Requiem for Reading Recovery
John J. Russell EdD

Reading Recovery (RR) is a short-term, school-based literacy intervention for first-grade students who experience difficulty learning to read. Developed by New Zealand researcher Marie Clay, Reading Recovery uses specially trained teachers to work with students individually for 30-minute lessons each school day for 12 to 20 weeks. The Reading Recovery Council, a not-for-profit association of Reading Recovery professionals and advocates, claims that RR is effective and grounded in research. In 2002 the Reading Recovery Council made its position clear, posting, “In the national debate about scientifically based research and accountability, Reading Recovery is a surprising target because no program is more accountable and has a stronger scientific base than Reading Recovery.” These claims have been supported by a number of researchers. The Reading Recovery Council cites research (2002) in peer-reviewed journals that the Council argues documents Reading Recovery’s effectiveness (Center, et al., 1995; Iversen & Tunmer, 1993; Pinnell, 1997; Pinnell et al., 1994; Sylva & Hurry, 1996; Wasik & Slavin, 1993). Further support comes from The Institute of Education Sciences (IES), the statistics, research, and evaluation arm of the U.S. Department of Education. Included in the What Works Clearinghouse of the IES is the following statement: “Reading Recovery was found to have positive effects on general reading achievement and potentially positive effects on alphabetics, reading fluency, and comprehension for beginning readers” (Author’s note: The inclusion of “potentially” should not be ignored.). 

Given this level of support, it is not surprising that Reading Recovery quickly spread across schools in the United States. Since 1984, when RR was first introduced, at least 2.4 million students have participated in the program (Hanford & Peak, 2022). According to the Reading Recovery website, in 2002 Reading Recovery was in approximately 20% of public elementary schools; it is currently in 2,000 schools in 41 states. This level of adoption required that thousands of teachers receive the specialized training deemed necessary to deliver the program. In order to accommodate this demand for specially trained teachers, more than 20 universities across the United States established RR training programs; included among them are The Ohio State University, University of Connecticut, Georgia State University, and Lesley University. These partnerships with well-regarded universities further bolster the perception that RR is an effective research-based program. 

However, these positive reviews of RR have been countered by an ever-increasing number of reports that are critical of RR and cast doubts on the claims that RR is effective and research based. Why are so many reading researchers and educators deeply troubled by what they state are persistent misleading assertions about Reading Recovery? 

Why are so many reading researchers and educators deeply troubled by what they state are persistent misleading assertions about Reading Recovery?

The theoretical basis for Reading Recovery is not supported by reading research. 

The Reading Recovery Council alleges that RR is built on the five essential components of reading instruction—phonemic awareness, phonics, vocabulary, comprehension, and fluency (Reading Recovery Council, 2002). RR lessons include reading and re-reading books containing predictable text, cutting up and rearranging sentences, identifying letters, and introducing new books. Students are taught to guess words, rely on pictures to understand text, and use a similar word in place of the actual word written in the text. Marie Clay (1998) specifically states that beginning readers “need to use their knowledge of how the world works; the possible meanings of the text; the sentence structure; the importance of order of ideas, or words, or of letters; the size of words or letters; special features of sound, shape, and layout; and special knowledge from past literary experiences before [emphasis added] they resort to left-to-right sounding out of chunks or letter clusters or, in the last resort, single letters.” The use of these instructional strategies to teach struggling readers has been widely discredited (Vellutino, 1991; Moats, 2000; Moats, 2007; Tunmer et al., 2013; Spear-Swerling, 2018). For example, research has shown that predicting words from context is a highly ineffective learning strategy that is preferred by poor readers, not proficient ones (Chapman & Tunmer, 2002; Hanford & Peak, 2019). 

Significantly, RR does not consistently incorporate explicit, systematic instruction in phonemic awareness and the use of letter–sound relations in its lessons. According to Chapman and Tunmer (2018), “Such instruction is essential for most students who struggle with literacy learning during their early years of schooling and especially important for students who experience the most difficulty with learning to read.” 

Reading Recovery does not result in sustained reading achievement. 

Examining schools in New Zealand by comparing the performance of students that received Reading Recovery to a control group, Center et al. (1995) found that on an evaluation administered at 15 weeks, the students in the RR group were superior to control students on all tests measuring reading achievement but not on two out of three tests which measured metalinguistic skills, which are defined simply as the ability to think about and reflect upon language (Gillon, 2004) and, in more detail by the American Speech-Language-Hearing Association (ASHA), as phonological awareness, morphological awareness, syntactic awareness, semantic awareness, and pragmatic awareness. Further, an evaluation at 30 weeks revealed that there were no longer any differences between the RR and control children on seven out of eight measures. Surprisingly, 12 months after discontinuation, about 35% of RR students had benefited directly from the program, and about 35% had not been "recovered." Center et al. maintain that the remaining 30% would probably have improved without such an intensive intervention, since a similar percentage of control and comparison students had reached average reading levels by this stage. 

More troubling, Chapman, Tunmer, and Prochnow (2001) found that children selected for placement in RR and successfully discontinued from the program were on average six months behind their same-age peers at the end of the program and 12 months below their same-age peers on standardized measures of reading performance one year after they had left RR. The RR children who were deemed no longer in need of the program performed no better following their exit from the program than a group of poor readers who did not receive RR. Moreover, the RR children’s performance on a number of measures showed no acceleration effects during or after the RR program. Similarly, and even more disturbing, Chapman et al., (2009) and Nicholas and Parkhill (2013) found that over 40% of children who were successful in Reading Recovery lost their gains within two to four years and read at levels significantly below average. 

In a large study of Reading Recovery in the United States, May and his colleagues (2011) found evidence of large positive gains in first grade, but whether the initial gains lasted and translated into better performance on state reading tests remained a question. In a follow-up study (2022), May discovered that children who received RR had scores on state reading tests in third and fourth grade that were below struggling readers who had not received the program, confirming the previous finding that Center et al. had documented in 1995 in New Zealand. 

May discovered that children who received RR had scores on state reading tests in third and fourth grade that were below struggling readers who had not received the program.

The research supporting the effectiveness of Reading Recovery is flawed. 

The research that Reading Recovery advocates cite as evidence of its effectiveness has been the subject of continuing criticism. Scientists who have conducted careful reviews of the evidence base for Reading Recovery have frequently come to disturbing conclusions. In 1987, Shanahan reviewed Marie Clay’s The Early Detection of Reading Difficulty and concluded that Clay’s research was fundamentally flawed in that it had been “designed in such a way that it is impossible to know whether or not the program was successful.” Following Shanahan’s review, Hiebert conducted a re-evaluation (Hiebert, 1994) of Reading Recovery research studies carried out in the United States, highlighting shortcomings in many of the evaluative studies frequently cited by Reading Recovery advocates. Similarly to Shanahan (1987), he concludes, “the impact of this program clearly requires further investigation” (Hiebert, 1994). 

Among the many subsequent studies that followed Hiebert’s recommendation for further evaluation, two in particular are worth noting. Center et al. (1995) criticized Clay’s studies for what they deemed to be “significant design flaws including a) no matched group of poor readers or a proper control group, b) inappropriate use of multiple t-tests for analyzing gain scores, c) inclusion of only those RR students who were considered successful rather than all RR students, d) failure to account for spurious regression-towards-the-mean effects, e) using only performance measures devised by Clay rather than independent standardized tests, and f) intervention and comparison groups not equivalent at baseline.” Then in 2016, Chapman and Tunmer reported that “(a) many of the lowest achieving students were excluded from participation in Reading Recovery; (b) the control group received a range of different experiences; (c) the successful completion rate of students in the program was modest; and (d) no data supported the claim that Reading Recovery leads to sustained literacy learning gains.” 

Shanahan revisited his earlier review (1987) and summed up his finding and those of other researchers by simply stating, “The flaws in Clay’s data misleadingly made the program appear more successful than it had been” (Shanahan, 2022). One specific flaw is cited by Tunmer et al. (2015). They report that up to 30% of Reading Recovery students do not complete the program but are “referred on” (removed from the program) instead for further assessment. Typically, these are children who are the lowest achieving students and are unlikely to respond to the Reading Recovery program. In most evaluations of Reading Recovery, these students are not included in the reported results, which, as Shanahan had previously noted, makes the program seem more successful than it has actually been. In an open letter entitled Experts Say Reading Recovery Is Not Effective, Leaves Too Many Children Behind, 30 international reading researchers expressed parallel concerns about Reading Recovery stating, “While research distributed by the developers of Reading Recovery indicates a positive effect of the program, analyses by independent researchers have found serious problems with these conclusions. Studies conducted by researchers associated with Reading Recovery typically exclude 25-40% of the poorest performing students from the data analysis” (Wrightslaw, 2022). 

RIP Reading Recovery? 

There are serious deficiencies in the Reading Recovery program; most notable among them: 

■ The program is grounded in the widely discredited whole language philosophy, not the Science of Reading (Hanford & Peak, 2022). For example, Reading Recovery teaches phonics and phonemic awareness, but the instruction is not sufficiently explicit. 

■ The assessments used in the program were developed by Reading Recovery in-house researchers rather than the norm-referenced tests that are commonly used in reading intervention research (Wrightslaw, 2022), casting serious doubts on the claims by RR proponents about the effectiveness of the program (Shanahan, 2022). 

■ The poorest readers instructed with Reading Recovery showed very little improvement (Elbaum et al., 2000), and students who do complete the Reading Recovery sequence in first grade lose much of their gains (Hiebert, 1994; May et al., 2022). 

For many of the same reasons stated in this article, the more than 30 researchers who were signatures to the open letter published in Wrightslaw conclude with what may in fact be the death knell for this controversial program: “Reading Recovery leaves too many students behind.”