Friday, December 20, 2013

Is Diane Ravitch missing something?

Diane Ravitch is a highly respected education commentator and historian in the USA. She initially supported "No Child Left Behind" but later reversed her position.
High-stakes testing, "utopian" goals, "draconian" penalties, school closings, privatization, and charter schools didn't work, she concluded. "The best predictor of low academic performance is poverty—not bad teachers." (Why I Changed My Mind About School Reform)
I noticed a couple of books she has written which are very relevant to my research (research update here). I have just ordered her most recent book, Reign of Error: The Hoax of the Privatization Movement and the Danger to America's Public Schools. It contains chapters on test scores (both local to the USA and international comparisons such as PISA), the achievement gap, poverty, merit pay, teacher tenure, Teach for America and Michelle Rhee. The gloves are off, it's education war in the USA.

This made me curious about her attitude to Direct Instruction, so I did an advanced search of her blog using that phrase, "Direct Instruction" site: It appears that Diane doesn't write much about Direct Instruction but she does publish opponents of DI, such as Stephen Krashen, on her blog.

But what I found most interesting was that a supporter of DI, with the moniker Eded, challenged Stephen Krashen and IMO Prof Krashen did not provide an adequate response. Here is the exchange, including the original Stephen Krashen material initially posted by Diane Ravitch, A Literacy Expert Opposes the Common Core Standards
Stephen Krashen is a professor emeritus at the University of Southern California, where he taught linguistics.

He comments here in response to an earlier post about the Common Core standards:
What this excessive detail also does is
(1) dictate the order of presentation of aspects of literacy
(2) encourage a direct teaching, skill-building approach to teaching.

Both of these consequences run counter to a massive amount of research and experience.

There is very good evidence from both first and second language acquisition that aspects of language and literacy are naturally acquired in a specific order that cannot be altered by instruction (C. Chomsky, 1969, The Acquisition of Syntax in Children from 5 to 10. Cambridge: MIT Press; Krashen, S. 1981, Second Language Acquisition and Second Language Learning, Pergamon Press, available at

There is also very good evidence that we acquire language and literacy best not through direct instruction but via “comprehensible input” – for literacy, this means reading, especially reading that the reader finds truly interesting, or “compelling.” (Krashen, S. 2010.The Goodman/Smith Hypothesis, the Input Hypothesis, the Comprehension Hypothesis, and the (Even Stronger) Case for Free Voluntary Reading. In: Defying Convention, Inventing the Future in Literacy Research and Practice: Essays in Tribute to Ken and Yetta Goodman. P. Anders (Ed.) New York: Routledge. 2010. pp. 46-60. Available at
Now here is the exchange between Eded and Prof. Krashen in the comments:
December 27, 2012 at 12:24 am
Regarding the second point, I’m not aware of any research to suggest that direct instruction is counter to research.

Please have a look at papers at, my books, as well as the work of Frank Smith, Kenneth Goodman and others.

Dr. Krashen – thanks very much for posting the link to your website in response to my comment earlier. I have reviewed a few of your articles, and unfortunately I’m not seeing any research demonstrating that direct instruction is ineffective. I do see arguments presented, with some research, that motivation is important when learning to read, along with opportunities to meaningfully engage with reading. However, I haven’t seen any studies which contradict the massive body of evidence supporting direct instruction in the “Big 5″ areas of reading (see for a good list of research).

Could you perhaps provide a reference to a research article which specifically examines direct instruction vs. non-direct instruction instructional methods, and shows a greater impact of non-direct instruction methods on general reading outcomes (e.g., measures of reading fluency, comprehension)? It may be more helpful to evaluate your claims more specifically, rather than talk in broad generalities.

To the general public reading, I would highly encourage you to view the research link above and draw your own conclusions regarding direct instruction, as Dr. Krashen (and apparently Dr. Ravitch) are in the extreme minority when it comes to perceptions regarding the literature base of direct instruction.

