simple. It does not need expensive resources. The entire Dunbartonshire Literacy project, which has, to date, reduced ‘reading failure’ to 6% , costs £93 per child taken over the first three years of education.
Lessons from abroad
England is not the first English-speaking country to reach the conclusion that Whole Language/ mixed methods teaching of reading has been the driving force behind the fall in standards of literacy. In the US the No Child Left Behind Act 2001 mandated the use of reading instruction programmes based on ‘scientifically based reading research’ (SBRR). The implementation of this mandate is proving difficult, as whole language/ balanced literacy programme developers attempt to qualify their programmes for Reading First funding by claiming that they contain the necessary components to satisfy the legislation. We have already seen that Marie Clay, the author of RR, has stated that she would “rewrite the description of Reading Recovery’s components in such a way as to bring it into compliance with the Reading First legislation” while refusing to amend her programme to include a systematic, explicit phonics component. Louisa Moats, one of the foremost reading experts in the US, in her recently published paper, ‘Whole Language High Jinks’  demonstrates that, despite the claims of programme developers and sellers, many ‘balanced literacy’ programmes are being given funding on: (a) the strength of the description of their components and (b) the lack of sufficient expertise in school administrators to be able to identify programmes which conform to the criteria. We should learn from the US experience, not ignore it.
The DfES has abandoned its original plan to ‘kitemark’ high quality phonics programmes, relying instead on the expertise of practitioners to apply the criteria and choose an appropriate programme. However, we can see no reason why the DfES should be ignoring its own criteria and actively promoting interventions which contravene them. The DfES refuses to name or promote high quality phonics programmes for initial reading instruction; it should maintain consistency by similarly refraining from naming or promoting intervention programmes. If it feels that it must alert teachers to intervention programmes it should confine its guidance to programmes in which the key element is high quality phonics and ensure that no other part of the programme contains elements which have been superseded by the Rose Report Findings.
 For further information see:
 Salvato, N. (2006), “R” Stands for Reading Rat Race - Ed.News Oct 2006 http://www.ednews.org/articles/3227/1/quotRquot-Stands-for-Reading-Rat-Race/Page1.html
 Every Child A Reader: the results of the first year… http://www.ioe.ac.uk/schools/ecpe/readingrecovery/ECR.pdf
 Adams, M. J., The Three Cueing System - EducationNews,
 HOC Education and Skills Committee: Teaching Children to Read
 Rose J. (2006), Independent Review of the Early Teaching of Reading, online at http://www.standards.dfes.gov.uk/rosereview/report.pdf
 Taken from titles of books in Band 1, RR levels 1/2
 Burkard, T. (2006), A World First for Dunbartonshire, Centre for Policy Studies
 Burroughs-Lange, S. (2006), Evaluation of Reading Recovery in London Schools: Every Child A Reader2005-2006 http://www.ioe.ac.uk/schools/ecpe/readingrecovery/ECAR%20London%20Evaluation%20Report%20Final.pdf
 RRF website discussion forum: http://rrf.org.uk/messageforum/viewtopic.php?t=2470&postdays=0&postorder=asc&start=0
 Sounds-Write research report Autumn 2006
 Johnston, R.S. and Watson, J.E. (2005) The Effects of Synthetic Phonics Teaching on Reading and Spelling Attainment, a seven year longitudinal study, http://www.scotland.gov.uk/Resource/Doc/36496/0023582.pdf
 Solity, J. (2003), Teaching Phonics in Context: A Critique of the National Literacy Strategy
 Sounds-Write research report (2006) http://www.sounds-write.co.uk/sw_research_report_06.pdf
 Op Cit
 Moats, L. (2007), Whole Language High Jinks, http://www.edexcellence.net/doc/Moats2007.pdf