Stephen Krashen

The Phi Delta Kappan (in press)
 

"An Urban Legend is usually a (good / captivating / titillating / engrossing / incredible / worrying) story that has had a wide audience, is circulated spontaneously, has been told in several forms, and which many have chosen to believe (whether actively or passively) despite the lack of actual evidence to substantiate the story."1
 

I wish to add another Urban Legend to those that already exist, a legend that in my opinion ranks with the legend of alligators in the sewers of New York City.2 I will refer to it as the "Plummet Legend" After whole language was introduced in California in 1987, test scores "plummeted" to the point where California's fourth grade children were last in the county in 1992. The sudden plummet never happened. It is an urban legend, a good/captivating story that has been told in several forms, and that many have chosen to believe them despite the lack of actual evidence.
 

The Plummet Legend has had serious consequences. It has led to the discrediting of the whole language approach to literacy and a strong movement promoting a "skill-building" approach.3 In this paper, I will try to show that the evidence does not support these this legend, and that it is inconsistent with the results of studies of literacy development.
 

DID TEST SCORES PLUMET IN CALIFORNIA?
 

Here is a more complete version of the Plummet Legend

In 1987 a group of whole language advocates took over the California Language Arts Framework Committee and brought in whole language. Phonics instruction and other forms of direct teaching were banned, and language scores plummeted to the point where California 's fourth graders scored last in the country in reading in 1992. California is now recovering from this damage, thanks to a rational, sensible phonics-based approach to reading.
 

This is not what happened. I served on the Language Arts Framework Committee in 1987. Phonics teaching was not banned. We simply proposed that language arts should be literature-based. This is hardly controversial. In fact, I regarded it as part of the definition of language arts. 
 

Did teachers change their ways in California? Nobody knows. There have been no empirical studies comparing methodology in language arts teaching before and after the 1987 committee met.
 

Did test scores decline? It is true that California fourth graders scored last in the country in the fourth grade NAEP examination in 1992. But this was the first time NAEP scores had been presented by state. It was assumed that there had been a decline, but there was no evidence that this was so no comparison with earlier test scores was made. McQuillan examined CAP reading comprehension scores from 1984 to 1990, which I present in table 1. There is no clear pattern of increases or decreases through these years, which leads to the conclusion that California 's reading problem existed well before "whole language" was introduced in 1987. There was no Great Plummet of 1987-1992. 4
 

TABLE 1. CAP scores in California1984-1990 

grade 1984 1985 1986 1987 1988 1989 1990

3 268 274 280 282 282 277 275

6 249 253 260 260 265 262 262

8 250 240 243 247 252 256 257

12 236 241 240 246 250 248 251

fromMcQuillan (1998) 
 
 
 

McQuillan also provides a convincing explanation for the low scoresThere is strong evidence that California 's poor performance is related to its print-poor environment. California ranks last in the country in the quality of its school libraries, and ranks near the bottom in public libraries. In addition, many of its children have very little reading material at home California ranked ninth in the country in the percentage of children ages 5-17 living in poverty in 1995, and near the bottom of the country in the percentage of homes with more than 25 books. 5 
 

McQuillan 's analysis was based on school library data published in 1990 and public library data published in 1995. There has been little change in California since that time
 

* A study published in 1990 reported that California's school libraries had 13 books per child, compared to the national average of 18. 6 In 2001, California elementary schools had only 12 books per child. Although some of this decline was due to pruning of old books, there is no sign of improvement. 7
 

* The same 1990 study reported 4595 students per school librarian in California, compared to the national average of 900 to 1.8 In 1998, California had 4673 students per school librarian. 9
 

* California's public libraries are not impressive. According to data published in 1997, California's public libraries have 1.9 volumes per capita, compared to the national average of 2.8. Only three states are worse. California's public libraries circulate 4.9 books per capita per year, compared to the national average of 6.6. Only ten states are lower.10
 

* California now ranks in the bottom eight of the country in terms of percentage of children ages 5-17 living in poverty.11
 

Print access variables are strongly correlated with NAEP reading scores. McQuillan reported a .85 correlation between measures of print access (books and other forms of print available in the home, school and community) and 1994 NAEP scores. 12 Even controlling for poverty, the correlation remained high (r = .63). 13 California 's problem is not whole language but a lack of reading material. 
 

Independent research supports McQuillan 's analysis. There is excellent evidence that children with more access to books read more and that children who read more make superior gains in literacy development. 
 

ACCESS > READING
 

There is a great deal of evidence showing that children with more access to books read more. Children with more books in the home read more. 14 Heyns reported that children who live close to public libraries read more than those who live far away.15 Morrow and Weinstein found that installing well designed library corners in kindergartens resulted in more use of books by the children during free play time.16 In a study of high school libraries, Houle and Montmarquette reported that students take more books out of school libraries that have more books and that stay open longer.17 McQuillan and Au (2001) reported that high school students did more reading when their teachers took them to the school library on planned library visits more often. 18,19
 

READING > LITERACY DEVELOPMENT
 

Until recently, it was considered to be obvious that actual reading itself helps readers get better at reading and helps them improve their vocabulary, grammar, spelling and writing. The U.S. Government disagrees. The National Reading Panel 20 concluded that there was insufficient research to support the hypothesis that reading itself is beneficial, and concluded that we should concentrate our efforts on phonemic awareness training, intensive phonics and having children read aloud so their errors can be corrected. 
 

The Panel overlooked a tremendous amount of research. The case for recreational reading is overwhelming. It consists of many case histories in which it is clear that reading was the causative factor in helping individuals increase their level of literacy development, cases such as Richard Wright, Malcolm X and Ben Carson.21 It includes correlational studies in which strong and consistent correlations are found between the amount of reading done and gains in reading development.22 It includes experiments in which readers show modest but reliable gains in vocabulary and spelling knowledge after only one or two exposures to an unfamiliar word in a meaningful context.23 
 

The case for recreational reading also includes studies of sustained silent reading. I reviewed the research on sustained silent reading and concluded that it worked. In 51 out of 54 comparisons, students who read for pleasure gained as much as or more than comparison students on tests of reading comprehension. Also, programs that lasted longer were more effectiveFor programs lasting one academic year or longer, those in sustained silent reading classes outperformed comparison students in eight out of ten comparisons, and in two other case there was no difference. The National Reading Panel report included only studies of sustained silent reading, included no long term programs, contained only a dozen comparisons, and misinterpreted and misreported some of the studies they did include.24
 

Studies done with younger readers. In an earlier publication I presented a narrative review of studies that claimed to compare the efficacy of whole language and "skills" approaches to teaching reading. I concluded that when whole language was defined correctly, that is as including a great deal of real reading, students in these classes performed as well as or better than children in skills classes on tests of reading comprehension, were equivalent to children in skills-based classes on tests of "skills" (e.g. reading nonsense words), and had more positive attitudes toward reading and read more on their own. 25
 

Once again, the federal government thinks otherwise. The National Reading Panel concluded that skills-based methods were superior to whole language methods (d = .31). 26 In my re-analysis, I considered performance on tests of reading comprehension, and also considered the amount of reading done and found an overall advantage for whole language (d = .17). For the four studies in which it was clear that one group did more real reading, the advantage for the readers was substantial (d = .70). 27
 

LIBRARIES > BETTER READING
 

If more access to books results in more reading, and more reading results in more literacy development, it follows that more access to books will result in more literacy development. Research confirms that this is the case. 28 A particularly important aspect of this research deals with the impact of libraries. Research on the impact of libraries over the last decade has shown that better school libraries, those with more books and better staffing, are associated with more literacy development. 
 

The seminal study in this area was done by Keith Curry Lance and his associates, who found that school libraries in Colorado with better staffing and better collections had higher reading scores, even when factors such as poverty and availability of computers were controlled. 29 These results were confirmed by other studies showing that states with better school and public libraries earned higher scores on the NAEP fourth grade reading examination. 30 Also, Warwick Elley reported a positive association between school library quality and reading achievement among children in 32 different countries. 31 Lance 's Colorado results have been replicated in several other states, by Lance himself as well as by other scholars. 32
 

THE ROLE OF PHONICS
 

The conclusions reached here do not exclude a role for the direct teaching of phonics. Frank Smith has argued that some conscious knowledge of sound-spelling correspondences can help make texts comprehensible. There are, however, severe limits on how much phonics can be taught directlythe rules are complex and have numerous exceptions. Smith argues that most of our knowledge of phonics is the result of reading, not the cause.33 Smith 's view is nearly identical to the view presented in Becoming a Nation of Readers often cited as supporting heavy, early phonics
 

" ... phonics instruction should aim to teach only the most important and regular letter-to-sound relationships ... once the basic relationships have been taught, the best way to get children to refine and extend their knowledge of letter-sound correspondences is through repeated opportunities to read. If this position is correct, then much phonics instruction is overly subtle and probably unproductive." 34
 

What does a low score on NAEP mean?
 

Recall that it was California fourth graders ' low scores on the NAEP examination that stimulated the movement to heavy phonics-based instruction. It is impossible to know for sure if a low score on the NAEP means that a child cannot read or lacks knowledge of important sound-spelling correspondences. In fact, I suspect that a substantial number of children who received low scores on the NAEP read reasonably well. The NAEP reading comprehension examination is not only a test of literacy, it is also a test of literature. A glance at the evaluation criteria reveals that readers have to be able to interpret passages the way an "educated" person would. It is quite possible to understand a passage perfectly well but have a nonstandard (or very creative) interpretation or way of answering questions. 35
 

An example of the scoring criteria for the 1992 NAEP reveals this. In a discussion of the "short constructed response" items of the NAEP, a sample passage was presented about the first paid woman umpire in baseball (Amanda Clement). Fourth grade students were asked "If she were alive today, what question would you like to ask Mandy about her career? Explain why the answer to your question would be important to know." Here are some answers that were considered "unacceptable"
 

"How old are you? Can I have a picture of you."
 

"Did you real like basket ball did you have any friends or fans. Was you ever at any basketball games? The reason I wold ask these questions is because I like basket, ball to. Was you ever a cherleader? What color is your hair because if she ever got lost or anything you or people would have to (know) what color here hair is."
 

The best answers, according to the authors, indicated "that the student has considered the more complex social or personal issues suggested by the passage." 36 
 

I am not arguing that the unacceptable examples presented here show high levels of literacy achievement. They don 't. But it is clear that the writers were not completely illiterate; they have obviously acquired basic sound-spelling correspondences. It should also be noted that 17% of California fourth graders wrote "unacceptable" answers to this question and 6% did not write anything. This means that 77% wrote answers considered of higher quality. 37
 

POST-SCRIPTWHAT HAPPENED AFTER 1992?
 

What has happened to California 's NAEP scores since 1992? Now that a heavy phonics and skill oriented approach is being aggressively pushed by the state government, have NAEP scores increased? Not so far. California's fourth graders scored 202 in 1992 on the NAEP, 197 in 1994 and 202 in 1998. (National norms for these years were 215, 212, and 215). 38,39
 

THE PLUMMET LEGENDSUMMARY
 

The Great Plummet of 1987-1992 never happened. California 's reading scores were low well before the Language Arts Framework Committee met in 1987. There is compelling evidence that the low scores are related California 's impoverished print environment. There is also strong and consistent evidence that the availability of reading material is related to how much children read, and how much children read is related to how well they read. The skills and testing hysteria that has gripped California and other states was unnecessary.
 
 
 

___________
 

1. Urban Legends Research Centre, www. urla.com

2. Other Urban Legends includeHumphrey Bogart was the original Gerber baby on their baby food ads, the FBI monitors public libraries and notes who reads "subversive" books, and my favoriteIf the entire population of China jumped up at the same time, the US would be swamped by a tidal wave. None of these are true. See www.urbanlegends.com for many others.

