May 14, 2022, I published what I thought would be my last word on Reading Recovery (RR) http://shanahanonliteracy.com/blog/me-and-reading-recovery.
Fat chance.
RR, if you don’t know, is a remedial reading program for first graders. It started in the 1970s in New Zealand and was widely adopted throughout the United States. Over the years, it has been both widely lauded and decried by reading authorities. From its inception, it was the focus of lots of research – much of which seemed to support its effectiveness, though I had expressed concerns about this (Shanahan, 1987; Shanahan & Barr, 1995), as had others (Iversen & Tunmer, 1992). The combination of its one-on-one instruction and its extensive teacher development requirements make it an especially expensive intervention.
In 2022, Henry May and his colleagues at the University of Delaware reported a fascinating study – extensive and high in quality. It was a follow up to an earlier study they had conducted evaluating the effectiveness of RR. That earlier study found clear learning advantages for the children enrolled in the program (May, Sirinides, Gray, & Goldworthy, 2016).
The study, first reported in 2022 and published in 2024, evaluated the long-term benefits for the program and the results weren’t pretty (May, Blakeney, Shrestha, Mazal, & Kennedy, 2024).
Reading Recovery advocates claimed that its students would become self-improving systems, no longer with a need for remedial support. A promise meant to allay the concerns about its costs. But according to the May study, these kids needed as much or more of those additional services as the untreated population. Evidently, there was no saving at all.
Even worse, the RR kids did less well in reading by grades 3 and 4. Their achievement levels were lower than those of the comparison kids. In other words, RR hadn’t improved their ability to keep learning. If anything, it appeared that it had somehow made them less able to keep up with their classmates.
That was what I wrote about in that blog that I thought would be my last word on RR.
However, in the past few weeks, I’ve received multiple queries about the May study. Each was as skeptical as I had been about the possibility that RR somehow disadvantaged these children. By the end of grade 1, the RR kids were outperforming the reading of similar kids who hadn’t had the advantage of RR instruction. How could that be a problem?
Some emails questioned the quality of the study. For example, as is often true with longitudinal research, attrition levels were high. Large potentially biasing losses of subjects can undermine the trustworthiness of a study. In this case? There were still 15,000 kids in the study and various analyses showed the unlikelihood that attrition had affected the results.
The results seem screwy – I’ve never seen this kind of long-term negative result from any instruction – but there was nothing screwy about the study.
One questioner raised the best query of all:
“Something very weird is going on when first graders who got a lot of one-on-one help from an expert teacher end up going backwards. What’s your theory?”
I didn’t really have an explanation. I had taken the study to mean that RR didn’t have the lasting power its advocates claimed and treated the underperformance as a quirky anomaly.
But what if it weren’t? What if RR made it harder for kids to keep learning well?
One explanation, this one particularly popular with RR opponents, is that intensively teaching students an approach to reading that is inconsistent with how we read, you know three-cueing, may have long-term negative consequences.
That may be, but one could level the same kind of argument at some phonics instruction.
Let’s face it, we don’t read by sounding out words laboriously, and studies do show that kids who come to expect too much regularity or consistency in our orthography don’t do as well as those who recognize English orthography as a more complex and conditional system.
The benefit of phonics is that it guides kids to pay attention to spelling patterns and alters their memory in ways that makes it possible for words to be easily remembered. Any youngster who perseverates on the sounding out part or who rigidly expects letters to always be related to certain phonemes wouldn’t be likely to make the same gains as kids who have a better purchase on the nuances of our spelling system.
I doubt that either three-cueing or sounding out usually have those negative consequences – kids wisely don’t follow what adults say that slavishly.
But what about a child who is working one-on-one with an adult for 30 minutes a day for 12 weeks? That might be somewhat a more persuasive circumstance. Perhaps with that kind of intense instruction more kids might be persuaded to approach words through context rather than through the spelling.
I’ve never been a big fan of this explanation, but it would be unreasonable to dismiss it out of hand.
Another possible explanation is that kids identified for inclusion in RR may differ in some unidentified qualitative way from those kids who are ever so slightly higher performing in reading. That could mean that RR hadn’t really done any harm in the long run, just that the kids it serves are in especially great need, which makes it hard for them to keep up.
