reflection

S235

Reading Research Quarterly, 55(S1) pp. S235–S247 | doi:10.1002/rrq.349 © 2020 International Literacy Association.

A B S T R A C T Recently, the term science of reading has been used in public debate to promote policies and instructional practices based on research on the basic cognitive mechanisms of reading, the neural processes involved in reading, computational models of learning to read, and the like. According to those views, such data provide convincing evidence that explicit decoding instruc- tion (e.g., phonological awareness, phonics) should be beneficial to reading success. Nevertheless, there has been pushback against such policies, the use of the term science of reading by “phonics-centric people”, and their lack of instructional knowledge and experience. In this article, although the author supports pedagogical decision making on the basis of a confluence of evidence from a variety of sources, he cautions against instructional over- generalizations based on various kinds of basic research without an adequate consideration of instructional experiments. The author provides several examples of the premature translation of basic research findings into wide-scale pedagogical application.

Science of reading is a term that has been used for more than 200 years. Throughout this history, it has been used most frequently to refer to the pronunciation and decoding of words on the basis of basic research. In this article, I situate the term historically and recom- mend caution in formulating agendas of instructional practice or policy primarily or solely on the basis of basic research. As such, I make claims about the nature of the empirical evidence that should comprise any sci- ence of reading instruction.

The purpose of research in any field of study, including reading edu- cation, is to increase our knowledge comprehensively about reading and its development. Yet, because reading is widely accepted as a public good, reading research also has a more specific responsibility. It should help determine how best to increase the literacy franchise, to raise levels of literacy performance, to ensure equal access, and to make the delivery of reading instruction more certain and more efficient. That first goal, the comprehensive expansion of knowledge, is not my focus here, but the second one, figuring out how best to teach reading and to use these results to inform instructional practice and public policy, is what I will explore.

In current use of the term science of reading, authors often try to make pedagogical and policy claims mainly on the basis of basic research done in the cognitive sciences and neurosciences, particularly with regard to beginning reading (e.g., Seidenberg, 2017). As such, the term is a bit of a misnomer because those using it today tend to reason directly

Timothy Shanahan University of Illinois at Chicago, USA

What Constitutes a Science of Reading Instruction?

S236 | Reading Research Quarterly, 55(S1)

from basic research to the prescription of instruction; the conversation seems to be less about a science of reading than a science of reading instruction. Unfortunately, this approach ignores why we use research in this endeavor at all; we seek to reduce uncertainty in the solving of practi- cal problems. We, as a society, are uncertain about how best to educate our children. We know that large percent- ages of students leave school with literacy levels too low to allow them full participation in the benefits of society (Shanahan & Shanahan, 2008). We know that socioeco- nomic status, racial identity, and language heritage too often determine how successful students will be in gain- ing reading ability in our schools (National Center for Education Statistics, 2013). In other words, how well reading is taught in schools is important, as it can have long-lasting impacts on the well-being of individuals. The purpose of bringing research to this enterprise is to try to increase the likelihood of success; we want to promote practices and policies that will have the greatest possibil- ity of ensuring equity and excellence in reading. We seek to reduce our uncertainty and to decrease the size and consequences of the inferences we must make to make that work.

For example, let’s say that a study reveals that readers are more likely to remember the information higher in the information hierarchy of a text than the information lower in the structure (Meyer, 1975). Such research provides use- ful information about how our minds work, how we gain information, and how we learn. It suggests that if we want to create better readers, then it would be wise to promote a more thorough or effective processing of subordinate infor- mation. Perhaps knowing that individuals are less likely to process subordinate information reduces our uncertainty as to how to proceed, but past experience suggests that it does not do so sufficiently to justify major pedagogical or policy investments. Sometimes basic research, including basic research in education that depends on descriptive, correlational, or qualitative data, identifies phenomena that may not be addressable pedagogically.

If we want to know how best to promote this better reading, then we need studies aimed at determining whether we can teach such abilities. That means developing cur- ricula and instructional regimes and then eva luating their effectiveness. With regard to improving readers’  recall of subordinate information, that has been researched, and indeed, this aspect of behavior can be improved, removing a significant inference from the causal chain (A. Adams, Carnine, & Gersten, 1982), and this success should increase our certainty with regard to future possibilities of success. If this instructional study were then replicated repeatedly with diverse groups of students, that should increase our certainty even more in the value of this approach to reading improvement (Hebert, Bohaty, Nelson, & Brown, 2016; Pyle et al., 2017; Wijekumar, Meyer, & Lei, 2017). Also, as will be explained, even that level of certainty may be

inadequate, both practically and ethically, for providing sufficient and responsible guidance for education policy and teaching.

Although we have much to learn from basic studies of reading processes and neurological functions, it is impor- tant not to overrely on such evidence in determining how best to teach reading. High-quality research reduces uncer- tainty, and a confluence of high-quality empirical data from multiple sources should go far in increasing our confidence that certain policies and practices will be effective and ben- eficial. In this article, I argue not for ignoring the fruits of basic research but for adjudicating matters of pedagogy chiefly on the basis of experimental studies of instructional practice itself. If our goal is to determine how best to teach reading, then we must rely on data that evaluate the effec- tiveness of teaching, rather than depending solely or even mainly on studies of reading processes or of other nonin- structional phenomena, which are then applied to teaching through analogy or logical deduction or from premature conclusions drawn from empirical investigations that do no more than describe or correlate. The role of basic research in shaping instruction quite appropriately lies either in identifying pedagogical innovations that can be evaluated through studies of instruction or in providing evidence that further buttresses or explains the results of such experi- mental pedagogical study.

Basic research by its very nature cannot reduce our uncertainty sufficiently to justify its use to determine public policy any more than the genetic mapping of COVID-19 should be used to promote a particular regi- men of wide-scale inoculation. In that sense, it is neces- sary but not sufficient. Basic research may point in helpful directions or warn us off from likely failures, but to be really certain of the effectiveness and safety of a vaccine, it is necessary to try it out on scale with varied populations. Identifying agents that might work effectively against a virus is one thing, and figuring out how such agents can actually be used effectively is something else altogether. Both of these kinds of research are essential in both medi- cine and education, but the key point here is that practice and policy, ultimately, must depend on evidence showing the practical effectiveness and safety of the approaches taken with patients or students, not on the basic research that may have generated the original insights that led to the development of those practices that worked.

The impact of the term science of reading as used today is as much bound up in its tone as in its meaning; it now often seems to be used as a rhetorical cudgel to challenge those not adhering to some particular conception of it, hence the arguments over who even has the right to use the term (Calkins, 2020; National Education Policy Center & Education Deans for Justice and Equity, 2020). One’s com- fort with today’s science of reading seems to depend on which instructional approaches one advocates and what one is willing to accept as determinative evidence. As such,

What Constitutes a Science of Reading Instruction? | S237

in this article, I delve into the nature of the kind of evidence that should be the basis of a science of reading instruction.

This consideration of what should be included in a science of reading instruction surfaces many philosophi- cal and practical issues, not all of which can be addressed here: Which research questions are worthwhile from a “science of reading instruction” perspective? Should we promote basic or applied science? Should pedagogy be governed by standards of instruction or professional autonomy? What are the nature and qualities of research most likely to contribute to a science of reading (e.g., types of studies, methodological rigor, criteria for amounts and types of evidence)?

What Is the Science of Reading? Science of reading is an old term. I conducted a Google Books Ngram analysis of its use in a corpus of 8 million books, limited to those written in English. This collection of books, fiction and nonfiction, represents 6% of all books ever published, drawn from more than 40 univer- sity libraries. Ngram searched these books for all appear- ances of the three-word string “science of reading.” Problems with Ngram analyses have been identified (e.g., scanning problems, overinclusion of scientific texts), but those should not be an issue here because our purpose is simply to see how long the term has been in use, to gain a sense of its popularity, and to identify its more common meanings, rather than any exhaustive or nuanced exami- nation of usage.

