Print this page

Broad Claims From Slender Findings

It seems, suddenly, de rigeur for advocates of particular approaches to early literacy instruction to assert that the "research says" particular curricular and instructional policies are necessary or, at least, appropriate.

While promoters of various sorts of educational efforts and initiatives have long used and abused "research" in their advocacy campaigns, only more recently have we seen legislation and other policy mandates that limit professional choices to "research-based" or "proven" methods, programs, and materials.

In our study of state educational policy making we have noted a number of instances where assertions have been made about what the "research says" in offering support for particular policies. However, in many of these instances the available research seems to have been distorted or exaggerated in order to better leverage particular policy proposals. The role of decodable texts in beginning reading provide but one example of such exaggeration.

Some advocacy on the role of decodable texts

The Learning First Alliance (12 professional groups, but not NRC, IRA, NCTE) white paper, Every Child Reading (1998), focused on "reading practices based on strong research findings..." (p. 54)

Early in first grade, a child's reading materials should feature a high-proportion of new words that use the letter-sound relationships they have been taught... p. 57

The report then summarizes the Juel & Roper/Schneider (1985) as follows:

The group of children who used texts with a high proportion of words they could sound out learned to read much better than the group who had texts in which they could rarely apply the phonics they were being taught. p. 57

The problem is that the Juel & Roper/Schneider study found no significant differences on the ITBS or the Bryant Phonics Test at the end of the year. They concluded:

The interpretation of the results of this study do not constitute advocacy for any one specific approach to beginning reading instruction. (p.150)

This suggests that the authors of the LFA report have a) never read the Juel & Roper/Schneider paper but relied on inaccurate citations by others who had distorted the results, b) purposely distorted the results themselves, or c) did not comprehend the research paper. In any event, the fact that a dozen professional organizations signed on to this "research-based" statement is amazing given the distortion of the research offered in this and other instances in the white paper.

The Center for the Future of Teaching and Learning (1996) issued and distributed a report entitled, 30 years of research on reading: What we now know. In that report, authored by Bonita Grossen, the issue of decodable texts again arises as can be seen in the following excerpt.

The findings of the NICHD research emphasize that children need extensive practice applying their knowledge of sound-spelling relationships to the task of reading as they are learning them. This integration of phonics and reading can only occur with the use of decodable text. Decodable text is composed of words that use the sound-spelling correspondences that children have learned to that point and a limited number of sight words that have been systematically taught. p.11

This excerpt suggests that the NICHD research reviewed in the paper demonstrated the efficacy of the use of decodable texts. However, no NICHD study has isolated or systematically manipulated decodable texts. In fact, as Fletcher and Lyon (1998) point out, the role of decodable and predictable texts are simply "not well understood" at this point in time. They also cite the CFTL report as an example of the exaggeration of research findings that they see as all too common.

The American Federation of Teachers (1998) in an editorial opening a special issue of their magazine, American Educator, notes:

We must find ways to ensure that research based on scientific principles reaches those in the front lines of education... It was with this need in mind that we planned this issue… p. 5

Inside, in an article by Louisa Moats, we find the following:

In systematic code instruction, decodable books are used that are aligned with the sound-symbol associations taught in the lesson...to provide practice reading words that have specific spelling patterns or letter-sound correspondences and to encourage sounding words out. p. 47

Summary: Advocacy

There seems no recent shortage of advocates for the use of decodable text – only a shortage of research to support such advocacy (Allington, 1984; Allington & Woodside-Jiron, 1998a, 1998b).

But advocacy is not necessarily linked in any direct way to educational policy making. In the case of decodable text advocacy, however, we found similar distortions of the research offered in state policy documents and then observed a translation of these into policy mandates.

Policy making on decodable texts

Texas educational policy making featured the distortion of research in shaping state policy on the use of decodable texts.

Texas Education Agency (1997), Beginning Reading Instruction: Components and Features of a Research-based Reading Program continued the "research supports the use of decodable texts" theme with the following quote:

Research asserts that most children benefit from direct instruction in decoding, complemented by practice with simply written decodable stories. p. 8

Later in the booklet, in a section on "Sound instructional materials" the assertion that research-based criteria would support the selection and use of decodable texts is again offered. However, no specific research studies are linked to the various assertions in this booklet. Thus, it is difficult to know which studies were misinterpreted or distorted in developing this advocacy for decodable texts.

Nonetheless, shortly after distribution of this booklet, the Texas State Board of Education approved the new state standards the Texas Essential Knowledge and Skills, wherein knowledge and skills component 110.3b7G states:

The student is expected to use letter-sound knowledge to read decodable texts.

This student standard was then translated into a textbook content standard for the upcoming basal reading series adoption in the 1997 Proclamation of the State Board of Education Advertising for Bids on Instructional Materials, English Language Arts and Reading, Grade 1 content requirements.

1.7G. Use letter-sound knowledge to read decodable texts (engaging and coherent texts in which most of the words are comprised of an accumulating sequence of letter-sound correspondences being taught). p. 7

So Texas first-grade teachers will soon receive a publicly funded supply of decodable texts (now defined in Texas as texts in which at least 51% of the words are decodable given the phonics skills taught). So too will the teachers in CA and other states, even though no research exists that actually demonstrates any longer-term benefits on reading achievement of decodable text use (See Allington & Woodside-Jiron, 1998, in press, for further analyses of CA and other states on this and related issues.)

Summary: State policies on decodable texts

The policy making process observed in Texas has been replicated (with minor variations) in other states. Yet even a cursory reading of the research studies cited would have shown that no bald assertions about the benefits of the use of decodable text were warranted.

In fact, in many cases (e.g. Juel & Roper/Schneider; Beck & Juel; Anderson, Hiebert, Scott & Wilkinson) the authors cited made this quite clear in the papers. I can only conclude that no one in a position of actual policy making authority actually bothered to read the research that was cited.

An alternative explanation is that distortion of the research and deception in policy making were the intended goals. If the research were not so clearly lacking in demonstrations of positive effects with use of decodable texts, if so many authors cited by the advocates of decodable texts had not indicated the lack of clear support for the use of decodable texts, and if there was no body of available research illustrating potential negative effects of reliance on decodable texts, then it would be easier to understand how such a distortion of the research could have been so easily and widely accepted.

But the research offers no clear answers, the authors generally said as much in their reports, and there exists a body of research pointing to potential negative effects. So what gives?

This seems but another example of educational faddism following the pendulum swings that have characterized reading education for the whole of this century (Langer & Allington, 1992). A more pointed question, perhaps, is why the professional associations for literacy researchers (NRC, IRA, NCTE, AERA) were so ineffectual in responding to the distortions, exaggerations and misrepresentations of the research.

What strategies might be developed to ensure a more adequate response the next time around?

Search