Evidence Check 1: Early Literacy Interventions - Science and Technology Committee Contents

2  Every Child a Reader: Reading Recovery

Q 1  The policy

9. The former Prime Minister Tony Blair famously told the Labour Party conference in 1996 that "education, education, education" were his three main priorities for government. The National Strategies,[4] developed in 1998 and expanded in 2005-06, advocate three Waves of Provision for addressing the range of educational needs in schools:

    Wave 1: Quality First Teaching. The majority of children achieve well through high quality classroom teaching. When children are being taught to read, Quality First Teaching provides high quality, systematic phonic work as part of a broad and rich curriculum that engages children in a range of activities and experiences to develop their speaking and listening skills and phonological awareness.

    Wave 2: Small group and one to one interventions. Some children require additional support to achieve well. This can often be provided through small group, time limited intervention programmes delivered by a member of the school's classroom based support team that will advance children's progress and help them achieve in line with their peers.

    Wave 3: Intensive support. This is for those children who require the personalised approach of a programme that is tailored to their specific, often severe, difficulties. It is usually taught as a one to one programme by a teacher or a member of the support staff who has undertaken some additional training for teaching children with reading difficulties.[5]

10. When in 2007 the Government formed the Department for Children, Schools and Families (DCSF), one of its first acts was to draw up The Children's Plan, which set out a bold vision for ensuring that "no child or young person is left to fall behind".[6] One of the flagship elements of this promise was the Every Child programmes—Every Child a Reader, Every Child a Writer and Every Child Counts—which provide targeted support, including one-to-one tuition, for those children who in their early school years have difficulties with reading, writing and numeracy.


11. Every Child a Reader (ECaR) is a programme of work aimed to help poor readers catch up with their peers. It consists of a programme of Reading Recovery (a type of one-to-one teaching intervention) for the lowest achieving 5% and less intensive intervention for the next lowest 15% delivered by teaching assistants and volunteers supported by the school's Reading Recovery teacher.[7] The programme was launched as a three-year pilot in 2005, although it commenced national rollout only a year later.[8] The national rollout is being managed by the National Strategies, which are professional development programmes for early years, primary and secondary school teachers, practitioners and managers.[9]

12. Reading Recovery, the bedrock of the ECaR Wave 3 programme, was developed by the late Marie Clay in the 1970s in New Zealand, where approximately two thirds of schools use it. It has also been widely adopted in Australia and the USA, but less so in the UK. Reading Recovery comprises 12-20 weeks of intensive, one-to-one, daily tuition by specially trained teachers. It is designed for year 1 and 2 children.[10]

Our expectations of the evidence base

13. The effectiveness of the Government's policy for ensuring that no children are "left behind" rests on a number of factors. We looked at two: the Government's decision to prioritise literacy in education, particularly the decision to focus on tackling literacy interventions early in school; and the Government's decision to use Reading Recovery.


14. An obvious initial question is whether early interventions are worthwhile. There are two parts to this. First, whether early interventions are more effective than later interventions in terms of how well an individual will read into adulthood. Second, whether the cost of intervention is, from a societal point of view, worthwhile. Neither are easy to calculate.

15. The first would, for the best evidence, involve longitudinal studies of children who variously receive or do not receive reading interventions at differing times during their education. However, this would be exceedingly difficult to do and certainly unethical since it would involve leaving some children with known reading difficulties to fend for themselves rather than receive the same level of help as their peers. What researchers are left with is the far-from-ideal situation of having to compare those interventions that are made early in a child's education with those that are made later. (Some interventions, like Reading Recovery, are aimed specifically at very early schooling, years 1 and 2, some are aimed at later years and some are flexible across a wide age range.) This situation is unsatisfactory because the confounding factors that can be controlled in a snap-shot trial are more problematic over a long period of time. For example, some children who receive later literacy interventions because "they did not get it the first time"[11] may not have received an intervention before, others may have received earlier interventions that did not work; there will be differences in the amount of help children get from their parents and their peers; the underlying reasons for the difficulty may be hard to control and may have changed over time; and underlying difficulties can be cognitive or pedagogical or both.