Please keep looking. Many of us have published research showing the extreme limitations of direct instruction in phonemic awareness, phonics, vocabulary, grammar, in direct response to the claims of the big 5 and National Reading Panel, and supporting the hypothesis that PA,phonics, vocabulary, grammar and competence in text structure are the result of reading (especially self-selected reading). Much of it is summarized in one place: Please see Comments on the LEARN Act, (A lot of it has been published in major journals, eg Phi Delta Kappan, Reading Research Quarterly, Garan’s papers in Language Arts, Kappan.)

Thanks for your response Dr. Krashen. Certainly one of the difficult aspects of this discussion is that it’s so broad. It’s not very easy to making sweeping generalizations about broad categories of interventions, as any particular side can start pointing at particular niches in the research, or studies within those niches, to prove points.

As such, I’ll respond to your comments on the LEARN Act specifically related to Phonemic Awareness (PA). Your main citation is a review of studies cited by the NRP about phonemic awareness, where you cite an insignificant effect of PA training on reading comprehension. In response, I’d direct you to this meta-analysis which shows a moderate effect size for PA on reading skills, and a large effect size for PA training on phonemic awareness skills:

I’d also point out a limitation of the parameters of your meta-analysis:

You only examine the effect of PA training on reading comprehension, as opposed to more component skills such as decoding, word reading, and reading fluency. It is entirely possible that PA training would have little or no direct impact on reading comprehension in later years of a child’s educational career, but have a more significant impact on more basic, foundational skills such as decoding. As such, it may be that phonemic awareness training is not sufficient in producing effects related to reading comprehension, and perhaps not even necessary with some (or many) kids, but it may nevertheless be a necessary component for some struggling readers in terms of acquiring beginning reading skills. As such, citing evidence that PA instruction fails to single-handedly produce long-term reading gains is not evidence that PA training is unnecessary.

Consider this analogy: a beginning swimmer receives instruction on how to breathe properly, but receives no additional swimming instruction. Is breathing instruction sufficient to producing good swimmers? No. Do all good swimmers breathe well? Mostly. Did all good swimmers learn, through explicit instruction, to breathe properly? No. Is any of this evidence that explicit instruction related to breathing properly is unnecessary or unhelpful to beginning swimmers, particularly those who struggle with breathing? Absolutely not. In fact, instruction on breathing may be absolutely critical to swimming, but may show little if any effect on a swimmer’s ability to swim a 500 meter butterfly stroke fluently, as beginning breathing is not sufficient to produce those gains.

As I mentioned before, it’s very difficult to discuss in blog comments a topic so wide as “direct instruction.” As such, my main point here is not to debunk your entire statement that there is no support for direct instruction as such a discussion would have to be much larger. Rather, my point is to highlight to other readers that your assertions (and Diane’s assertions) about direct instruction are not “givens” in research, that most folks do support the use of direct instruction, and that your research links/comments are not without challenge.

I’d also like to add a note of thanks for your willingness to engage in discussion on this blog – too often there is a gap between research and practice, and your willingness to engage in discussion with the “average reader” is a testament to your desire for research to be actually used rather than simply created. I’d also welcome follow-up comments and challenges, as I think those reading this blog post would be most informed by a more specific discussion of the research, as opposed to general statements about broad categories of interventions.

The failure to find a clear relationship between PA training and reading (reading comprehension) is consistent with the meta-analysis you cited. Also, there are other arguments, eg: some people learn to read quite well with very little PA, PA develops without instruction, adult illiterates have low PA, then their PA improves after they learn to read. Also we have to ask how millions of people learned to read before experts “discovered” PA. (We have made similar arguments for PA in second language development,Krashen, S. and Hastings, A. 2011. Is Phonemic Awareness Training Necessary in Second Language Literacy Development? Is it Even Useful? International Journal of Foreign Language Teaching 7(1). Available at

Again, PA training may not have a significant, direct impact on the general outcome of reading comprehension, but may have an impact on component, foundation skills such as word reading. Simply because an intervention is not sufficient in producing a general outcome does not mean that it isn’t a helpful component. There has been a causal link established between PA training and development of word reading skills, between word reading skills and reading fluency, and between reading fluency and reading comprehension. As such, it appears that PA training is mediated by variables such as word reading and reading fluency, and thus does have an impact on reading comprehension, if only indirectly.