The term was used first to refer to text reading during the 18th century, coinciding with the birth of linguistics, the scientific study of language. The original agenda of scientific linguistics was the determination of proper pronunciations of ancient languages (Allan, 2010). Thus, during the early 19th century, the term was used to refer to how one should read the Koran or the Bible, particu- larly with regard to the pronunciation of the words (e.g., Stewart, 1809).

Science of reading was first used pedagogically during the 1830s. Consistent with its original meaning in linguis- tics, it was used in education to discuss teaching students to sound out words properly (Experience, 1836; Pitman, 1843). Use of the term has waxed and waned over two centuries, with occasional torrents of use in the 1840s, 1880s, and 1920s (see Figure 1). Although that historically early reading research was not limited to issues of decod- ing (e.g., Huey, 1908/1968), the term science of reading usually has been reserved for decoding, often with an emphasis on noninstructional research (including studies of eye movements and linguistic analyses of the English spelling system).

It is worth noting that efforts to apply research to read- ing instruction have increased markedly since the 1950s and that these efforts rarely employed the term science of reading. This increased emphasis seems to be beyond argu- ment, but it can be documented in many ways: increases in amount of reading research (e.g., the 72 volumes of Annual Summary of Investigations Relating to Reading published 1925–1998: Roser & Weintraub, 2009), the appearance of specialized journals of reading research (e.g., Reading Re ­

FIGURE 1 Results of the Google Books Ngram Analysis of the Use of the Term Science of Reading From 1700 to 2008

S238 | Reading Research Quarterly, 55(S1)

search Quarterly, Journal of Reading Behavior/Journal of Literacy Research, Scientific Studies of Reading), increases in national funding for reading research (e.g., Center for the Study of Reading, National Reading Research Center), insti- tutional efforts to interpret reading research for practitio- ners and policymakers (e.g., M.J. Adams, 1991; Anderson, Hiebert, Scott, & Wilkinson, 1985; August & Shanahan, 2006; National Early Literacy Panel [NELP], 2008; National Reading Panel [NRP], 2000; RAND Reading Study Group, 2002; Snow, Burns, & Griffin, 1998), and government policies that require practical concordance with reading research: Reading Excellence Act (U.S. Department of Education, 2000), No Child Left Behind (U.S. Department of Education, 2002b), Reading First (U.S. Department of Education, 2002c), Early Reading First (U.S. Department of Education, 2002a), Striving Readers (U.S. Department of Education, 2005), Striving Readers Comprehensive Literacy Program (U.S. Department of Education, 2015b), and the Every Student Succeeds Act (U.S. Department of Education, 2015a). These efforts to increase the application of research in reading have included a heavy emphasis on decoding but usually have gone well beyond that, considering domain knowledge, vocabulary, reading comprehension, metacog- nition, oral language, and a plethora of other instructional issues (e.g., comprehension strategies, oral reading fluency, read-alouds to students, text readability).

Now, the term science of reading has returned to vogue. This resurgence has included its appearance in scholarly publications (e.g., Snowling & Hulme, 2005). However, the surge of attention to the science of reading is attributable to its use by the media and in the public debate. Analysis of the term in indexes of newspapers and other popular publications (via Nexis Uni) over a three- decade period reveals an upsurge in the use of the term since 2018. This increase is specifically due to coverage of policy initiatives of the International Dyslexia Association, the publication of Seidenberg’s (2017) book Language at the Speed of Sight: How We Read, Why So Many Can’t, and What Can Be Done About It, and an exploration of issues raised in that book by Hanford (2018), a radio jour- nalist for American Public Media.

Seidenberg (2013, 2017) did not explicitly define science of reading in his book or in an earlier article, but it is easy enough to piece together what counts and what does not. His insights about instruction are based almost entirely on conclusions drawn from basic scientific research concerning the mechanisms of skilled reading, neural circuits, computer simulations of learning, and the like. Instructional studies are barely mentioned, and when they are, it is not to support the pedagogy that he promotes. Hanford (2018) focused public attention on Seidenberg’s book and used the term much as he did to refer to a wide range of empirical studies drawn from cognitive science and neural science showing the importance of phonological processing in proficient reading. Although I agree with Seidenberg’s and Hanford’s

conclusion—the science of reading actually supports the teaching of explicit phonics—I do so not because I agree that basic research has provided the greatest certainty of that but because basic research findings have concurred with the preponderance of evidence drawn from the direct study of that instructional question.

Given the well-documented contentiousness of the reading field (Stanovich, 2000; Taylor, 1998), an embrace of a term like science of reading by some will only arouse opposition among others, as has been the case here (Calkins, 2020; National Education Policy Center & Education Deans for Justice and Equity, 2020). Those opposed to the current use of this term argue that it is used too narrowly (Hanford, 2018, mentions decoding and related terms 86 times and all other aspects of reading only once) and that the instructional practices it promotes are overemphasized and often inappropriate. The critics also take an ad hominem swing. They question the value of ideas communicated by journalists and noneducators who do not know classroom instruction or who are not scholars of reading and literacy. Unfortunately, these opponents of current uses of the term are no more likely to rely on appropriate instructional research evidence to support their pedagogical claims.

A Science of Reading Instruction Historically, basic science has been distinguished from applied science on the basis of its apparent distance from practical problems: Basic science aims at answering fun- damental questions, whereas applied science tries to solve practical ones. This distinction is about neither who does the science (neurologists, psychologists, ethnographers vs. educators) nor their motivation (even those conduct- ing the most basic forms of science hope that their insights will have some practical value). In this article, basic sci­ ence refers to any investigation that is not directly aimed at evaluating whether a particular approach to instruction works. That includes those brain science and computer simulation studies that are the basis of public debate in reading education today. They are basic because they tell us something about how the reading brain works and how we may learn to recognize words, but these studies provide no direct test of any instruction, because no instruction is provided in them. This definition of basic science includes descriptive studies that can reveal who is doing well in reading and who struggles with status quo reading instruction. It also includes various correlational studies and complex statistical modeling that exposes the relations that various cognitive abilities may have with reading, or ethnographic observations in classrooms that describe patterns of interaction that might be useful.

In 2002, the National Research Council (NRC) issued a prestigious report in book form called Scientific Research

What Constitutes a Science of Reading Instruction? | S239

in Education. The purpose of the report was both to pro- mote the use of research as the basis of educational policy and to attempt to heal the rift between quantitative and qualitative researchers. The NRC said that “the distinction between basic and applied science [in education] has out- lived its usefulness” (p. 20). That may suggest to some that the point being made here is inappropriate or outdated, which is not the case. The NRC argument was not that all research is equal when it comes to educational policy but that all research well implemented was scientific, a posi- tion I heartily endorse. “What makes research scientific is not the motive for carrying it out, but the manner in which it is carried out” (p. 20). NRC was saying that basic research was not any more scientific than applied research and that both should be accepted as being integral to the scientific endeavor. What is being argued here is not whether basic research is scientific or even useful (indeed, it is both), but whether it is proper, either intellectually or ethically, to prescribe pedagogical practice and policy on its basis alone.

No matter how scientific basic research may be, ulti- mately any science of instruction will have to depend on applied studies of teaching, that is, those studies that require smaller inferences to application. This is not a rehash of the ad hominem concerns of the critics of current sources on science of reading but a matter of practical epistemology. No matter how sound the studies of neural processing, perception, and memory, we must recognize the possibility that they, at least in some cases, could be irrelevant, inconsequential, or misleading with regard to teaching.

A famous example is the first U.S. study of reading, or more properly of word perception (Cattell, 1886). Cattell (1886) found that readers recognized words more quickly than letters, which was interpreted to mean that we read words as wholes rather than decoding them. The peda- gogical interpretation of this was that word memorization rather than decoding should be taught, and this study was cited well into the 20th century as evidence of that. What Cattell’s study revealed was accurate and reliable—people actually recognize letters within words faster than isolated letters—but the interpretation of this finding and its application to teaching was neither accurate nor reliable. Studies quite consistently have found decoding instruc- tion to be advantageous (M.J. Adams, 1990; Chall, 1996; NELP, 2008; NRP, 2000).

Another example of pedagogical conclusions from noninstructional studies can be found in analyses of oral reading errors. Goodman (1967) analyzed errors and found that they revealed the use of orthographic/phonemic, syn- tactic, and/or semantic information. A reader might, for instance, read the word automobile as car (relying on the semantic and syntactic context or cues). Goodman con- cluded that this was how readers read, figuring words out as much from meaning and context as from letters and

sounds. This is what is meant by the three-cueing systems. Never mind that later studies only found such responses when readers erred, not when reading words proficiently and that this dependence on semantic and syntactic infor- mation was prominent with poor readers, not with good ones (Stanovich, 1980).

Goodman (1967) was not the first to recommend this kind of guessing on the basis of minimum visual informa- tion, nor was he the first to do so without any instruc- tional evidence showing that it conferred a learning advantage. Earlier, Bond and Bond (1943), in their popu- lar reading textbook, made the same recommendation, in their case on the basis of Gestalt psychology. According to Gestalt psychology, perceptions arise from patterns or gestalts rather than from an analysis of constituent parts. Thus, Bond and Bond concluded that readers should not devote much energy to figuring out words letter by letter, which is a fair generalization from the basic psychological studies but one not particularly in much agreement with direct studies of reading instruction.

Again, in Goodman’s (1967) case, his empiricism was sound. Readers, when distracted or struggling, try to compensate for this failure by inferring words that might make sense in context. However, no one has shown that teaching students to compensate in this way improves reading achievement, and other basic research has weak- ened the original claim because proficient readers look at pretty much every letter during reading, and where they look is not affected by semantics or syntax (Rayner, Binder, Ashby, & Pollatsek, 2001). (Although no research has shown that learning benefits from teaching cueing systems, there likewise is no evidence showing such teaching to be harmful; Hanford, 2019.)

There has been much basic physiological and psycho- logical study into what distinguishes struggling readers from those who learn literacy more easily. These studies have led to myriad theories of how best to address the needs of struggling readers. An early research-based theory was the claim that reading disability was linked to left-handedness (Kushner, 2017). Fortunately, one is no longer likely to see educational prescriptions based on this long-standing scientific correlation. A century ago, it was widely embraced by psychologists, physicians, and educators, which led to retraining practices in which chil- dren were forced to use their right hands through binding or corporal punishment. Needless to say, these practices were never found to improve reading ability (e.g., Fagan, 1931).

On the basis of various comparisons, correlational studies, and theories of brain–behavior relations, scien- tists have categorized students into disability subtypes, with the idea that each subtype requires special instruc- tion addressing particular deficits (Johnson & Myklebust, 1967; Kephart, 1960; Kirk & Kirk, 1971). These defi- ciencies  were not specific to reading but referred to

S240 | Reading Research Quarterly, 55(S1)

limitations in underlying cognitive or neural processes, including perceptual deficits and problems with hemi- spheric differentiation (Holmes, 1965; Kavale & Forness, 1985; Werner & Strauss, 1939). For example, there were several decades of provocative research into visual perception and visuomotor processing and their roles in  learning to read (e.g., Robinson, Mozzi, Wittick, & Rosenbloom, 1960; Snyder, & Freud, 1967). Frostig’s (Frostig & Horne, 1964) theories of perceptual-motor coordination and their implications for reading instruc- tion emerged from such research. This research also led to the widespread use of perceptual measures in class- rooms and clinics, including the Bender Visual-Motor Gestalt Test, Frostig’s Developmental Test of Visual Perception, and the Metropolitan Reading Readiness Test, all aimed at trying to predict who might have trouble in learning to read or to identify what perceptual training was required. Likewise, at least for a couple of genera- tions, young students practiced picture and shape dis- crimination in kindergarten and first grade in the name of reading readiness because of the misapplication of such psychological data (Durkin, 1980).

Some of the earliest reading studies tracked eye move- ments during reading (Javal, as cited in Huey, 1908/1968). Although the original methodology for this research was somewhat medieval (and required the administration of cocaine), the basic findings continue to be confirmed in more recent technologically advanced eye motion studies (Rayner et al., 2015). These data reveal much about read- ing, but attempts to translate them to instruction have been problematic. Because studies have shown that poorer readers have less efficient eye movement, this has led to optometric training of the eyes to read better (Heath, Cook, & O’Dell, 1976; Keogh & Pelland, 1985); such interventions have come and gone since the 1930s but are still with us today (Apperson, 1940; Murphy, 2017). These days, there are even computerized systems that supposedly transform one into a faster, more efficient reader by banishing eye movements altogether, as studies have shown that eye movements require too much of a reader’s time (Spritz, n.d.).

Another line of theory and research drawn from clini- cal data and studies of perception can be found in Orton’s (1925) theories of hemispheric dominance that encouraged tracing and other multisensory rehearsals of words, with the idea that this would result in more apt memory storage in the appropriate neural hemisphere. Unfortunately, this neurological theory, although quite influential, lacked a basis as empirically accurate as the earlier examples. Orton concluded that the hallmark of dyslexia was the visual reversal of words and letters, which he attributed to inappropriate dominance of the neural hemispheres. More rigorous studies have concluded that reversals play no spe- cial role in dyslexia (Fischer, Liberman, & Shankweiler, 1978), and to this day, no convincing studies have shown

positive learning effects from tracing letters or words, yet those practices hang on.

Yet another theory of instruction to be derived from basic psychological study is the idea of learning styles. Although this idea is most associated with Carbo (e.g., 1983), who elaborated on it greatly, parts of it have had a very long empirical life (Henmon, 1912). The basic prem- ise is that individuals learn in different ways and that teachers must identify the type of learner a student is so instruction can be modified appropriately. This is not a theory focused on differences in reading ability; thus, pre- scribing phonics instruction to poor decoders or repeated reading to the disfluent would not capture its essence. No, the idea here is that students vary in their ability to learn through different modalities; there are visual and audi- tory learners, and teachers must teach to the strongest modality. This is another interesting theory that has not panned out (Rogowsky, Calhoun, & Tallal, 2020).

There are many other examples of how basic research, even when empirically accurate, may mislead instruction. Educators, for example, used to delay reading instruction rather than provide early interventions (Morphett & Washburne, 1931). Psychological studies led to the discour- agement of oral reading instruction in U.S. classrooms for approximately 50 years (Singer, 1981), although research has found learning benefits from some oral reading approaches (NRP, 2000). Laboratory studies focused on the teaching of artificial characters have found no benefit to let- ter name teaching, which led to the recommendation that we should not teach the ABCs to beginning readers (Samuels, 1970), yet classroom studies with the actual alphabet have revealed learning benefits (Byrne, Fielding- Barnsley, & Ashley, 2000). There are more examples of this kind of overgeneralization, misinterpretation, or premature application of basic research, but these should be sufficient to make my point.

The examples provided here should not be misun- derstood. Their point was not to expose basic research as having no value, nor to suggest that it is usually wrong. Basic research can provide insightful hypotheses that may eventually lead to the development of valuable peda- gogical developments and can help us better understand why a pedagogical approach may be effective. The pur- pose of these examples is to illustrate the dangers of attempting to move directly from basic research findings to the formulation of public policy or to the widespread adoption of particular instructional practices without direct, rigorous, and repeated evaluations of the ability of those insights to improve instructional practice. In each of these examples, later studies either called into question the accuracy of the original empiricism or convincingly exposed the inadequacies of the pedagogical inferences drawn from sound data (Berninger, Lester, Sohlberg, & Mateer, 1991; Luchow & Shepherd, 1981; Pryzwansky, 1972; Williams, 1969).

What Constitutes a Science of Reading Instruction? | S241

The cognitive and neuropsychological studies that Seidenberg (2017) examined on how readers read words are an important part of a science of reading instruction but not just or even mainly because they are high-quality studies. These studies are valuable because they have contributed to our understanding of reading instruction through their consistency with the findings of extensive instructional experiments that have demonstrated consistently and over- whelmingly that explicit and systematic teaching of decod- ing is beneficial (NELP, 2008; NRP, 2000). Without those instructional studies, Seidenberg’s results could offer no more than provocative hypotheses that would still need to be evaluated through instructional studies. With those instructional studies, Seidenberg’s results provide insightful explanations of why systematic phonics instruction is so advantageous—a real contribution but not the one being touted publicly in the debates over the science of reading.

Making predictions about what kinds of instruction will be effective on the basis of basic research is a fraught enterprise. When the predictions are incorrect, they encourage poor pedagogy. When they are sound, their value can only be determined by their consistency with the findings of instructional studies. As such, the predic- tions reinforce what we learn from instructional studies, strengthening our trust in those pedagogical findings through their consonance with the predictions. Again, this does not denigrate the value of basic research for identifying potential pedagogical innovations or insight- ful explanations that could lead to even greater future innovation. Yet, no matter how good the ideas of basic research, they must be tried out instructionally and shown to be beneficial in improving reading ability or its dispersion in some way before they should be recom- mended to educators and policymakers.

A functional magnetic resonance imaging study has shown that phonics instruction alters particular areas of the brain where reading is known to take place (Temple et al., 2003), but remember that this kind of study cannot tell us how to teach reading. This particular study could not possibly do so because it did not provide an evalua- tion of any kind of teaching. The instruction in this study had already been shown to be effective; the students who received it outperformed those who did not, which was why their brains were observed: The instructional group had already shown that their brains were changed (learn- ing had happened), and the study was aimed at revealing why that instruction may have been so successful. This study, and others like it, may enlighten us as to why pho- nics may be more effective than word memorization by divulging the neural correlates of the learning that we already knew had taken place. Maybe it will encourage future improvements to instruction, but for now, the main reason for teaching phonics explicitly, thoroughly, and well in the primary grades is because in approxi- mately 100 instructional experiments (NELP, 2008; NRP,

2000), such teaching has consistently given students a clear learning advantage.

Of course, cognitive science has come up with more than the centrality of decoding to successful reading. Cognitive science has implicated other processes in read- ing as well. For instance, there is an extensive body of evi- dence suggesting the importance of theory of mind in reading comprehension (e.g., Y.-S.G. Kim, 2020). This concept refers to the ability to understand the mental states or perspectives of others. The readers’ theory of mind is what allows them to draw inferences about intents, goals, and emotions, and it has been shown to be correlated with comprehension abilities (both listening and reading). Then, there is also rapid automatic naming, a measure of the speed at which individuals can process cognitive infor- mation; studies have consistently shown this to be one of the most robust correlates of reading ability (Araújo, Reis, Petersson, & Faísca, 2015). Also, a great deal of research has demonstrated the importance of working memory and other executive functions in reading (Castles, Rastle, & Nation, 2018; Ober, Brooks, Homer, & Rindskopf, 2020). Working memory and its limitations clearly play a crucial role in reading comprehension.

As much research as there has been on each of these aspects of reading, and as persuasive as that research’s findings may be because of the strength and consistency of the correlations, the role these variables play in complex statistical models, and the elegant theories that connect them to reading, those variables still cannot be a part of a science of reading instruction. This distinction is made because no one has yet found that they can successfully teach these variables in ways that improve reading. Unlike with decoding, for which there is a rich and extensive collection of such studies, these insights from cognitive science are, at this point, disconnected from the instruc- tional enterprise. They are provocative but not proven, and as such should not be recommended to teachers.

Let’s consider a counterexample: the earlier noted research on text structure. There is some basic science underlying this work but not a great deal (Meyer, 1975). None of the publicly prominent “science of reading” advo- cates even mention the value of teaching students about discourse structure. Nevertheless, there is an extensive and rigorous body of instructional research showing that the teaching of text structure promotes higher levels of reading comprehension with a diverse group of students under a wide array of circumstances (Hebert et al., 2016; Pyle et al., 2017; Wijekumar et al., 2017). It might be worthwhile for cognitive psychologists and neuroscientists to try to pro- vide deeper understanding of how we process text structure in our minds and brains, but until they do, based on the preponderance of available instructional evidence, it seems obvious that teaching text structure should be encouraged, even given the dearth of cognitive and neural evidence but- tressing these instructional studies. By contrast, we should

S242 | Reading Research Quarterly, 55(S1)

not teach theory of mind, rapid naming, or working mem- ory despite their strong cognitive and neural foundations, because there is not yet evidence that we know how to teach them effectively. Teachers benefit from knowing both what works and why it works, but it is the what, ultimately, that leads to improved reading ability, at least most immediately. Education is necessarily an applied science, not a basic one.

As noted earlier, some oppose this renewed call for a reliance on a science of reading, but the basis of their opposition seems to be mainly to the specific instruc- tional practices being promoted rather than to the nature of the research that is the basis of this promotion. Their concern is not with the epistemological premise of draw- ing instructional conclusions from basic research (they do this, too, just focusing on different assemblages of other nonconclusive evidence). From this, it appears that if this same cognitive science and neuroscientific data were used to support their own cherished, although often untested, instructional theories, they would gladly accept this basic scientific support (even if journalists were the ones making this information public).

In the end, the only way to know if any instructional approach is effective is to try it out in classrooms and to measure its impact on student learning, the same approach adopted by various research review panels (August & Shanahan, 2006; NELP, 2008; NRP, 2000), research insti- tutions (e.g., Institute of Education Sciences, National Institute of Child Health and Human Development, What Works Clearinghouse), and independent scholars (e.g., Castles et al., 2018; Graham & Hebert, 2011). Of course, the better designed such educational experiments are and the more often they are independently replicated, the more likely that the instructional approach under study can be made to work in other classrooms.

As some of the critics of the science of reading have pointed out, a great deal of relevant research has gone well beyond phonics instruction (Calkins, 2020). Instructional studies have identified the importance of explicit teaching of phonological awareness (NELP, 2008; NRP, 2000), oral read- ing fluency (Kuhn & Stahl, 2003; NRP, 2000), reading com- prehension strategies (NRP, 2000; Shanahan et al., 2010), text structure (Hebert et al., 2016; Pyle et al., 2017), vocabu- lary (NRP, 2000; Stahl & Fairbanks, 1986), the use of com- plex text (Shanahan, 2019), the impact of writing on reading (Graham & Hebert, 2011), and several other curricular and instructional approaches and interventions (e.g., Hattie, 2009). Consequently, those too must be considered part of any acceptable definition of a science of reading instruction.

The Pieces That Do Not Fit Yet, what of those instructional practices, supported by pedagogical research but inconsistent with basic research findings? Are these practices like the fabled hummingbird

that supposedly, according to aeronautical research, could not fly? The idea is to use discrepancies between basic research and successful instructional approaches to raise questions about how to make an approach more successful.

An example of this is Reading Recovery, a program of beginning reading instruction aimed at struggling read- ers. This instructional scheme has often been challenged by critics (e.g., Baker et al., 2002; Greaney, 2001, 2011; Wood, 1994) unhappy because of the inconsistency of that program with what is known about effective decod- ing instruction. Despite this inconsistency, qualitative syntheses (e.g., Shanahan & Barr, 1995), meta-analyses (e.g., D’Agostino & Harmey, 2016), and specific high- quality studies (What Works Clearinghouse, 2008) have all concluded that Reading Recovery improves reading. The research on Reading Recovery has often been flawed by seriously misleading research practices, and these limi- tations have been well documented (Chapman & Tunmer, 1991, 2016; Shanahan, 1987; Shanahan & Barr, 1995; What Works Clearinghouse, 2008). However, even when those questionable studies are dismissed, Reading Recov- ery has been found to provide learning benefits, albeit at what to many is a prohibitively high cost.

Reading Recovery appears to be a hummingbird! Although we know the program works, we cannot be cer- tain about why it does. Often, instructional approaches are complex, including a variety of content and instruc- tional activities. In this, Reading Recovery is not an exception. In its daily 30-minute lessons, students read little books according to their reading levels, engage in the rereading of books they have read previously, receive feedback on their oral reading errors based on the three- cueing systems, write sentences, and engage in other learning activities under the one-to-one supervision of a highly trained teacher. Unfortunately, studies have indi- cated that the entire constellation of content and activity can be advantageous, but have not identified which fea- tures of the program are active ingredients. Is it possible, for instance, that any of those components could be omit- ted without lessening the positive results?

This kind of component analysis has not been common in education, but when it has been used, such as with vocab- ulary instruction, it has been very informative (McKeown, Beck, Omanson, & Pople, 1985) both in helping to tailor better instruction and, as with basic research, by providing clues to what makes this instruction work. Although we lack component analyses of Reading Recovery, some rele- vant investigations have provided clues. For example, the one-to-one teaching approach may neither explain the pro- gram’s success nor be necessary for its successful implemen- tation (Iversen, Tunmer, & Chapman, 2005). Investigations have shown that Reading Recovery has no impact on pho- nological awareness (Chapman, Tunmer, & Prochnow, 2001), and in a study in which Reading Recovery was

What Constitutes a Science of Reading Instruction? | S243

supplemented with explicit phonics teaching, learning ac - celerated significantly (Iversen & Tunmer, 1993).

It seems clear from the neurosciences (D’Mello & Gabrieli, 2018) that in terms of brain function, we all read in the same way, no matter how we were taught. It is not clear, however, what readers learn that enables this uni- versal process. We know about the coordination of pho- nological processing and visual processing, and we know that teaching a broad array of sound–symbol relations and spelling patterns enhances reading achievement, but we do not know what is coded into memory. Do nascent readers learn rules, patterns, or the words themselves? Rules refers to a system of constraints that limit structural possibilities, such as i after e except after c. Some psychol- ogists think that reading requires the mastery of such rules (Pinker, 1999). Another possibility is that we learn to recognize patterns (Venezky, 1995). Patterns here refers to the allowable sequences of letters that relate spelling to sound. According to this idea, learners come to recognize legal sequences of letters (i.e., those that occur with regu- larity). Finally, there is the possibility that we learn words more directly, as templates perhaps, which enables us to recognize similar shapes and sequences in the future (Pinker, 1984), making reading a remarkable memory task.

However, what is going on when instruction does so little to explicitly familiarize students with the relations between orthographic patterns and phonology, such as in Reading Recovery? Somehow, students who are being taught in this way are still ending up reading much as the kids who receive explicit decoding instruction. The same could be said of approaches to reading that only teach words (Barr, 1974). As already noted, such approaches do not do as well as explicit decoding instruction in improv- ing reading, yet how do students learn from them at all? According to basic research studies, they should not work; that they do should be a matter of more than intel- lectual curiosity.

What physicists and engineers knew about aerody- namics was not consistent with the flight behaviors they could observe in hummingbirds (Ransford, 2008). That led them to a great deal of study of hummingbirds, expanding what we now know about flying and hovering. Instead of assuming that the basic knowledge of aerody- namics was complete and correct, researchers decided that it was worth probing those instances where practice was not in accord with empirically grounded theory.

Another example of discrepancies between successful instruction and basic science concerns the role of text complexity in reading instruction. Teachers have long been told to teach students to read at their so-called instructional levels and that, by doing so, students would make optimum progress (Betts, 1946). Accordingly, a large number of instructional programs that successfully teach many students to read use this approach (e.g.,

Ransford-Kaldon et al., 2010). Yet, there has been a great accumulation of basic cognitive data showing that under at least some circumstances, engaging students in more complex texts leads to greater learning (McDaniel & Einstein, 2005). Instructional studies are starting to bear this out as well (Shanahan, 2020).

These examples demonstrate the value that basic research in cognitive science and the neurosciences brings to reading. Instructional experiments may indicate that an instructional approach is beneficial. When such results appear to contradict basic cognitive research, it is worth- while to pursue the discrepancy empirically, as this may lead to further refinements of successful practices or to the development of more effective alternatives. Those new ideas themselves, then, must eventually be evaluated through the use of teaching experiments.

As the NRC (2002) concluded, a guiding principle that should underlie all scientific investigation is the need to employ research methods that permit direct investiga- tion of a question: “[Research] methods can only be judged in terms of their appropriateness and effectiveness in addressing a particular research question” (p. 3). In other words, if we are trying to find out the most effective way to teach reading or the educational practices most likely to provide the fairest distribution of reading ability, then we need to directly investigate those questions with appropriate methodology. Given the nature of the kinds of practical questions of reading pedagogy under discus- sion here (e.g., what is it that schools or teachers can do that will result in…?), direct investigation would require that research rigorously evaluate the effectiveness of the recommended pedagogy.

The NRC (2002) went a step further, too:

Particular research designs and methods are suited for specific kinds of investigations and questions, but can rarely illuminate all the questions and issues in a line of inquiry. Therefore, very different methodological approaches must often be used in various parts of a series of related studies. (p. 4)

The NRC’s immediate purpose was to highlight the value of investigations that combine quantitative and qualitative methods. Nevertheless, this point also sup- ports the kinds of combinations of basic and applied research promoted here. Basic research is not appropri- ate for answering the practical pedagogical questions raised by public policy exigencies (despite recent “sci- ence of reading” claims), but basic research will continue to play an important role when used in combination with applied research.

Issues of Research Quality If such teaching experiments are necessarily the episte- mological heart of the science of reading (because they

S244 | Reading Research Quarterly, 55(S1)

answer the questions most directly and with methodol- ogy appropriate to the questions), then all of the quality criteria for conducting and reporting such studies must be adhered to. Teachers and groups of students in the var- ious conditions have to be truly comparable (through randomization or relevant pretesting), and the conditions being compared must be equivalent, too (e.g., all groups receiving equal amounts of teaching, fidelity checks). The measures used to evaluate learning must be reliable and valid, and replication matters, too (J.S. Kim, 2019). If the findings of multiple independent studies do not concur, then there can be no certainty. These multiple studies should be meta-analyzed properly and consistently. Finally, the insights drawn from implementation science must be honored (Kelly & Perkins, 2012; Wasik & Snell, 2019); there is a great distance between cognitive studies produced in a laboratory and instructional studies imple- mented in classrooms, but there is an equally wide chasm between instructional effectiveness studies and successful large-scale implementations.

Clearly, the idea that certain data will be more apt for answering particular kinds of questions challenges those who claim that it is better to treat all research paradigms as equal. Placing instructional experiments above basic cognitive research, or above instructional studies that are descriptive, qualitative, or correlational, may seem to some to be unfair or antidemocratic (Cunningham, 2001; Pearson, 2004). The issue, however, is not one of fairness but rather of what kinds of questions can be answered sat- isfactorily by particular empirical methods (NRC, 2002; NRP, 2000). If one claims that an approach to teaching confers learning advantages, that is a causal claim. The trustworthiness of a causal claim about instruction will differ depending on whether the empirical evidence was drawn directly from a systematic evaluation of its use or from the measurement of some distant, underlying phe- nomena that then must be linked back to instruction logi- cally rather than empirically (as is often done with basic research). Likewise, our ability to depend on an instruc- tional claim should be enhanced when the approach has been applied and found to improve learning. Certainly, such evidence should be more persuasive than qualitative or quantitative correlations or descriptions. At best, these can suggest the possibility that something may be effec- tive. Correlations may point us in the right direction, but given the importance of literacy and the inequality of its distribution in society, it would be best to adopt practices with a high certainty of effectiveness above those that provide no more than provocative possibilities.

Current proponents of the science of reading are cor- rect that there is a substantial body of high-quality cogni- tive and neuroscientific evidence, and it is evident that instruction consistent with that research has not been emphasized much in teacher education recently (Cohen, Mather, Schneider, & White, 2017; Joshi et al., 2009). Yet,

these arguments have characterized this problem too nar- rowly, ignoring most issues of reading instruction beyond decoding and beginning reading. These arguments have underestimated the challenges inherent in applying any research findings on scale and may have overestimated the likely payoffs from such applications. Consider the National Assessment of Educational Progress gains ob - tained during the past phonics awakenings in the 1990s and early 2000s, which were positive for sure yet modest considering the current rhetoric (National Center for Education Statistics, 2013).

Basic research has an important role to play in reading science but can never be the final determinant of practice or policy; that should always depend on studies that directly evaluate the effectiveness of a practice or policy. Reports such as those by the NRP (2000) and the NELP (2008) are the most promising foundation for practice and policy determinations because the panels directly evaluated the effectiveness of instruction. Likewise, the research agendas of the Institute of Education Sciences and the  National Institute of Child Health and Human Development are promising both because they encourage high-quality basic research studies and because they require that these lines of investigation eventually result in experimental evaluations of practical applications. Such programs of research should allow us to take pedagogical action with greater certainty and with a higher possibility of success.

REFERENCES Adams, A., Carnine, D., & Gersten, R. (1982). Instructional strategies

for studying content area texts in the intermediate grades. Reading Research Quarterly, 18(1), 27–55. https://doi.org/10.2307/747537

Adams, M.J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT Press.

Allan, K. (2010). The Western classical tradition in linguistics (2nd ed.). Sheffield, UK: Equinox.

Anderson, R.C., Hiebert, E.H., Scott, J.A., & Wilkinson, I.A.G. (1985). Becoming a nation of readers: The report of the Commission on Read­ ing. Washington, DC: National Institute of Education, U.S. Depart- ment of Education.

Apperson, S.V. (1940). The effectiveness of orthoptic training as a means of remedial instruction of reading. The Journal of Experimen­ tal Education, 9(2), 160–166. https://doi.org/10.1080/00220 973.1940. 11010201

Araújo, S., Reis, A., Petersson, K.M., & Faísca, L. (2015). Rapid automa- tized naming and reading performance: A meta-analysis. Journal of Educational Psychology, 107(3), 868–883. https://doi.org/10.1037/ edu00 00006

August, D., & Shanahan, T. (Eds.). (2006). Developing literacy in second­ language learners: Report of the National Literacy Panel on Language­ Minority Children and Youth. Mahwah, NJ: Erlbaum.

Baker, S., Berninger, V.W., Bruck, M., Chapman, J., Eden, G., Elbaum, B., … Wolf, M. (2002). Experts say Reading Recovery is not effective, leaves too many children behind: An open letter from reading researchers. Retrieved from https://www.wrigh tslaw.com/info/read. rr.ltr.exper ts.htm

Barr, R. (1974). The effect of instruction on pupil reading strategies. Reading Research Quarterly, 10(4), 555–582.

What Constitutes a Science of Reading Instruction? | S245

Berninger, V.W., Lester, K., Sohlberg, M.M., & Mateer, C. (1991). Inter- ventions based on the multiple connections model of reading for developmental dyslexia and acquired deep dyslexia. Archives of Clin­ ical Neuropscyhology, 6(4), 375–391.

Betts, E. (1946). Foundations of reading instruction: With emphasis on differentiated guidance. Chicago, IL: American Book.

Bond, G.L., & Bond, E. (1943). Teaching the child to read. New York, NY: Macmillan.

Byrne, B., Fielding-Barnsley, R., & Ashley, L. (2000). Effects of pre- school phoneme identity training after six years: Outcome level dis- tinguished from rate of response. Journal of Educational Psychology, 92(4), 659–667. https://doi.org/10.1037/0022-0663.92.4.659

Calkins, L. (2020). No one gets to own the term “the science of reading”. New York, NY: Teachers College Reading and Writing Project, Columbia University. Retrieved from https://readi ngand writi ngpro ject.org/news/no-one-gets-to-own-the-term-the-scien ce-of-reading

Carbo, M. (1983). Research in reading and learning style: Implications for exceptional children. Exceptional Children, 49(6), 486–494. https://doi.org/10.1177/00144 02983 04900601

Castles, A., Rastle, K., & Nation, K. (2018). Ending the reading wars: Reading acquisition from novice to expert. Psychological Science in the Public Interest, 19(1), 5–51. https://doi.org/10.1177/15291 00618 772271

Cattell, J.M. (1886). The time taken up by cerebral operations. Mind, 11(44), 524–538. https://doi.org/10.1093/mind/os-XI.44.524

Chall, J.S. (1996). Learning to read: The great debate (3rd ed.). Fort Worth, TX: Harcourt Brace.

Chapman, J.W., & Tunmer, W.E. (1991). Recovering Reading Recovery. Australia and New Zealand Journal of Developmental Disabilities, 17(1), 59–71. https://doi.org/10.1080/07263 86910 0034271

Chapman, J.W., & Tunmer, W.E. (2016). Is Reading Recovery an effec- tive intervention for students with reading difficulties? A critique of the i3 scale-up study. Reading Psychology, 37(7), 1025–1042. https:// doi.org/10.1080/02702 711.2016.1157538

Chapman, J.W., Tunmer, W.E., & Prochnow, J.E. (2001). Does success in the Reading Recovery program depend on developing proficiency in phonological-processing skills? A longitudinal study in a whole language instructional context. Scientific Studies of Reading, 5(2), 141–176. https://doi.org/10.1207/S1532 799Xs sr0502_2

Cohen, R.A., Mather, N., Schneider, D.A., & White, J.M. (2017). A comparison of schools: Teacher knowledge of explicit code-based reading instruction. Reading and Writing, 30(4), 653–690. https:// doi.org/10.1007/s11145-016-9694-0

Cunningham, J.W. (2001). The National Reading Panel report [Review]. Reading Research Quarterly, 36(3), 326–335. https://doi.org/10.1598/ RRQ.36.3.5

D’Agostino, J.V., & Harmey, S.J. (2016). An international meta-analysis of Reading Recovery. Journal of Education for Students Placed at Risk, 21(1), 29–46. https://doi.org/10.1080/10824 669.2015.1112746

D’Mello, A.M., & Gabrieli, J.D.E. (2018). Cognitive neuroscience of dyslexia. Language, Speech, and Hearing Services in Schools, 49(4), 798–809. https://doi.org/10.1044/2018_LSHSS-DYSLC-18-0020

Durkin, D. (1980). Teaching young children to read. Boston, MA: Allyn & Bacon.

Experience. (1836). On teaching reading. In W.C. Woodbridge (Ed.), American annals of education and instruction, for the year 1836 (Vol. 6, pp. 65–69). Boston, MA: John Allen.

Fagan, L.B. (1931). A case study of dextral training of a left-handed boy and its effect on speech, reading and writing. The Psychological Clinic, 19, 291–293.

Fischer, F.W., Liberman, I.Y., & Shankweiler, D. (1978). Reading rever- sals and developmental dyslexia: A further study. Cortex, 14(4), 496– 510. https://doi.org/10.1016/S0010-9452(78)80025-2

Frostig, M., & Horne, D. (1964). The Frostig program for the develop­ ment of visual perception: Teacher’s guide. Chicago, IL: Follett.

Goodman, K.S. (1967). Reading: A psycholinguistic guessing game. Journal of the Reading Specialist, 6(4), 126–135. https://doi.org/10. 1080/19388 07670 9556976

Graham, S., & Hebert, M. (2011). Writing to read: A meta-analysis of the impact of writing and writing instruction on reading. Harvard Educational Review, 81(4), 710–744. https://doi.org/10.17763/ haer.81.4.t2k0m 13756 113566

Greaney, K. (2001). An investigation of teacher preferences for word identification strategies. Australian Journal of Language and Liter­ acy, 24(1), 21–30.

Greaney, K.T. (2011). The multiple cues or “searchlights” word reading theory: Implications for Reading Recovery®. Perspectives on Lan­ guage and Literacy, 37(4), 15–19.

Hanford, E. (2018, September 10). Hard words: Why aren’t kids being taught to read? APM Reports. Retrieved from https://www.apmre ports.org/episo de/2018/09/10/hard-words-why-ameri can-kids-arent- being-taught-to-read

Hanford, E. (2019, August 22). At a loss for words: How a flawed idea is teaching millions of kids to be poor readers. APM Reports. Retrieved from https://www.apmre ports.org/episo de/2019/08/22/whats-wrong- how-schoo ls-teach-reading

Hattie, J. (2009). Visible learning: A synthesis of over 800 meta­analyses relating to achievement. New York, NY: Routledge.

Heath, E.J., Cook, P., & O’Dell, N. (1976). Eye exercises and reading efficiency. Academic Therapy, 11(4), 435–445. https://doi.org/10. 1177/10534 51276 01100408

Hebert, M., Bohaty, J.J., Nelson, J.R., & Brown, J. (2016). The effects of text structure instruction on expository text comprehension: A meta-analysis. Journal of Educational Psychology, 108(5), 609–629. https://doi.org/10.1037/edu00 00082

Henmon, V.A.C. (1912). The relation between mode of presentation and retention. Psychological Review, 19(2), 79–96. https://doi.org/10. 1037/h0072813

Holmes, J.A. (1965). Basic assumptions underlying the substrata- factor theory. Reading Research Quarterly, 1(1), 5–28. https://doi.org/ 10.2307/746971

Huey, E.B. (1968). The psychology and pedagogy of reading. Cambridge, MA: MIT Press. (Original work published 1908)

Iversen, S., & Tunmer, W.E. (1993). Phonological processing skill and the Reading Recovery Program. Journal of Educational Psychology, 85(1), 112–126. https://doi.org/10.1037/0022-0663.85.1.112

Iversen, S., Tunmer, W.E., & Chapman, J.W. (2005). The effects of vary- ing group size on the Reading Recovery approach to preventive early intervention. Journal of Learning Disabilities, 38(5), 456–472. https:// doi.org/10.1177/00222 19405 03800 50801

Johnson, D., & Myklebust, H.R. (1967). Learning disabilities: Educa­ tional principles and practices. New York, NY: Grune & Stratton.

Joshi, R.M., Binks, E., Graham, L., Ocker-Dean, E., Smith, D.L., & Boulware-Gooden, R. (2009). Do textbooks used in university read- ing education courses conform to the instructional recommenda- tions of the National Reading Panel? Journal of Learning Disabilities, 42(5), 458–463. https://doi.org/10.1177/00222 19409 338739

Kavale, K.A., & Forness, S.R. (1985). Learning disability and the history of science: Paradigm or paradox? Remedial and Special Education, 6(4), 12–24. https://doi.org/10.1177/07419 32585 00600404

Kelly, B., & Perkins, D.F. (Eds.). (2012). Handbook of implementation science for psychology in education. New York, NY: Cambridge Uni- versity Press.

Keogh, B.K., & Pelland, M. (1985). Vision training revisited. Journal of Learning Disabilities, 18(4), 228–236. https://doi.org/10.1177/00222 19485 01800410

Kephart, N.C. (1960). The slow learner in the classroom. Columbus, OH: Merrill.

Kim, J.S. (2019). Making every study count: Learning from replication failure to improve intervention research. Educational Researcher, 48(9), 599–607. https://doi.org/10.3102/00131 89X19 891428

S246 | Reading Research Quarterly, 55(S1)

Kim, Y.-S.G. (2020). Theory of mind mediates the relations of language and domain-general cognitions to discourse comprehension. Jour­ nal of Experimental Child Psychology, 194, article 104813. https:// doi.org/10.1016/j.jecp.2020.104813

Kirk, S.A., & Kirk, W.D. (1971). Psycholinguistic learning disabilities: Diagnosis and remediation. Chicago: University of Illinois Press.

Kuhn, M.R., & Stahl, S.A. (2003). Fluency: A review of developmental and remedial practices. Journal of Educational Psychology, 95(1), 3– 21. https://doi.org/10.1037/0022-0663.95.1.3

Kushner, H.I. (2017). On the other hand: Left hand, right brain, mental disorder, and history. Baltimore, MD: Johns Hopkins University Press.

Luchow, J.P., & Shepherd, M.J. (1981). Effects of multisensory training in perceptual learning. Learning Disability Quarterly, 4(1), 38–43. https://doi.org/10.2307/1510711

McDaniel, M.A., & Einstein, G.O. (2005). Material appropriate diffi- culty: A framework for determining when difficulty is desirable for improving learning. In A.F. Healy (Ed.), Experimental cognitive psy­ chology and its applications (pp. 73–85). Washington, DC: American Psychological Association.

McKeown, M.G., Beck, I.L., Omanson, R.C., & Pople, M.T. (1985). Some effects of the nature and frequency of vocabulary instruction on the knowledge and use of words. Reading Research Quarterly, 20(5), 522–535. https://doi.org/10.2307/747940

Meyer, B.J.F. (1975). The organization of prose and its effects on mem­ ory. Amsterdam, Netherlands: North-Holland.

Morphett, M.V., & Washburne, C. (1931). When should children begin to read? The Elementary School Journal, 31(7), 496–503. https://doi. org/10.1086/456609

Murphy, R. (with Heiting, G.). (2017, April). Are learning-related vision issues holding your child back? All About Vision. Retrieved from https://www.allab outvi sion.com/paren ts/learn ing.htm

National Center for Education Statistics. (2013). The Nation’s Report Card: Trends in academic progress 2012 (NCES 2013-456). Washing- ton, DC: National Center for Education Statistics, Institute of Educa- tion Sciences, U.S. Department of Education.

National Early Literacy Panel. (2008). Developing early literacy: Report of the National Early Literacy Panel. Washington, DC: National Institute for Literacy.

National Education Policy Center & Education Deans for Justice and Equity. (2020). Policy statement on the “science of reading”. Boul- der: National Education Policy Center, University of Colorado Boulder. Retrieved from https://nepc.color ado.edu/publi catio n/ fyi-readi ng-wars

National Reading Panel. (2000). 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: National Institute of Child Health and Human Development.

National Research Council. (2002). Scientific research in education. Washington, DC: National Academy Press.

Ober, T.M., Brooks, P.J., Homer, B.D., & Rindskopf, D. (2020). Execu- tive functions and decoding in children and adolescents: A meta- analytic investigation. Educational Psychology Review. Advance online publication. https://doi.org/10.1007/s10648-020-09526-0

Orton, S.T. (1925). “Word-blindness” in school children. Archives of Neurology and Psychiatry, 14(5), 581–615. https://doi.org/10.1001/ archn eurps yc.1925.02200 17000 2001

Pearson, P.D. (2004). The reading wars. Educational Policy, 18(1), 216– 252. https://doi.org/10.1177/08959 04803 260041

Pinker, S. (1984). Visual cognition: An introduction. Cognition, 18(1– 3), 1–63. https://doi.org/10.1016/0010-0277(84)90021-0

Pinker, S. (1999). Words and rules: The ingredients of language. New York, NY: Basic.

Pitman, I. (1843). Report of the proceedings at the Birmingham Pho- nographic Festival, held in the assembly room of Dee’s Royal Hotel, July 18, 1843. The Phonotypic Journal, 2(20), 109–120.

Pryzwansky, W.B. (1972). Effects of perceptual-motor training and manuscript writing on reading readiness skills in kindergarten. Jour­ nal of Educational Psychology, 63(2), 110–115. https://doi.org/10. 1037/h0032648

Pyle, N., Vasquez, A.C., Lignugaris/Kraft, B., Gillam, S.L., Reutzel, D.R., Olszewski, A., … Pyle, D. (2017). Effects of expository text structure interventions on comprehension: A meta-analysis. Reading Research Quarterly, 52(4), 469–501. https://doi.org/10.1002/rrq.179

Ransford, M. (2008, March 4). Animals that hover. Popular Science. Retrieved from https://www.popsci.com/scite ch/artic le/2008-03/ anima ls-hover/

Ransford-Kaldon, C.R., Flynt, E.S., Ross, C.L., Franceschini, L., Zob- lotsky, T., Huang, Y., & Gallagher, B. (2010). Implementation of effec­ tive intervention: An empirical study to evaluate the efficacy of Fountas & Pinnell’s Leveled Literacy Intervention system (LLI). Memphis, TN: Center for Research in Educational Policy, University of Memphis.

Rayner, K., Abbott, M.J., Schotter, E.R., Belanger, N.N., Higgins, E.C., Leinenger, M., … Plummer, P. (2015). Keith Rayner eye movements in reading data collection. San Diego: University of California, San Diego Library Digital Collections. https://doi.org/10.6075/J0JW8BSV

Rayner, K., Binder, K.S., Ashby, J., & Pollatsek, A. (2001). Eye move- ment control in reading: Word predictability has little influence on initial landing positions in words. Vision Research, 41(7), 943–954. https://doi.org/10.1016/S0042-6989(00)00310-2

RAND Reading Study Group. (2002). Reading for understanding: Toward a R&D program in reading comprehension. Santa Monica, CA: RAND.

Robinson, H.M., Mozzi, L., Wittick, M.L., & Rosenbloom, A.A. (1960). Children’s perceptual achievement forms: A three year study. Ameri­ can Journal of Optometry and Archives of American Academy of Optometry, 37(5), 223–237. https://doi.org/10.1097/00006 324-196 00 5000-00001

Rogowsky, B.A., Calhoun, B.M., & Tallal, P. (2020). Providing instruc- tion based on students’ learning style preferences does not improve learning. Frontiers in Psychology, 11, article 164. https://doi.org/ 10.3389/fpsyg.2020.00164

Roser, N.L., & Weintraub, S. (2009). From Gray to Google: Learning within a profession. In J.V. Hoffman & Y.M. Goodman (Eds.), Changing literacy for changing times: An historical perspective on the future of reading research, public policy, and classroom practices (pp. 17–33). New York, NY: Routledge.

Samuels, S.J. (1970). An experimental program for teaching letter names of the alphabet. Washington, DC: Office of Education, U.S. Depart- ment of Health, Education, and Welfare.

Seidenberg, M.S. (2013). The science of reading and its educational implications. Language Learning and Development, 9(4), 331–360. https://doi.org/10.1080/15475 441.2013.812017

Seidenberg, M. (2017). Language at the speed of sight: How we read, why so many can’t, and what can be done about it. New York, NY: Basic.

Shanahan, T. (1987). The Early Detection of Reading Difficulties (3rd ed.), by Marie Clay [Book review]. Journal of Reading Behavior, 19(1), 117–119.

Shanahan, T. (2019). Why children should be taught to read with more challenging texts. Perspectives on Language and Literacy, 45(4), 17– 19, 22–23.

Shanahan, T. (2020). Limiting children to books they can already read. American Educator, 44(2), 13–17, 39.

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

Shanahan, T., Callison, K., Carriere, C., Duke, N.K., Pearson, P.D., Schatschneider, C., & Torgesen, J. (2010). Improving reading compre­ hension in kindergarten through 3rd grade: A practice guide (NCEE 2010-4038). Washington, DC: National Center for Education Eval ua-

What Constitutes a Science of Reading Instruction? | S247

tion  and Regional Assistance, Institute of Education Sciences, U.S. Department of Education.

Shanahan, T., & Shanahan, C. (2008). Teaching disciplinary literacy to adolescents: Rethinking content-area literacy. Harvard Educational Review, 78(1), 40–59. https://doi.org/10.17763/ haer.78.1.v6244 4321 p 602101

Singer, H. (1981). Teaching the acquisition phase of reading develop- ment: An historical perspective. In O.J.L. Tzeng & H. Singer (Eds.), Perception of print: Reading research in experimental psychology (pp. 291–311). Hillsdale, NJ: Erlbaum.

Snow, C.E., Burns, M.S., & Griffin, P. (Eds.). (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.

Snowling, M.J., & Hulme, C. (2005). The science of reading: A hand­ book. Malden, MA: Blackwell.

Snyder, R.T., & Freud, S.L. (1967). Reading readiness and its relation to maturational unreadiness as measured by the spiral aftereffect and other visual-perceptual techniques. Perceptual and Motor Skills, 25(3), 841–854. https://doi.org/10.2466/pms.1967.25.3.841

Spritz. (n.d.). Reading reimagined. Retrieved from https://spritz.com/ Stahl, S.A., & Fairbanks, M.M. (1986). The effects of vocabulary instruc-

tion: A model-based meta-analysis. Review of Educational Research, 56(1), 72–110. https://doi.org/10.3102/00346 54305 6001072

Stanovich, K.E. (1980). Toward an interactive-compensatory model of individual differences in the development of reading fluency. Read­ ing Research Quarterly, 16(1), 32–71. https://doi.org/10.2307/747348

Stanovich, K.E. (2000). Progress in understanding reading: Scientific foundations and new frontiers. New York, NY: Guilford.

Stewart, C. (1809). A descriptive catalogue of the Oriental library of the late Tippoo sultan of Mysore. Cambridge, MA: University Press.

Taylor, D. (1998). The spin doctors of science: The political campaign to change America’s mind about how children learn to read. Urbana, IL: National Council of Teachers of English.

Temple, E., Deutsch, G.K., Poldrack, R.A., Miller, S.L., Tallal, P., Mer- zenich, M.M., & Gabrieli, J.D.E. (2003). Neural deficits in children with dyslexia ameliorated by behavioral remediation: Evidence from functional MRI. Proceedings of the National Academy of Sciences of the United States of America, 100(5), 2860–2865. https://doi.org/10. 1073/pnas.00300 98100

U.S. Department of Education. (2000). Reading Excellence Act. Retrieved from https://www2.ed.gov/pubs/promi singi nitia tives/ rea.html

U.S. Department of Education. (2002a). Early Reading First. Retrieved from https://www2.ed.gov/progr ams/early readi ng/index.html

U.S. Department of Education. (2002b). No Child Left Behind: Elemen­ tary and Secondary Education Act (ESEA). Retrieved from https:// www2.ed.gov/nclb/landi ng.jhtml

U.S. Department of Education. (2002c). Reading First. Retrieved from https://www2.ed.gov/progr ams/readi ngfir st/legis lation.html

U.S. Department of Education. (2005). Striving Readers. Retrieved from https://www2.ed.gov/progr ams/striv ingre aders/ resou rces.html

U.S. Department of Education. (2015a). Every Student Succeeds Act (ESSA). Retrieved from https://www.ed.gov/ESSA

U.S. Department of Education. (2015b). Striving Readers Comprehen­ sive Literacy Program. Retrieved from https://www2.ed.gov/progr ams/striv ingre aders-liter acy/index.html

Venezky, R.L. (1995). From orthography to psychology to reading. In V.W. Berninger (Ed.), The varieties of orthographic knowledge (pp. 23–46). New York, NY: Springer.

Wasik, B.A., & Snell, E.K. (2019). Synthesis of preschool dosage: How quantity, quality, and content impact child outcomes. In A.J. Reyn- olds (Ed.), Sustaining early childhood learning gains: Program, school, and family influences (pp. 31–51). New York, NY: Cambridge University Press.

Werner, H., & Strauss, A. (1939). Types of visuomotor activity in their relation to low and high performance ages. Proceedings of the Ameri­ can Association on Mental Deficiency, 44(1), 163–168.

What Works Clearinghouse. (2008, December). Reading Recovery® [WWC intervention report]. Washington, DC: Institute of Educa- tion Sciences. Retrieved from https://readi ngrec overy.org/wp-conte nt/uploa ds/2016/12/wwc_readi ng_recov ery_report_08.pdf

Wijekumar, K.K., Meyer, B.J.F., & Lei, P. (2017). Web-based text struc- ture strategy instruction improves seventh graders’ content area reading comprehension. Journal of Educational Psychology, 109(6), 741–760. https://doi.org/10.1037/edu00 00168

Williams, J.P. (1969). Training kindergarten children to discriminate letter-like forms. American Educational Research Journal, 6(4), 501– 514. https://doi.org/10.3102/00028 31200 6004501

Wood, K. (1994). Reading Recovery: What is it? How effective is it? Edu­ cational Psychology in Practice, 10(1), 3–13. https://doi.org/10.1080/02667 3 6940 100101

Submitted April 20, 2020 Final revision received July 3, 2020

Accepted July 16, 2020

TIMOTHY SHANAHAN is a Distinguished Professor Emeritus at the University of Illinois at Chicago, USA; email shanahan@ uic.edu. His research focuses on how to improve reading achievement, disciplinary literacy, and reading–writing relations.