16. Whether or not early interventions are better than later interventions is only part of the story. The bigger question is whether ensuring that every single child can read is really necessary. As a society, we choose to prioritise reading, but teaching children to read—all children—is expensive and teaching the very worst readers is very expensive indeed. Reading Recovery, for example, costs approximately £2,600 per child.[12] Is it worth it? To answer this question, we need to know the cost of doing nothing: it will either be more expensive to society, about the same or less. We did not put preconditions or stringent expectations on the nature of the research we were looking for in this area. Calculating societal costs is complicated and inevitably involves a large dose of estimation and generalisation. We discuss the quality of this evidence in paragraphs 26-29.


17. Educationalists put literacy interventions to the test to see whether they work or not, and if they do work, how effective they are. Here are the basics:

  • In order to test whether a reading intervention has worked, one needs to know the starting reading standard of a child (pre-test assessment) and the finishing standard of a child (post-test).
  • The gold standard research model is a randomised controlled trial (RCT). It is randomised so that decisions about who receives a reading intervention are taken out of the hands of the person who will be doing the teaching to avoid bias. It is controlled in the sense that those children who receive an intervention are compared against a matched group of children who do not.
  • The silver standard is a quasi-experiment in which randomisation does not take place, but groups of children in the experimental group are matched against control groups of children who do not receive an intervention.

18. Of course, this is a simplification. One does not necessarily need to compare children who receive an intervention with those who do not. In the case of reading interventions it is preferable to compare different reading interventions directly against each other. There are three reasons for this:

a)  it is well established that one-to-one support for a child yields better results than normal classroom teaching;[13]

b)  what is more pertinent is how the variety of reading interventions compare in terms of effectiveness; and

c)  there are ethical problems with assigning children who have a known reading difficulty into an experimental category where they receive no extra help at all.

19. All of this is overlooking a very obvious question: how does one measure improvements in literacy? Dr Chris Singleton extols the virtues of standard (or standardised) scores on the grounds that "they are age-independent and test-independent and enable a proper comparison between different groups and different studies".[14] Additionally, standardised scores are designed to make it possible to carry out what statisticians call 'parametric' statistics, which is a way of saying that the interactions between variables can be quantified and statistical significance calculated. In other words, when standardised tests are used it is possible to bring together numerous studies to improve the power of the data and to control for a variety of complicating variables.

20. The alternative to standardised scores are 'reading ages', which are metrics for how well normally progressing readers will develop. A normal reader aged 6 years 2 months will have a reading age of 6 years 2 months. This measure is less statistically robust than standardised scores.

21. The size of the difference between literacy skills at the start of an intervention and at the end of an intervention, or years later, can be calculated in two ways. The best is to calculate the 'effect size' between two standardised scores (where the mean is usually 100 and the standard deviation 15). By measuring the difference between two scores and dividing by the standard deviation, researchers can calculate statistically robust measures of success (or otherwise).

22. An alternative, less rigorous approach is to calculate the 'ratio gain'. In this measure, the reading (or spelling) age of a child during a set time frame is expressed as a ratio of that time frame. For example, a normal child aged 5 years 2 months will, three months later, have a reading age of 5 years 5 months. This is a ratio gain of 1.0. A child who progresses only three months in a year will have a smaller ratio gain, 0.25. What we are looking for is children who have fallen behind their peers but who receive extra help and improve quickly. Professor Greg Brooks, Research Director of the Sheffield arm of the National Research and Development Centre at the University of Sheffield, has suggested that ratio gains of 1.4 or higher are of "educational significance".[15] Ratio gains are less satisfactory than effect sizes. They do not offer the same statistical rigour and they are particularly problematic for research on the lowest attaining readers because, for example, rapid improvement in reading age from a very low base will yield high ratio scores even if the children are still very poor readers.

The Evidence Check

The value of intervention

23. The Government's position is that the earlier an intervention can be made for a struggling reader, the better. We asked our panel of expert witnesses[16] about the evidence base for early interventions and they were in unanimous agreement:

    Chairman: So, in your view, the evidence is there to say early intervention works?

    Professor Slavin: I think without any doubt.

    Chairman: Jean, do you support that?

    Jean Gross: Yes, I do. I support that completely.

    Chairman: Professor Brooks?