With your “other route” concept – that some people learn to read quite well without PA training, consider mapping directions from your house to the mall. There are likely multiple routes, and the existence of one route does not imply the lack of existence of all others. You might take the highway, or the back roads. Those supporting PA training are not claiming that PA training is the only way to become a proficient reader, but that it is one route, particularly for struggling readers. In fact, it’s a common assertion that MANY readers do not require direct, explicit instruction in PA, phonics, etc., and that other, informal processes are at play.

In terms of PA developing without instruction, consider the case of a diabetic not producing insulin. The fact that many healthy people produce sufficient insulin is not evidence that diabetics do not. Similarly, that some children develop PA in a healthy manner is not evidence that other children do not.

In terms of PA developing as a result of other reading processes developing, I agree that there is not necessarily a unidirectional influence of PA (or many reading variables). For example, phonics instruction contributes to PA. However, that phonics instruction contributes to PA is not evidence that PA training does NOT contribute to phonics skills.

Again, bringing this discussion back to a point of relevance to this blog post, my hope is that those reading this discussion will not take for granted Diane’s comment that direct instruction “run[s] counter to a massive amount of research and experience.”

Credit/blame where credit/blame is due: the comment that direct instruction “run[s] counter to a massive amount of research and experience” comes from me, not Diane Ravitch.

Ah, I apologize – I was reading what I thought was her interpretation. Again, I’ve enjoyed the discussion. Looking forward to more exchanged in the future hopefully…

RE: PA, may have an impact on component, foundation skills such as word reading. Maybe word reading is not foundation skill but also a result of reading experience. (The comprehension hypothesis).

I think I understand you theoretically, but how do you make sense of evidence that phonics instruction improves word reading, that better word reading results in better fluency with connected text, and that fluency with connected text is what (in part) enables comprehension?

Complex phonics, word reading, fluency are all the RESULT of real reading for comprehension.

Dr. Krashen, in response to your last comment in our discussion above about phonics, word reading, and fluency being the result of comprehension as opposed to building toward comprehension, I’d again return to my “multiple pathways” comment: Some if not many children do not require explicit phonics (or other) instruction to read fluently and comprehend – they may independently acquire those skills, facilitated in part by being provided motivational and engaging reading contexts. However, with struggling readers (and others as well), research has suggested that explicit instruction in foundational reading skill areas (e.g., phonics) can lead to acquisition of more advanced skills such as reading fluency.

In other words, we both seem to be right, in that kids seem to be able to learn to read with both direct instruction and non-direct instruction. The question then becomes which modality to use in different situations, which would be directly answerable by research investigating the differential effects of DI vs. non-DI approaches in different instructional contexts. I am familiar with a variety of studies which support DI in across contexts, and am not familiar with any studies which examine DI and non-DI approaches side-by-side, and find greater effects for non-DI approaches. Could you provide any links to studies that would suggest favorable results for non-DI approaches over DI approaches?

Stephen Krashen
December 27, 2012 at 9:09 pm EdEd to avoid clogging up Diane’s blog, please write me off line and I will send you some sources and papers. My email:
Reading wars morph into research wars and it's hard to keep track of it all. But I thought that Eded had the better of this exchange, with Stephen copping out at the end. From my reading of the evidence it does favour Direct Instruction over the Whole Language views of Stephen Krashen. For instance see Kerry Hempenstall's essay, Literacy assessment based upon the National Reading Panel’s Big Five components (long, 46pp), for a very thorough review of the evidence.

Diane Ravitch's position is that poverty trumps teaching methodology. That is a powerful argument but we can't put on hold better education until the poverty question is solved. The education establishment and teacher unions have a duty to study the evidence and deliver the best possible education to poor students in the here and now.

No comments: