Is In-School Free Reading Good for Children? Why the National Reading Panel Report is (Still) Wrong
Submitted for Publication
May be shared and cited
In Krashen (2001) I reviewed the section of the National Reading Panel (NRP) report (NCIHD, 2000) that dealt with fluency. I argued that in their review of in-school free reading research, the NRP missed a number of studies (they included only 14 comparisons of in-school free reading and regular instruction; I found 53), and made serious errors in reporting the studies they did include. I noted that some of the studies showing no difference between readers and comparisons involved students that were already advanced and had already established a reading habit. I pointed out that the NRP did not include long-term studies, which I found to be more supportive of SSR than short-term studies, and they also included one study in which students were highly constrained on what they could read. I argued that the case for free reading rests on more than experimental studies, that case histories also provided compelling evidence for the power of reading. I concluded that the evidence in support of free reading in school was strong, contrary to the panelís conclusion.
Shanahan (2004) and Stahl (2004) both members of the NRP, attempted to respond to my criticisms. Their papers contain serious misrepresentations of my position as well as inaccuracies in reporting the literature.
"JUST LETTING KIDS READ ..."
Shanahan (p. 245) states my position as follows: "Krashen (2001) argued that just letting kids read is better than instruction ... ".
This is not an accurate description of what I wrote. I discussed "sustained silent reading" (SSR) and other in-school free reading programs. These programs include more than "just letting kids read." They set aside time to make sure children have a chance to read, provide access to good books, and do things that encourage reading. I strongly suspect that that "just letting kids read" is better than many kinds of instruction, but SSR does more than that.
There is evidence that many of the strategies used in SSR do in fact result in more reading. Modeling reading, or teachers reading while children are reading, results in more reading (Wheldall, and Entwhistle, 1988), as do providing more access to reading material, and activities done in conjunction with SSR, such as teachers recommending books, and teachers reading aloud to children (Von Sprecken and Krashen, 1998).
Did I say that free reading is better than instruction? I actually concluded that it was at least as good and often better than "regular" instruction, the instruction provided to comparison groups in the studies, which is generally instruction based on skill-building. Shanahanís interpretation of my position can be interpreted as saying that I am opposed to all instruction, which is false.
Similarly, According to Shanahan (p. 246) I claim that "teachers prefer just having students read instead of teaching them ...". I did not say that. I said teachers prefer free reading to what usually goes on in language arts and reading classes, i.e. regular instruction.
Shanahan refers to the comparison groups in these studies as utilizing "some impoverished form of teaching, e.g. just assigning random worksheets" (p. 246). I think it is highly likely that skill-building was involved in the comparison groups, but there is no evidence that the worst of these classes were selected to serve as comparisons in these studies, as Shanahan suggests.
WHAT "NO DIFFERENCE" MEANS
Many studies comparing in-school reading to regular instruction show no difference in gains in reading comprehension between the two groups. For Shanahan, a finding of "no difference" between in-school free reading programs and comparison groups is "uninformative," because the failure to find a difference could be due to factors other than the efficacy of one of the treatments, e.g. poor application, not enough of a treatment, etc.. I agree that such factors can play a role, but such results are an invitation to closer analysis. What is interesting is that studies that do show a difference consistently show the same results. This is true in my analysis, in which I found negative results (in-school free reading worse) in only three comparisons out of 53, and in the National Reading Panel report, which found no negative studies.
I have concluded that long-term studies are more likely to show positive results, results in favor of in-school reading. Findings of "no difference" are more common in short-term studies. Thus, Shanahan may be correct in suggesting that not enough of a treatment is the problem: The "no difference" result in some studies appears to be due to not providing enough reading. The NRP did not consider any studies lasting longer than one academic year, nor did they consider the possibility that length of treatment may be a factor.
This generalization makes sense. Some in-school reading programs are very short, as short as two months or ten weeks. I have seen SSR programs and I now know what most teachers know: The first few weeks there is little reading going on Ė children havenít found a good book yet.
Another reason, as mentioned earlier, is that subjects in some studies were already avid readers; in-school reading programs are, it is predicted, of greatest benefit to less mature readers, to provide reading exposure and to get them interested in reading.
THE SCORE: 24 to 3
If we discard the "no difference" studies in the NRP report, we are still left with a good case for in-school free reading: 4 comparisons positive, none negative. If we discard all the no difference studies from my review, we are left with 24 positive studies and three negative studies. And all three negative studies had a short duration, less than seven months. This is a very strong case for in-school free reading.