    Professor Brooks: Yes, it is better if they come in early, when children are first identified as struggling, but there are also programmes for those who are picked up rather later.[17]

24. However, it is difficult to test directly and we have therefore had to rely on comparative evidence. We turned to Professor Greg Brooks' extensive review, What works for pupils with literacy difficulties?, which was commissioned by the Government.[18] The first edition came out in 1998, with subsequent editions in 2002 and 2007. The tables of effect sizes and ratio gains for all of the different interventions that he reviewed provide interesting reading. They suggest that on average the earliest interventions (years 1 and 2) are beneficial and provide approximately the same benefits, give or take a few effect size or ratio gain fractions, as later interventions (years 3 and beyond).[19] Both the very earliest interventions and later interventions are, on average, of "useful impact", to use Professor Brooks' language.[20]

25. It is clear that literacy interventions, both at the very earliest stages of formal education and later on, can be effective. This is all the evidence we need, since it logically follows that the earlier struggling readers are identified, the longer the time educationalists have to tackle the their problems and teach them to read. (It is possible that there are also additional benefits of bringing poor readers back to parity with their peers as early as possible—for example, boosting the self-esteem of the individuals and enabling them to take part more fully in the wider curriculum—although we have not been directed to evidence that supports these suppositions.) The Government's policy that literacy interventions should take place early on in formal education is in line with the evidence.

26. Another angle on the same topic, is to ask whether these literacy interventions are worthwhile from a societal point of view. The evidence we were referred to in order to answer this problem was research carried out by the KPMG Foundation[21] and the Every Child a Chance Trust,[22] although it is fair to say that this cannot be described as wholly independent given its interest in the work. These studies provided estimates of the cost of literacy difficulties to society. Some of these costs are simple to calculate because they are a direct consequence of literacy difficulties; for example, adult literacy classes. But the majority of costs require a subtle way of making estimations because the relationship between literacy difficulties and the cost-factors are complicated; for example, truancy, unemployment, teenage pregnancy, depression, obesity and crime.

27. The KPMG Foundation and Every Child a Chance Trust chose to make these complicated estimates by using frequency differentials. For example:

    10% of women with average literacy levels experience depression and 36% of women with very low literacy skills experience depression. The differential frequency is therefore 26%.

    Applying this figure of 26% to the 12,300 females in the year group population of 38,700 gives 3,198 more women with very poor literacy skills who can be expected to experience depression than would be expected if they were average readers.

    79% of these (2,526 women) can be assumed to escape depression because early literacy intervention has successfully lifted them out of very low literacy levels.

    We assumed that the differential rate of depression applies for adult life, and cannot be limited to a particular age range, so those with poorer literacy levels will be more likely to experience depression whatever their age.

    The costs of depression were identified as £194 (inflated to 2008 prices) per year per depressed person.

    This cost was then applied throughout the adult lives (ages 18-37, and over a lifetime) of the identified 2,526 subjects to obtain total cost savings for depression for women.[23]

28. These kinds of estimates are fraught with assumptions and generalisations that will undoubtedly lead to miscalculation. This particular methodology is likely to offer an overestimation of the cost, because its simple calculation of a differential frequency on the basis of two variables—normal literacy skills versus low literacy skills—means that a large number of co-occurring variables have not been considered. To be fair, it would be very difficult to control for all these variables properly and both reports are proportionately littered with cautions about the data and methodologies. For example, one of the most difficult costs to calculate was crime. The Every Child a Chance Trust notes that there is a correlation between crime and literacy difficulties, but warns:

    [The d]ifferentials [are] based on empirical data on [the percentage] of children with literacy difficulties who also have behaviour problems (and empirical data about costs of behaviour problems to criminal justice system) but [there are] no controls for other factors (such as general cognitive ability, social class) that might explain the link.[24]

29. With all the appropriate caveats and warnings, the two reports make estimates of the long term costs of literacy difficulties to society. The most recent Every Child a Chance Trust report, estimates that literacy difficulties cost England in the region of £2.5 billion every year (see Table 1).

30. The authors have suggested that these costs are "conservative"[25] because they have not included items such as soft costs (like illness and loss of income) or "social services costs, social housing costs, the costs of generally poorer health, the costs of substance abuse over the age of 18, the costs of women's involvement in the criminal justice system and lost tax on pension income".[26]

31. Although in this section we have discussed the value of interventions in general, it is worth noting that the reports were commissioned explicitly to ascertain whether the Every Child a Reader programme represented value for money. The conclusion of the reports was that it did. The second edition calculates that for every pound spent on the Every Child a Reader programme, there is an overall return on investment of between £11 and £17.[27]

32. Estimating the cost of literacy difficulties is clearly not easy to do, but we believe that that should not stop researchers from making the best estimates they can. We were impressed by the KPMG Foundation and Every Child a Reader Trust's efforts. While the figures quoted are unlikely to be correct, they clearly show that there is a substantial cost associated with literacy difficulties. Spending money on literacy interventions is a cost effective thing to do. The Government's position that early literacy interventions are an investment that saves money in the long run is evidence-based.

Table 1. The long term costs of literacy difficulties.[28]


33. We are satisfied that the Government is right to support literacy interventions, but the question remains as to which literacy interventions are most appropriate. There are many choices. Professor Brooks reviewed the evidence in 1998, 2002 and 2007 for the Government. In the 2007 edition, he reviewed large and small group teaching, one-to-one teaching, paired reading, and teaching using information and communication technologies (ICT). He also analysed the impact of teaching phonological skills, comprehension and self-esteem. In all, he reviewed evidence for 48 different kinds of reading intervention.[29]

34. Given the large range of options, we asked the Government what other kinds of interventions were considered before the decision was taken to make Reading Recovery the bedrock of the Every Child a Reader programme. Jennifer Chew, a retired English teacher, summarised the Government's response to the Committee's question about the cost effectiveness of different literacy interventions:

    Other literacy interventions are available which are cheaper than Reading Recovery, which are more consistent with Wave 1 teaching, and which arguably produce better results […] If the government did consider alternatives before providing funding for Reading Recovery, it should be able to say which the alternatives were, how they were investigated, and what evidence led to the conclusion that they were less cost-effective than Reading Recovery.[30]

35. In oral evidence, we asked Ms Johnson and Ms Willis what alternative interventions were considered. Ms Johnson told us she did not know the answer,[31] and Ms Willis pointed us to Professor Brooks 2002 review.[32] We asked for a supplementary memorandum from the Government on this point. The Government responded:

    Interventions other than Reading Recovery we considered

    The choice of Reading Recovery as the core intervention of the ECAR programme was made during the pilot phase led by the Every Child A Chance Trust. The Department saw no reason to change this when taking on the programme for national roll-out.[33]

36. In other words, the Government did not formally consider any other kind of intervention. We pressed Ms Willis on this point who reassured us that: "it is my strong interest and my purpose in the Department to ensure that policy is based on sound evidence".[34] She continued:

    We have commissioned an independent evaluation just recently from IFS [Institute for Fiscal Studies], the University of Nottingham, and NatCen to look very carefully at how the programme is implemented, to undertake a cost/benefit analysis and to look at the value for money of the way in which ECAR is being implemented. […] What we will need to do […] is compare that with other emerging evidence from alternative interventions.[35]

37. Ms Willis is right to acknowledge the need to compare Reading Recovery with alternative interventions. We conclude that, whilst there was evidence to support early intervention, the Government should not have reached the point of a national roll-out of Reading Recovery without making cost-benefit comparisons with other interventions.


38. We have already discussed what evidence we were looking for in dealing with this problem: randomised controlled trials that use standardised test scores. We were alarmed to discover that both are lacking in the UK literacy research base. On data quality, Sir Jim Rose noted:

    Most of the US studies included in Dr Chris Singleton's review[36] have reported standard scores; unfortunately few of the UK studies have done so, sometimes because the tests used do not provide tables of norms in standard score form.[37]

39. Professor Brooks, in the 2007 version of What Works, summarised some of the problems with the data used in UK literacy intervention trials:

    Three particular problems arose from the tests used in the 121 studies. Firstly, some of the tests were old even when used in the relevant studies.

    Secondly, most of the tests provided only reading/spelling age data and not standardised scores. Though apparently easier to interpret, reading and spelling ages are statistically unsatisfactory […] Reading and spelling age data do allow the calculation of the ratio gain—but this is in itself not a very useful statistic, especially for low-attaining groups. […]

    Thirdly, for many of the tests used it was impossible to calculate effect sizes, which are statistically much more satisfactory than ratio gains. If a standardised test is used, an effect size can be calculated even in the absence of an explicit comparison group; but if a non-standardised test is used then an effect size can be calculated only if comparison group data, including the standard deviation, are reported.[38]

40. We are concerned by the low quality of data collection in UK trials on literacy interventions. Government-funded trials should seek the best data so as to make the results as powerful as possible. Running trials that do not collect the best data is a failure both in terms of the methodological approach, but also value for money.