3. The more recent skill-building approaches to literacy have been labeled "balanced" approaches, balancing reading for meaning and skills. Coles (Gerald Coles, Reading UnmentionablesDamaging Reading Instruction While Seeming to Fix it. Portsmouth, NHHeinemann. in press) points out, however, that "a close look reveals that the comprehension end of the seesaw remains close to the ground for a long time"; the approach is essentially skill-building, with some real reading utilized as a means of practicing skills. For evidence, see the State of California 's Every Child A Reader, the report of the Reading Task Force, which contains very little mention of real reading, and birth-to-grave (actually K through 8) timelines for phonics instruction and the teaching of other skills.

4. Jeffrey McQuillan, The Literacy CrisisFalse Claims and Real Solutions. (PortsmouthHeinemann, 1998)

5. Ibid.

6. Howard D. White, "School library collections and servicesRanking the states," School Library Media Quarterly, vol. 19, 1990. pp. 13-26.

7. http//www.cde.ca.gov/library/libstats.html

8. White, op.cit.

9. http//www.cde.ca.gov/library/libstats.html

10. National Center for Education Statistics, Digest of Educational Statistics, (US Department of Education, 2001). table 422.

11. Ibid, table 20.

12. McQuillan, op. cit.

13. Poverty has profound effects of literacy development. This is consistent with the claim (see text) that access leads to more reading and more reading leads to better literacy development; children of poverty have far less access to books than more privileged children do (see eg. Susan B. Neuman and Donna Celano, 2001. "Access to print in low-income and middle-income communities." Reading Research Quarterly, 36(1)8-26.). McQuillan 's finding that literacy development is related to access even when poverty is controlled shows that access itself is an important factor in lack of literacy development.

14. Leslie Morrow, "Home and school correlates of early interest in literature," Journal of Educational Research, vol. 76, 1983. pp. 221-230; Susan B. Neuman, "The home environment and fifth-grade students ' leisure reading." Elementary School Journal, vol. 86, 1986, pp. 335-343; Vincent Greaney, and Mary Hagerty, M. "Correlations of leisure time reading." Journal of Research in Reading, vol. 10, 1987, pp. 3-20.

15. Barbara Heyns, Summer Reading and the Effects of Schooling. New York Academic Press, 1978). 

16. Leslie Morrow and Carol Weinstein, "Increasing children 's use of literature through program and physical changes." Elementary School Journal, vol. 83, 1982, pp.131-137.

17. Rachel Houle and Claude Montmarquette,. "An empirical analysis of loans by school libraries," Alberta Journal of Educational Research, vol. 30, 1984, pp. 104-114.

18. Jeff McQuillan and Julie Au. "The effect of print access on reading frequency," Reading Psychology, vol. 22, 2001, pp. 225-248.

19. Simply providing access, of course, is not always enough. Pack, in a study of children 's after-school activities, identified a group of children he labeled "library latch-key kids," children whose parents used the public library as a "free source of after-school care" from one to six hours per day. Pack reported that the children did "little more than ëhang out ' at the library" (p. 166). They did not read, but passed the entire time socializing with other children and playing on the computer (Sam Pack, "Public library use, school performance, and the parental X-factorA bio-documentary approach to children 's snapshots," Reading Improvement, vol. 37, 2000, pp.161-172). 

Modest interventions on the part of teachers and librarians, however, can remedy this. Many readers report that one positive experience with reading was enough to make them dedicated pleasure readers (a "home run" book experience; see Jim Trelease, The Read-Aloud Handbook. New York Penguin, 2001. Fifth Edition. For empirical evidence, see Debra Von Sprecken, Jiyoung Kim, and Stephen Krashen, "The home run bookCan one positive reading experience create a reader?" California School Library Journal, vol. 23 (2), 2000, pp. 8-9; Jiyoung Kim and Stephen Krashen, "Another home run." California English, vol. 6(2), 2000, p. 25.