A statement by Shanahan (p. 248) could be interpreted as my showing support for Accelerated Reader, a reading management program that tests children on what they have read and rewards children for points earned on the tests. It is not clear that Shanahan intended to say that I support Accelerated Reader, but just for the record, I donít. I have been critical of Accelerated Reader, and have concluded that there is no clear evidence supporting it (Krashen, 2004).
NO INDEPENDENT REVIEW?
Shanahan claimed that eight of the studies I claimed were omitted from the National Reading Panel report were not published in refereed journal papers. In Shanahanís view, this means "they had no independent quality review" (p. 248) I counted only five in Krashen (2001) not published in journals. Of the five, four underwent peer review: One appeared in the National Reading Conference Yearbook, one in conference proceedings of the International Reading Association, another from an edited collection published by the International Reading Association, and one from the Claremont Conference Yearbook.
AN ABSURD ACCUSATION
Shanahan claims that I selected studies that agreed with my views but ignored those that did not (p. 248, 249), an absurd and false accusation. This would be a serious violation of professional ethics. It would also be a stupid thing to do, as any critic with access to a library could easily find the studies with counterevidence.
Shanahan dismisses studies I cited that were done in Singapore, India, Fiji and Malaysia, saying that "Krashenís approach assumes that cultural differences donít matter" (p. 249). It is hard to see how cultural differences could modify the impact of reading on language and literacy development. In addition, the NRP panel cheerfully accepted the results of studies of phonemic awareness done in Israel, Norway, and Spain using languages other than English. The studies I cited all examined the impact of reading in English (on the development of proficiency in English as a foreign language). It is also of interest that these studies provided powerful support for the efficacy of free reading.
Shanahan also dismisses results of individualized reading programs published in the 1950ís and 1960ís because, he claims, they did not necessarily use SSR and also included "a great deal of reading instruction" (p. 249). These programs did indeed include some direct instruction, in the form of brief student-teacher conferences. What is crucial is that students in these classes certainly did in- school free reading, far more than students in the comparison classes (Duker, 1965).
BOND AND DYKSTRAíS STUDY
Shanahan claims that I failed to cite Bond and Dykstraís study (Bond and Dykstra, 1997) which, he says, also compared individualized reading and skills and failed to find a superiority for free reading. Bond and Dykstra did not compare skills and free reading in any of their comparisons.
LETíS INCLUDE OLIVER (1973)
Shanahan concludes that the NRP really missed only one study, Oliver Ďs 1973 study which found no difference between SSR and comparisons. As Shanahan notes, the duration of this study was only one month. Shanahan once again ignores my conclusion that SSR effects are more positive with longer-term studies.
Shanahan (p. 249) concludes that if Oliver (1973) were added to the tally, it "would have given the NRP 15 evaluations of methods for encouraging reading with only 3 of these coming out positive." He does not point out that the NRP did not find any studies in which SSR was worse than comparisons.
ARE THE "BEST STUDIES" CORRELATIONAL?
Stahl (2004) also comments on research on in-school free reading and my criticisms. He states that "In this area, the best studies are not experimental but are correlational (see Krashen, 2001)" (p. 206). I did not discuss correlational studies in the paper Stahl cites. My entire paper was devoted to experimental studies, with some case histories mentioned briefly at the end of the paper. My claim was that the experimental studies present a strong case for the efficacy of free reading in school.
Stahl claims that the NRP report on fluency shows that "nonmonitored reading, in the form of sustained silent reading (SSR) or similar approaches, was not shown to be effective in the experimental studies the panel reviewed" (p. 205). The official view of the NRP, however, is that the NRP "had no findings" about these programs, because of a what the NRP considered to be the low quality of the research (Shanahan, 204, p. 245). This is not the same as "not shown to be effective."
DO THEY REALLY READ DURING SSR?
Stahl also questions whether children in SSR actually engage in reading. He cites his own observations of one pair of children "who were taking turns and turning pages in a shared book, looking as if they were sharing reading. When I came close enough to listen to what they were saying, however, I heard them talking about what they were going to do that weekend" (p. 206).
First, Stahl is discussing shared reading, not sustained silent reading. Second, several studies have shown that if SSR classes are observed in the middle of the school year, about 90% of the students are reading (Von Sprecken and Krashen, 1998; Cohen, 1999; Herda and Ramos, 2001) and the probability of studentsí actually reading is increased when several factors are present, e.g. access to interesting reading material, teachers reading while children are reading, etc. (Von Sprecken and Krashen, 1998). Third, Stahlís conclusion is based on one observation of two children.
TOO EASY, TOO HARD?
Stahl also mentions another "failing" of SSR: " ... I have seen children read books that are very easy or very difficult for them, neither of which would aid in their reading development" (p. 206).
Concerning "easy books": Easy books can do readers a lot of good. First, what looks easy isnít always so easy. If a reader who reads at the sixth grade level reads a book that is "officially" at the fourth grade level, that book will still contain a substantial amount of language at or above the studentís level. Reading level is an average (thanks to Kathleen Sesplaukis for pointing this out to me).
Second, an easy book can provide the taste and background knowledge that will lead to and facilitate reading other books. Third, readers donít stick to light and easy reading Ė their reading tastes gradually develop (LaBrant, 1958). In addition, it has been argued that reading has to feel effortless for it to result in language development; studies indicate that a text needs to be about 98% comprehensible in order for it to help the reader acquire new vocabulary (e.g Hu and Nation, 1995).
Stahl prefers "hard" reading "scaffolded through repeated reading and/or teacher assistance" (p. 207). If that were necessary or even optimal, nobody would ever want to improve in reading. The way all of us improved in reading was extensive reading of texts that did not necessitate strain and suffering, and that were so interesting we were completely absorbed in the message. Delayed gratification is not necessary to learn to read and to improve in reading.
Concerning "hard" books: Children who appear to be reading books that are "too hard" may in fact be finding highly comprehensible sections of these books that are of real interest to them, skipping the parts they donít understand but getting meaning and enjoyment from the parts they focus on.
I have attempted to deal with all of Shanahanís and Stahlís points related to my criticisms. I remain unrepentant in my conclusion that in-school free reading is a very good idea.
Bond, G. and Dykstra, R. 1997. The cooperative research program in first-grade reading instruction. Reading Research Quarterly, 32 (4): 345-427
Cohen, K. 1999. Reluctant eighth grade readers enjoy sustained silent reading. California Reader 33(1): 22-25.
Duker, S. 1965. Individualized Reading: Readings. Metuchen, NJ: Scarecrow Press.
Herda, R., and F. Ramos. 2001. How consistently do students read during sustained silent reading? California School Library Journal 24(2): 29-31.
Hu, M. and Nation, I.S.P. (2000) Vocabulary density and reading comprehension. Reading in a Foreign Language 13(1): 403-430.
Krashen, S. 2001. More smoke and mirrors: A critique of the National Reading Panel Report on fluency. Phi Delta Kappan 83(2): 119-123.
Krashen, S. 2003. The (lack of) experimental evidence supporting the use of accelerated reader. Journal of Childrenís Literature 29 (2): 9, 16-30.
LaBrant, L. 1958. An evaluation of free reading. In Research in the three Rís, ed. C. Hunnicutt and W. Iverson. New York: Harper and Brothers, pp. 154-161.
National Institute of Child Health and Human Development (NICHD), 2000. Report of the National Reading Panel: Teaching Children to Read: An Evidence-Based Assessment of the Scientific Research Literature on Reading and Its Implications for Reading Instruction: Reports of the Subgroups. Washington, DC: NIH Publication 00-4654.
Oliver, M. 1973. The effect of high intensity practice on reading comprehension. Reading Improvement 10: 16-18.
Shanahan, T. 2004. Critiques of the National Reading Panel Report: Their implications for research, policy, and practice. In P. McCardle and V. Chhabra (Eds.) The Voice of Evidence in Reading Research. Baltimore: Brookes Publishing Company, pp. 235-265.
Stahl. S. 2004. What do we know about fluency? Findings of the National Reading Panel. In P. McCardle and V. Chhabra (Eds.) The Voice of Evidence in Reading Research. Baltimore: Brookes Publishing Company, pp. 187-211.
Von Sprecken, D. and Krashen, S. 1998. Do students read during sustained silent reading? California Reader 32(1): 11-13.
Wheldall, K., and J. Entwhistle. 1988. Back in the USSR: The effect of teacher modeling of silent reading on pupilsí reading behaviour in the primary school classroom. Educational Psychology 8, 51-56.