41. Professor Brooks also commented on the different types of trial model used to assess literacy interventions. He noted that there are very few randomised trials "and some of those were so small as to be hardly worth carrying out".[39] There were a few quasi-experiments with matched groups, but the bulk were either of unmatched groups or just one-group studies.

Research design of 121 studies analysed by Professor
Brooks in the 2007 version of his review.[40]
Research design
randomised controlled trial
matched groups quasi-experiment
unmatched groups pre-test/post-test study
one-group pre-test/post-test study

42. Given the Government's support for Reading Recovery we were alarmed to discover that none of the nine randomised controlled trials were on Reading Recovery.[41] We were reassured by Ms Willis that there have been RCTs of Reading Recovery in the United States:

    Reading Recovery has a strong evidence base. It has been reviewed for example by the What Works Clearinghouse in the States and includes a number of randomised controlled trials.[42]

43. However, when we sought these trials, we discovered that they took place in 1988,[43] 1994,[44] 1997,[45] and 2005.[46] More than 15 years separate the first trial from the most recent. This is important because Reading Recovery has evolved considerably over time: between the '90s and mid-'00s, Professor Brooks observed, "Reading Recovery changed considerably, to reflect international research, and now includes a large amount of phonological awareness and phonics".[47] In other words, three of the four RCT trials were on a kind of Reading Recovery that is not used in the UK.

44. Ms Willis stated that the Government has accepted the United States What Works Clearinghouse report as supporting the effectiveness of the Reading Recovery programme, but was unaware of evidence from Mary Reynolds et al in the International Journal of Disability, Development and Education, which contradicts this work and concludes other methods would probably work better.[48]

45. The Government should be careful when selecting evidence in support of educational programmes that have changed over time. Reading Recovery today differs from its 1980s and 1990s ancestors. Evidence used to support a national rollout of Reading Recovery should be up-to-date and relevant to the UK. The Government's decision to roll out Reading Recovery nationally is not based on the best quality, sound evidence.


46. To recap, the Government is rolling out Reading Recovery nationally without having subjected it to a randomised controlled trial in the UK school system. The Government had an opportunity to commission a randomised controlled trial of Reading Recovery when it started the three-year pilot for Every Child a Reader in 2005. It would have been particularly sensible in the light of calls from leading experts that RCTs were needed. In 2001, Carole Torgerson and David Torgerson from the Institute for Effective Education, University of York, made a general call for more RCTs in educational research:

    Educational researchers have largely abandoned the methodology they helped to pioneer. This gold-standard methodology should be more widely used as it is an appropriate and robust research technique. Without subjecting curriculum innovations to a RCT then potentially harmful educational initiatives could be visited upon the nation's children.[49]

And Professor Brooks made a specific call for an RCT on Reading Recovery when he was a member of the advisory group to the Every Child a Reader study.[50]

47. We asked Ms Willis, DCSF's Chief Scientific Adviser, why an RCT had not been undertaken. She told us that RCTs are "the gold standard in terms of research" but that "I do not believe it is always essential to have a randomised controlled trial".[51] She noted a number of problems with running an RCT:

  • there is a "logistical issue around picking schools and then randomly allocating them and then having […] teacher leaders in local authorities […] going out to train up the teachers in individual schools";[52]
  • it is difficult to get schools to agree to take part in research generally and randomised controlled trials in particular;[53] and
  • DCSF has a "limited research budget" and "matched comparison groups can be undertaken at less cost and deliver very similar quality results".[54]

48. She added, however, that:

    I should add the Department is not against RCTs. We are running a number. We have one underway looking at foster care and we have one underway looking at how to reduce teenage pregnancy, a very interesting randomised controlled trial there, although that has taken over a year just to try and get the methodology right and actually convince people to take part. Of course we have the Every Child Counts randomised controlled trial.[55]

49. We cannot understand why these issues warranted the extra time, effort and money, and a crucially important issue such as literacy did not. We put it to Ms Willis that the Government should have criteria for determining whether or not a research project requires an RCT,[56] for example, in the Magenta Book, which provides guidance on social science research and policy evaluation.[57] We were pleased that Ms Willis found our suggestion "interesting".[58] We recommend that the Government should draw up a set of criteria on which it decides whether a research project should be a randomised controlled trial.