There are many ways to help ensure that home run experiences will happen, e.g. readalouds (Trelease, op. cit.), modeling reading (e.g. reading while children are reading during sustained silent reading time; see Kevin .Wheldall, and Judy Entwhistle, "Back in the USSRThe effect of teacher modeling of silent reading on pupils ' reading behaviour in the primary school classroom," Educational Psychology, vol.8, 1988, 51-56) and interesting book discussions (the core of language arts), as well as providing time to read. There is consistent evidence showing that when students are provided time to read, they take advantage of it. When observations of sustained silent reading classes are made in the middle of the school year, and when students have adequate access to interesting reading, the vast majority of students in sustained silent reading programs are involved in reading during SSR time (Debra Von Sprecken and Stephen Krashen, "Do students read during sustained silent reading?" California Reader, vol. 32(1), 1998, pp. 11-13; Kera Cohen, Reluctant eighth grade readers enjoy sustained silent reading. California Reader, vol. 33(1), 1999, pp. 22-25; Rene Herda and Francisco Ramos, "How consistently do students read during sustained silent reading?" California School Library Journal, vol. 24(2), 2001, pp. 29-31.)

20. Reports of the National Reading PanelTeaching Children to Read. www.nichd.nih.gov/publications/nrp/report.htm.

21. Reviewed in Stephen Krashen, The Power of Reading, (Eglewood, CO Libraries Unlimited, 1993).

22. Stephen Krashen, "Do we learn to read by reading? The relationship between free reading and reading ability." In Deborah Tannen (Ed.) Linguistics in ContextConnecting Observation and Understanding. (Norwood, NJAblex, 1988). pp. 269-298.

23. Reviewed in Krashen, 1993, op. cit.

24. Stephen Krashen, "More smoke and mirrorsA critique of the National Reading Panel report on fluency." Phi Delta Kappan. vol.83,2001. pp. 119-123.

25. Stephen Krashen, Three Arguments Against Whole Language and Why They are Wrong, (Portsmouth, NHHeinemann, 1999).

26. Reports of the National Reading Panel, op. cit.

27. Stephen Krashen, The National Reading Panel comparison of whole language and phonicsIgnoring the crucial variable in reading. Talking Points. In press.

28. Krashen, 1988, op. cit.

29. Keith Curry Lance, L. Welborn, and Chrisine Hamilton-Pennell. The Impact of School Library Media Centers on Academic Achievement. (Castle Rock, CO Hi Willow Research and Publishing, 1993).

30. Stephen Krashen, "School libraries, public libraries, and the NAEP reading scores," School Library Media Quarterly, vol. 23, 1995, 235-238; McQuillan, The Literacy Crisis, op, cit.

31. Warwick Elley, How in the World do Children Read? (Hamburg International Association for the Evaluation of Educational Achievement, 1992).

32. Replication studies by Lance and his colleagues include a second Colorado study (Keith Curry Lance, Marcia Rodney and Christine Hamilton-Pennell, How School Librarians Help Kids Achieve StandardsThe Second Colorado Study. (San JoseHi Willow Research and Publishing, 2000a), and replications in Alaska (Keith Curry Lance, Christine Hamilton-Pennell, Marcia Rodney, Lois Petersen and Clara Sitter, Information EmpoweredThe School Librarian as an Academic Achievement in Alaska Schools. (JuneaAlaska State Library,1999), Oregon (Keith Curry Lance, Marcia Rodney, and Christine Hamilton-Pennell, Good Schools have Good Librarians. (Oregon Educational Media Association, 2001) and Pennsylvania (Keith Curry Lance, Marcia Rodney and Christine Hamilton-Pennell, Measuring to StandardsThe Impact of School Library Programs and Information Literacy in Pennsylvania Schools. (Greensburg, PA Pennsylvania Citizens for Better Libraries,2000b). In addition, similar studies have been done in Texas (Ester Smith, Texas School Libraries Standards, Resources, Services, and Students' Performance. (Texas State Library and Archives Commission. http//www.tsl.state.tx.us/ld/pubs/schlibsurvey/index.html, 2001), Massachusetts (James Baughman, School Libraries and MCAS Scores. http//artemis.simmons.edu/~baughman/mcas-school-libraries, 2000) and Indiana (NCEL, 2000. A Study of the Differences between Higher- and Lower-performing Indiana Schools in Reading and Mathematics. (North Central Regional Educational Laboratory, Oak Brook, Illinois, 2000). In all of these studies, poverty was a predictor of achievement. The number of books per student and amount of library staffing were also consistent predictors of achievement. In most studies, these relationships held even when poverty was controlled (Colorado replication, Oregon, Indiana, Massachusetts and Texas studies); in the Alaska and Pennsylvania studies, only staffing predicted reading scores, and in the Alaska study this was the case only when poverty was not controlled. The Colorado replication, Pennsylvania, Texas and Oregon studies utilized tests of reading comprehension only. In Alaska and Indiana, math was included and in Massachusetts, the measure included language arts, math and science. 33. Frank Smith, Understanding Reading. (Hillsdale, NJErbaum, 1994)

34.Richard Anderson, Elrieda Heibert, Judith Scott, and Ian Wilkinson, Becoming a Nation of Readers. (WashingtonNational Institute of Education, 1985), p. 38.

35. Judith Langer, Jay Campbell,, Susan B. Neuman, Ina Mullis, Hillary Persky,and Patricia Donahue, P. Reading Assessment RedesignedAuthentic Texts and Innovative Instruments in NAPE 's 1992 Survey. (Washington, DC National Center for Educational Statistics, Report No. 23-FR-07, 1995). p. 74.

36. ibid. p. 73.

37. ibid.

38. Patricia Donahue, Kristin Voelkl, Jay Campbell, and John Mazzeo, J. NAEP 1998 Reading Report Card for the States. NCES 1999-500. (Washington, DCUS Department of Education, 1999)

39. SAT9 scores have increased in California since 1998. Does the increased emphasis on phonics deserve the credit? Recall that NEAP scores up to 1998 reveal no increases for fourth graders in California, and there have been no NEAP results analyzed for individual states since that time. One could argue that the impact of phonics only became felt after 1998, because it took time for changes to be made. The new approaches to reading (California Department of Education. Every Child a Reader. SacramentoCalifornia Department of Education, 1995) were only announced in 1995. It is premature to grant intensive phonics instruction credit for SAT9 increases, however. (1) There are other plausible reasons for the gains. The SAT9 was introduced in California in 1998. Research has shown that after new tests are introduced, test scores rise, which is why commercial tests need to be recalibrated every few years (Robert Linn, Elizabeth Graue, and Nancy Sanders, "Comparing state and district test results to national normsThe validity of claims that ëeveryone is above average. '" Educational MeasurementIssues and Practice, vol. 10, 1990, pp. 5-14.). Typical test score inflation is about 1.5 to two points per year, which accounts for a great deal of the gains seen in California. "Test inflation" is especially prevalent in California where the same test has now been given for four years in a row, punishments for lower scores are severe, and rewards for higher scores are generous. This pressure has resulted in districts using unusual and extraordinary means for raising test scores, some of which have nothing to do with increased competence. Among the bogus means of increasing test scores are extensive training in certain test-taking skills and selective testing, i.e. excluding low scoring children from taking the test. San Francisco Chronicle reporter Nanette Asimov ("Test scores up, test-takers downLink between participation, improvement on school exam promotes concern," San Francisco Chronicle, Saturday, July 22, 2000) reported that selective testing may have occurred in California. Asimov reported that in many cases in which SAT9 scores increased from year to year, the number of students tested decreased. According to Asimov, "questionable pairings" appeared in 22 San Francisco Area school districts. And of course some test-taking skills will raise scores without an increase in competenceIf there is no penalty for guessing, for example, simply encouraging guessing will raise scores. Use of these means to raise scores is like claiming to raise the temperature of the room by lighting a match under the thermometer. (2) There has been no study made of teaching practice before and after the new phonics emphasis, and no attempt to search for a specific link between increased phonics teaching and test score improvements. A detailed look at experimental research comparing intensive versus "regular" phonics teaching shows that the phonics advantage is limited only to phonetically regular words presented in isolation, with only a weak impact on tests of reading comprehension for older children (grades 2-6), tests that contain words with irregular sound-spelling correspondences (see Garan 's discussion of the National Reading Panel report; Elaine Garan, "Beyond the Smoke and MirrorsA Critique of the National Reading Panel Report on Phonics," Phi Delta Kappan, vol. 82, 2001, pp. 500-506.); SAT9 tests are given only to grades 2 and higher.