Of course, that would also mean that RR is not a sufficiently effective program for those lowest performing kids, which is a serious problem since those are the kids it targets.
In 1995, I pointed out that reading problems have two possible sources: something internal to the child (organic or neurological differences) or something external to the child (environmental problems including poor teaching). It was evident to me at the time that RR was increasing students’ ability to read but that it wasn’t altering either students’ ability to learn nor the kids’ environment. If youngsters had perceptual problems or an especially weak memory, they would still suffer these ills, even though they would have slightly higher reading scores. Likewise, if mom and dad weren’t encouraging academic success at home, RR wasn’t altering that either.
Despite the claims of RR advocates there was no reason to think the program would make kids successful in learning to read from then on.
But that wouldn’t explain why they would do even worse than comparable kids who didn’t get the treatment. That’s why it would be necessary to posit some unidentified factor that makes those lowest dozen kids different from everyone else. Whatever that unknown difference might be – internal or external – it should be obvious from the May study that RR was insufficient to adequately address it. That would argue against the heavy concentration of resources demanded by RR, since students with such stubborn learning problems should be expected to need ongoing instructional support.
Currently, we rarely plan for continuous need. Interventions usually aren’t designed to provide multi-year support, so our lowest progress kids tend to end up with a patchwork of programs, much of it repetitious, and all of it poorly coordinated.
Personally, I think this is the most likely of my explanations. Nevertheless, this one, like the others is just a hunch, not a research finding.
A third possibility for the odd results reported in the May et al. study could be that the degree of support offered by RR is so extensive that it ultimately undermines kids’ confidence, leading to “learned helplessness” and its consequences (Maier & Seligman, 2016). These kids know they are getting a substantial amount of help – 60 or more half-hour lessons delivered one-on-one is substantial. That so much help fails to alter their course – despite RR they still require remedial assistance – that could be especially discouraging.
It’s easy to imagine students concluding from their lack of success from so much help, “I must be really dumb.”
That explanation highlights something important about any remedial teaching. It is essential that we set specific goals for remediation, goals shared with the students; goals that we are serious about accomplishing. That way kids will have very specific ideas of the potential value of what they are being provided, and they’ll be aware of any real learning gains that they make.
What we shouldn’t do is make the kinds of claims offered by Reading Recovery advocates – that the students’ reading problems will be extinguished once and for all. Likewise, I’m wary of any program that touts learning guarantees. In medicine, it’s considered unethical to promise cancer cures and I think we should be equally cautious when it comes to assuring learning gains.
So those are three conjectures about why a program like RR might appear to slow student learning in the long run: (1) it may have slowed student learning by inculcating bad reading habits; (2) it may only look like it interfered with learning because it failed to address the real needs some students have; (3) it may somehow have undermined student confidence making them more likely to withdraw or quit trying.
I don’t know if any of these explanations holds water – but thinking about those possibilities suggests some important insights about remedial instruction.
In any event, kids enrolled in Reading Recovery did less well in the long run than similar kids who didn’t receive the program. No matter the reason, those sad results recommend caution. If I were calling the shots in a school district again, I’d discontinue the program until and unless I received a satisfactory explanation.
References
Maier, S. F., & Seligman, M. E. (2016). Learned helplessness at fifty: Insights from neuroscience. Psychological Review, 123(4), 349–367. https://doi.org/10.1037/rev0000033
May, H., Blakeney, A., Shrestha, P., Mazal, M., & Kennedy, N. (2024). Long-term impacts of Reading Recovery through 3rd and 4th Grade: A regression discontinuity study. Journal of Research on Educational Effectiveness, 17(3), 433–458. https://doi.org/10.1080/19345747.2023.2209092
May, H., Sirinides, P., Gray, A., & Goldsworthy, H. (2016, March). Reading Recovery: An evaluation of the four-year i3 scale-up. A research report. Newark, DE: Consortium for Policy Research in Education.
Shanahan, T. (1987). Review of “The Early Detection of Reading Difficulties” (3rd ed.), (by Marie Clay). Journal of Reading Behavior, 19, 117–119.
Shanahan, T., & Barr, R. (1995). Reading recovery: An independent evaluation of the effects of an early instructional intervention for at-risk learners. Reading Research Quarterly, 30(4), 958–996. https://doi.org/10.2307/748206
Comments
See what others have to say about this topic.