50. But still the question remains as to whether a randomised trial was necessary. Jean Gross, Director of the Every Child a Chance Trust, thinks it was: "Looking back, I wish at the time we had been able to find money to do a randomised control trial".[59] And as we already mentioned, Professor Brooks thought so too.[60] We agree. The arguments put forward by Ms Willis for not doing an RCT do not stand up to scrutiny.

51. The argument that it is difficult to persuade schools to engage in research is made on the basis that "Some local authorities or schools perceive it as unfair that some of their pupils will be getting some sort of intervention that others are not".[61] This is focussing on a problem that should not exist. We have already said that because it is so clear that one-to-one interventions are better than doing nothing it is not worth conducting trials that compare one-to-one interventions against no intervention. Rather, it would be more beneficial to conduct randomised trials that compared two or more kinds of intervention, in order to determine which was better. This would avoid the ethical problem of some children not receiving help when they need it and should make it easier to persuade schools to take part in research (provided the interventions offered reasonable prospect of improvement).

52. The logistical difficulties of conducting randomised trials for literacy interventions are no different than for any other kind of educational randomised trial; they are not prohibitive and accordingly neither are the costs (approximately £250,000).[62] As we have already discussed, early literacy interventions appear to be highly cost effective. In our view, proven cost effective measures that have the potential to be even more cost effective should be subjected to the highest quality research, not marginally cheaper substitutes. The Government spent a lot of money setting up the pilot programme for Every Child a Reader, which failed to gather the highest quality data or use a gold-standard trial design.

53. Finally, Professor Brooks raised another problem with running randomised controlled trials: time.[63] Although running an RCT does not take longer than any other kind of well designed trial, setting one up can take a long time. Ms Willis told us that it has taken over a year to set up an RCT on reducing teenage pregnancies.[64] But this is not a good enough excuse: the pilots for Every Child a Reader were due to run for three years from 2005. That would have been more than enough time to set up and run a large randomised controlled trial on a more than just Reading Recovery.

54. We conclude that a randomised controlled trial of Reading Recovery was both feasible and necessary.

55. But we are where we are. The Government is rolling out Every Child a Reader nationally. However, we do not consider this a reason to abandon further research. Reading Recovery is an expensive component of the ECaR programme and if there is an equally effective but cheaper alternative, it should be sought out. Research in this area should be ongoing, not stop when the Government decides to roll out a particular programme. We recommend that the Government identify some promising alternatives to Reading Recovery and commission a large randomised controlled trial to identify the most effective and cost-effective early literacy intervention.


56. We have one final concern. The teaching of systematic phonics (see box below) became a requirement by National Curriculum Order in 2007.[65] But, according to Dr Singleton, Senior Research Fellow at the University of Hull, Reading Recovery is "a pedagogical sibling to the 'whole-language' theory of reading, which maintains that reading skills arise naturally out of frequent encounters with interesting and absorbing reading materials".[66] This view of learning has been "increasingly contested" said Dr Singleton in his 2009 review of the evidence,[67] and over time the accumulating evidence on the benefits of systematic phonics work has influenced the way in which Reading Recovery is delivered.[68] Despite this trend, Dr Singleton concludes that:

Sir Jim Rose, in his 2006 report on the teaching of early reading, quotes Linnea Ehri:

    Phonics is a method of instruction that teaches students correspondences between graphemes in written language and phonemes in spoken language and how to use these correspondences to read and spell words. Phonics instruction is systematic when all the major grapheme-phoneme correspondences are taught and they are covered in a clearly defined sequence.[69]

—  Synthetic phonics: readers are taught to break down words into their constituent parts and work out how to pronounce words for themselves (e.g., 'shrink' would be broken down into 'sh', 'r', 'i', 'n' and 'k' and blended together).

—  Analytical phonics: readers are taught consonant blends as units (e.g., 'shr' is taught as a whole unit) and analyse sound-symbol relationships but do not blend the words together.

—  Embedded phonics: readers are taught phonics as part of a whole-word approach to reading, not as separate lessons.

He went on to recommend, and the Government accepted, that "the case for systematic phonic work is overwhelming and much strengthened by a synthetic approach".[70]

    [N]either Reading Recovery as part of [Every Child a Reader] nor Reading Recovery in the UK more generally provides systematic phonics instruction. [… D]espite these reported changes to the reading recovery programme, a fundamental conflict still remains between its approach and the revised National Literacy strategy, in which systematic teaching of phonics is now a central feature.[71]

57. We received evidence that supported this position. Elizabeth Nonweiler, Teach to Read, highlighted Dr Sue Bodman's method of delivering a Reading Recovery lesson:

    Bodman (2007)[[72]] describes a Reading Recovery lesson, which, she claims, 'links the teaching actions to the ideas of synthetic phonics': After reading a book, a child observes his teacher reading the word 'can' 'whilst demonstrating a left to right hand sweep'. Then he builds 'can' with magnetic letters and reads it himself. It is clear that the child was asked to read a text before acquiring the phonic knowledge and skills involved, and to read a word after being told the pronunciation. With synthetic phonics children read texts after learning the phonic knowledge and skills involved and they are not told the pronunciation of a new word before being asked to read it.[73]

58. We put it to the Minister, Ms Johnson, that the Government's use of Reading Recovery to help the poorest readers does not square with its support of the use of systematic phonics, particularly for children diagnosed with dyslexia. She told us:

    Perhaps I ought to just say that the Reading Recovery that we are looking at, in terms of the evidence of Reading Recovery over the last 20 or 30 years, has changed and obviously phonics is now much more embedded within Reading Recovery than it was in the earlier examples of Reading Recovery.[74]

59. Phonics is "embedded" in the modern Reading Recovery, but systematic, synthetic phonics, as we have discussed above, is not. Teaching children to read is one of the most important things the State does. The Government has accepted Sir Jim Rose's recommendation that systematic phonics should be at the heart of the Government's strategy for teaching children to read. This is in conflict with the continuing practice of word memorisation and other teaching practices from the 'whole language theory of reading' used particularly in Wave 3 Reading Recovery. The Government should vigorously review these practices with the objective of ensuring that Reading Recovery complies with its policy.

4   www.nationalstrategies.org.uk Back

5   Sir Jim Rose, Identifying and Teaching Children and Young People with Dyslexia and Literacy Difficulties, 2009, p 60 Back

6   DCSF, The Children's Plan: Building brighter futures, 2007, p 4 Back

7   Ev 3 [Every Child a Chance Trust], para 3.10 Back

8   Policy Exchange, Rising Marks, Falling Standards, 2009, p 37 Back

9   www.literacytrust.org.uk/database/readingrecovery.html. According to www.nationalstrategies.org.uk: "Since 1998, the National Strategies have taken the form of a professional development programme providing training and targeted support to teachers (and increasingly practitioners). The training and support programmes have been supplemented by materials and resources provided free of charge and widely accepted as being of high quality. During 2005 and 2006, the National Strategies took on an expanded brief, which increased the age range covered and the way in which system-wide challenge and support is provided through the Strategies. The expanded functions include: Local Authority whole school improvement; School Improvement Partners; Behaviour & Attendance; Early Years from 0 to 5 years; 14-19." Back

10   Chris Singleton, Intervention for Dyslexia, 2009, p 95 Back

11   Q 20 [Professor Greg Brooks] Back

12   Ev 6 [Every Child a Chance Trust], para 6.6 Back

13   Greg Brooks, What works for pupils with literacy difficulties?, DCSF, 2007; Sir Jim Rose, Identifying and Teaching Children and Young People with Dyslexia and Literacy Difficulties, 2009; Chris Singleton, Intervention for Dyslexia, 2009 Back

14   Chris Singleton, Intervention for Dyslexia, 2009, p 26 Back

15   Greg Brooks, What works for pupils with literacy difficulties? DCSF, 2007, p 270 Back

16   Professor Bob Slavin, Director of the Institute for Effective Education, University of York, and Director of the Center for Research and Reform in Education, Johns Hopkins University; Jean Gross, Director of the Every Child a Chance Trust; Professor Greg Brooks, Research Director of the Sheffield arm of the National Research and Development Centre, University of Sheffield. Back

17   Qq 3-5 Back

18   Greg Brooks, What works for pupils with literacy difficulties? DCSF, 2007 Back

19   We calculated this by averaging the effect size and ratio gain in the experimental groups on accuracy for those interventions that focussed on years 1 and 2 (the early cohort) and years 3 and beyond (the later cohort). We discounted those studies that assessed both early and later cohorts. The average effect size for the early interventions was 0.5 (excluding the highest effect size, 3.5, a clear outlier); for the later interventions it was 0.5. The average ratio gain for the early interventions was 2.5; for the later interventions it was 2.8 (excluding the highest ratio gain of 16.1, a clear outlier). This is not a rigorous test-for example, we did not weight according the quality of the trials-and is intended to be indicative only. Back

20   Greg Brooks, What works for pupils with literacy difficulties? DCSF, 2007, p 270 Back

21   KPMG Foundation, The long term costs of literacy difficulties, 2006 Back

22   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009 Back

23   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009, p 21 Back

24   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009, p 46 Back

25   As above Back

26   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009, p 21 Back

27   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009, p 5 Back

28   Every Child a Chance Trust, The long term costs of literacy difficulties, 2nd edition, 2009, p 6 Back

29   Greg Brooks, What Works for Children with Literacy Difficulties? The Effectiveness of Intervention Scheme, Department for Education and Skills, 2002, p 15 Back

30   Ev 67 [Jennifer Chew], para 7 Back

31   Q 184 Back

32   Q 185 Back

33   Ev 54 Back

34   Q 119 Back

35   Qq 119-120 Back

36   Chris Singleton, Intervention for Dyslexia, 2009 Back

37   Sir Jim Rose, Identifying and Teaching Children and Young People with Dyslexia and Literacy Difficulties, 2009, p 176 Back

38   Professor Greg Brooks, What works for pupils with literacy difficulties? Department for Children, Schools and Families, 2007, p 110 Back

39   Q 21 Back

40   Greg Brooks, What Works for Children with Literacy Difficulties? The Effectiveness of Intervention Scheme, Department for Education and Skills, 2002, p 124 Back

41   Greg Brooks, What Works for Children with Literacy Difficulties? The Effectiveness of Intervention Scheme, Department for Education and Skills, 2002, pp 124-125 Back

42   Q 115; see also Q 171 Back

43   G. S. Pinnell, D. E. DeFord & C. A. Lyons, "Reading Recovery: Early intervention for at-risk first graders", Educational Research Service Monograph, (1988) Back

44   G. S. Pinnell, C. A. Lyons, D. E. DeFord, A. S. Bryk & M. Seltzer, "Comparing instructional models for the literacy education of high-risk first graders", Reading Research Quarterly, Vol. 29 (1994), pp 8-39 Back

45   N. Baenen, A. Bernhole, C. Dulaney & K. Banks, "Reading Recovery: Long-term progress after three cohorts", Journal of Education for Students Placed at Risk, Vol. 2 (1997), p 161 Back

46   R. M. Schwartz, "Literacy learning of at-risk first-grade students in the Reading Recovery early intervention", Journal of Educational Psychology, Vol. 97 (2005), pp 257-267 Back

47   Greg Brooks, What works for pupils with literacy difficulties? 2007, p 74 Back

48   Reynolds, M., & Wheldall, K. (2007). "Reading Recovery 20 years down the track: Looking forward, looking back. International Journal of Disability, Development and Education", Vol. 54, pp 199-223 Back

49   C. J. Torgerson & D. J. Torgerson, "The need for randomised controlled trials in educational research", British Journal of Educational Studies, Vol. 49 (2001), pp 316-328 Back

50   Q 56 Back

51   Q 134 Back

52   Q 177 Back

53   Qq 163, 177 Back

54   Q 134 Back

55   As above Back

56   Q 168 Back

57   Ev 56; Government Social Research Unit, The Magenta Book: guidance notes for policy evaluation and analysis, 2007 Back

58   Q 168 Back

59   Q 24 Back

60   Q 56 Back

61   Q 163 Back

62   Q 24 Back

63   Q 54 Back

64   Q 134 Back

65   Ev 55 Back

66   Chris Singleton, Intervention for Dyslexia, 2009, p 96 Back

67   Chris Singleton, Intervention for Dyslexia, 2009, p 97 Back

68   As above Back

69   Sir Jim Rose, Independent review of the teaching of early reading, Department for Education and Skills, 2006, p 17 Back

70   The Rose Report, p 20 Back

71   Chris Singleton, Intervention for Dyslexia, 2009, p 99 Back

72   S. Bodman, "Skilful teaching of phonics in Reading Recovery", The Running Record, Issue 12 (2007), pp 3-5 Back

73   Ev 74 [Elizabeth Nonweiler], para 5 Back

74   Q 114 Back

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2009
Prepared 19 December 2009