Does Close Reading Reject the Science of Reading?

  • 03 April, 2021

Teacher question:

I recently read an article suggesting that the research findings on reading comprehension have been modified, distorted and ignored (Dewitz & Graves, 2021, What concerned me most is that close reading and the CCSS came under heavy fire. Although the article ends with suggestions for bridging the research to practice gap, it leaves practicing teachers using the CCSS wondering whether to modify their reading comprehension instruction and use of close reading. Since you have written about close reading and the CCSS in other blogs, what are your suggestions?

Shanahan responds:

This article claims that today’s reading comprehension instruction is not in close accord with the science of reading. The commentary presents no new evidence but cites a number of past publications that have showed gaps between what we know and what we do. Nothing for me to disagree with there. I’ve written several articles and blogs about the role of strategies, value of knowledge building, benefits of vocabulary teaching, and so on. Instruction is, indeed, often far afield from the research.

This article goes on to assign blame for the discrepancy: commercial reading programs, public testing, the emphasis on close reading, and educational research itself. This critique isn’t unreasonable – though I’d apportion things a bit differently.

A real plus of the Dewitz and Graves article is that it has the research community shouldering some of the blame. I’d go even further. Researchers have produced too many investigations based on fuzzy reasoning and without adequate concern for applicability.

A particular worry is researchers’ continued failure to distinguish comprehension from learning to comprehend. No small difference that.

One of Michael Grave’s early studies is an example of what I’m talking about (Graves, Cooke, & LaBerge, 1983). That studied examined the influence pre-reading preparation (reviewing background knowledge, previewing key concepts to be presented in the text, preteaching vocabulary) has on comprehension. Students given this support comprehend a text better than do students without preparation.

This tells us nothing about how to make students better readers. Would a daily regimen of reading with lead to better independent comprehension? Possibly, but this study didn’t address that, nor did anyone else. I find it hard to castigate anyone for not scrupulously following such science.

Even when researchers have looked at comprehension instruction, the studies have almost always been so brief (6-8 weeks) it would be malpractice to base 13 years of comprehension instruction on those alone (something Dewitz and Graves properly complain about). Just because two months of a particular approach to instruction improved learning in a study, one should not conclude that several years of such teaching would continue to be either efficient or effective. Again, hard to criticize anyone for not adhering slavishly to that research.

At times I thought these authors were looking back with rose-colored glasses. Like them, I like the idea of having instruction, at times, focus on sets of related texts. However, claims that the science of reading has proven that approach to be best has no real foundation. If you doubt that go back and read the hemming and hawing over it in the Rand Report or look at the texts used in the bulk of comprehension studies (a large part of the research community apparently missed that memo). My point isn’t that there is no benefit to multi-text approaches, just that instructional research hasn’t shown that to be the best way to go – so it’s not the science of reading.

So, what about close reading?

The article says, “At the practical level, the instructional routine of close reading ignored years of research on comprehension instruction.” I don’t believe that to be the case – though I’ve pointed out the lack of research on close reading and have cautioned against investing too much time on it. As last week’s blog entry pointed out, I’m especially bothered by the misinterpretation of the close reading concept. Close reading is not a synonym for reading comprehension.

Unlike Dewitz and Graves, I don’t see close reading as an “instructional routine.” It is, rather, an approach to understanding text. The New Critics never talked about as “reading comprehension” but as a particular form of interpretation or criticism. They were encouraging a particular social response to text – an analytical one that seeks unity between what a text expresses and how it goes about expressing it. Close reading promotes attention to particular content and text features.

Dewitz and Graves cast the disagreement as being pedagogical, with research championing the role of prior knowledge and close reading advocating a laser-like focus on text. There is something to that – hence, my cautions to not overdo the attention to close reading (there are times when external information may be essential and times when it should be eschewed).

But there are interpretive benefits to treating a text as being self-contained object, separate from historical or cultural contexts. It makes great sense to teach students to try to understand a text through consideration of the words and structures in the text, as opposed to doing so with the external information that teachers may provide.

Prior knowledge plays a necessary role in comprehension (research says), and some of the rhetoric used to promote close reading made it sound like that wasn’t the case. However, prior knowledge is a two-edged sword.

Readers must both learn to use knowledge, and to suppress its application. Critical thinking requires that we not allow our assumptions, beliefs, presuppositions, or biases dominate our reading – and studies have shown that in at least some situations an instructional emphasis on prior knowledge may derail the comprehension process by allowing readers’ biases to run amok.

To me the real argument here is over conceptions of reading comprehension not instructional methods. My friend Jan Hasbrouck recently pointed out an apt quotation on this: “there is consistent interpretation across studies of what it means to actually comprehend a text and what the outcomes of comprehension should and could look like..." (Smith, Snow, Serry, & Hammond, 2020).

Reading research has been a bit all over the map when it comes to what constitutes comprehension. Studies have usually ignored the types of outcomes mandated by Common Core; comprehension research hasn’t been much interested in critical reading outcomes, for instance.

Deciding what should be our desired educational outcomes is not an empirical research enterprise – those issues are bound up in our values (implicating philosophy, social class, religion, individual experience, and so on).

The role for science is to help us figure out how best to accomplish these goals.

Thanks for your letter, and Drs. Dewitz and Graves, thank you for your provocative article.

My take-aways: I’d oppose skipping close reading. I want kids to be close readers. I admire people who are close readers and I think teachers should strive to accomplish the standards their states have established.

But take a gimlet-eyed look at what it is that you are teaching. Is it really close reading? Or is it just a traditional questioning scheme that is now labeled that way?   

I would also encourage you to make sure your instruction is consistent with what we have learned from instructional studies. Unfortunately, following the science of reading research in this area is fraught because it mainly shows possibilities without offering any research-based guidance as to dosage or how best to combine these elements.

Comprehension strategy instruction is effective – at least in rather small doses (Shanahan et al., 2010). There are several studies showing that this teaching can be successful in the context of science or social studies instruction (e.g, Duke et al., 2021; Kaldenberg, Watt, & Therrien, 2015; Swanson et al., 2014) -- increasing knowledge and improving reading proficiency simultaneously.

Likewise, there are some studies (Murphy et al., 2009) showing that guiding students in deep conversations about the ideas in text may bear fruit (none specifically focusing on close reading, per se). Having students write about text seems to be similarly beneficial, perhaps for the same reasons. [One study has competed these deep discussions with explicit strategy teaching and found them to be equally beneficial – making it unclear whether we should teach both.]

Comprehension research has been supportive of teaching vocabulary and other aspects of written language (e.g., syntax, cohesion, discourse structure). I’d certainly find a place for that.

Finally, research has been accumulating showing the positive impacts of teaching students with texts that they cannot already read well. That is something required in most state standards but ignored in most classrooms. As with strategy instruction and the like – these studies don’t provide much direction as to the most productive pedagogical schedule of text complexity.


Dewitz, P., & Graves, M.F. (2021). The science of reading: Four forces that modified, distorted, or ignored the research finding on reading comprehension. Reading Research Quarterly. doi:10.1002/rrq.389

Duke, N.K., Halvorsen, A.-L., Strachan, S.L., Kim, J., & Konstantopoulos, S. (2021). Putting PjBL to the Test: The Impact of Project-Based Learning on Second Graders’ Social Studies and Literacy Learning and Motivation in Low-SES School Settings. American Educational Research Journal, 58(1), 160–200.

Graves, M.F., Cooke, C.L., & Laberge, M. J. (1983). Effects of previewing difficult short stories on low ability junior high school students' comprehension, recall, and attitudes. Reading Research Quarterly, 18(3), 262–276.

Kaldenberg, E.R., Watt, S.J., & Therrien, W.J. (2015). Reading instruction in science for students with learning disabilities: A meta-analysis. Learning Disability Quarterly, 38(3), 160–173.

Murphy, P. K., Wilkinson, I. A. G., Soter, A. O., Hennessey, M. N., & Alexander, J. F. (2009). Examining the effects of classroom discussion on students’ comprehension of text: A meta-analysis. Journal of Educational Psychology, 101(3), 740–764.

Shanahan, T., Callison, K., Carriere, C., Duke, N. K., Pearson, P. D., Schatschneider, C., & Torgesen, J. (2010). Improving reading comprehension in kindergarten through third grade: A practice guide. Washington, DC: National Center for Education Evaluation and Regional Assistance, Institute of Education Sci­ences, U.S. Department of Education.

Smith, R., Snow, P., Serry, T. & Hammond, L. (2021.) The role of background knowledge in reading comprehension: A critical review. Reading Psychology, 42:3, 214-240, DOI: 10.1080/02702711.2021.1888348

Swanson, E., Hairrell, A., Kent, S., Ciullo, S., Wanzek, J.A., & Vaughn, S. (2014). A synthesis and meta-analysis of reading interventions using social studies content for students with learning disabilities. Journal of Learning Disabilities, 47(2), 178–195.  doi:10.1177/0022219412451131


See what others have to say about this topic.

Timothy Shanahan
Apr 11, 2021 03:59 PM

Yes... the CCSS document was stating an opinion about what they think works instructionally, and I was stating an opinion on where we are at this time with research. There is no direct research showing that doing a better job of teaching social studies or science improves reading comprehension. Some people believe that increases in science and/or social studies knowledge will improve reading achievement -- but they don't have studies showing that to be the case (they believe it, but they lack evidence).



Peter Dewitz
Apr 12, 2021 11:50 AM

I appreciate the attention that Tim Shanahan and his readers have devoted to our recent article, “The Four Forces That Modified, Distorted or Ignored the Research Finding on Reading Comprehension. I would like to reiterate a few points from the article and respond to the discussion. Our main quarrel was not with Common Core itself (OK, we did cite several just criticisms of those standards) but with the revised publishers’ criteria (Coleman & Pimentel, 2012) written two years after the standards. If the construction-integration model of Kintsch (1998) is correct and we use our knowledge base to create a situation model of the text, then it is not possible for a reader to stay “within the four corners of the text.” We all bring knowledge to what we read and the failure to develop some knowledge before reading may discriminate against readers who lack the cultural capital that others possess.

Whether close reading is an instructional routine or a “particular form of interpretation or criticism,” close reading became a routine in the lessons written in almost all core reading programs published after the CCSS and in the articles that describe and clarified close reading to many of us (see the work of Fisher & Frey in the Reading Teacher). What I found irksome was the failure to acknowledge, by Coleman and Pimentel and others, that instructional routines like questioning the author (McKeown, Beck and Blake, 2009) or reciprocal teaching (Palincsar & Brown, 1984) meet the goals of reading a text closely, and both routines demonstrate that students’ comprehension improves and transfers to new texts. The evidence for the efficacy of these two approaches is in the data and the transcripts of teachers and students’ discussions. Read a transcript from Palincsar or McKeown and you will realize that these students are engaging in a close reading. This is why Mike Graves and I recommend that all educators study the history of our field, but without rose colored glasses, and build on that knowledge.

I do agree with Shanahan that readers must learn to “use and suppress knowledge.” The case for using prior knowledge is clearly established both in theory and research. The key for the teacher is figuring out when you have told the students too much, denying them the opportunity to do the work of reading comprehension. Other researchers have demonstrated the importance of helping readers suppress knowledge especially when existing knowledge disrupts learning new ideas (see Sinatra, & Broughton, 2011).

As a final point I want to return to the work of Hodge, Gabriel and Chenelle (2020). They conducted a close reading of the Common Core, particularly all the references cited by the writers of the Common Core. They found ample research support for the problems the Common Core attempts to address. They found no research for the solutions that the Common Core proposes, including close reading. Will our current focus on close reading produce a cohort of readers who read closely? We do not know.

Thank you all for your attention to our work.

Apr 12, 2021 04:19 PM

Thank you, Peter, for your explanation. You say:

If the construction-integration model of Kintsch (1998) is correct and we use our knowledge base to create a situation model of the text, then it is not possible for a reader to stay “within the four corners of the text.”

Don't questioning the author and reciprocal teaching stay within the four corners of the text?

Peter Dewitz
Apr 12, 2021 06:36 PM


Kintsch would argue that the use of knowledge to build understanding is at some level automatic. We connect ideas within a text and connect the text to prior knowledge. Given that I suspect that the use of knowledge does happen in both questioning the author and reciprocal teaching. I am going to go back and reread some of the transcripts to see if that actually happens.

Thanks, Peter

Timothy Shanahan
Apr 12, 2021 08:58 PM


Neither Questioning the Author nor Reciprocal Teaching focus on the same information that Close Reading encourages students to pay attention to or to construct. Nor do either of them encourage even conditional suppression of prior knowledge. I have long written that close reading as a reading technique has no research base--but the goals of close reading (and the emphasis on symbolism, dual meanings, ambiguity, tone, etc.) have not been addressed in any of those techniques.



Jake Downs
Apr 15, 2021 02:58 PM

This is a very important discussion I love that CCSS, Kintsch, Rand, and QTA/RT are all in the mix.

Perhaps the issue is as much when close reading is valuable as opposed to if close reading is valuable. Kintsch states that comprehension occurs in a multi-stage process; first construction, followed by integration (Kintsch, 1988). My read of Kintsch suggests that it is indeed impossible to stay within the 'four corners of the text'- background knowledge is essential. However, we can learn from texts from which we have limited background knowledge. During the construction process the readers forms a coherent basic representation of the text using mainly bottom-up processing (Kintsch, 2019). Pearson (2013) interpreted construction as focusing on 'what the text says' (p. 252). This level of comprehension is rather basic and can be accomplished with varying degrees of background knowledge (Kintsch, 1996). Perhaps close reading is valuable in the sense that it can emphasize the 'four corners of text' to build a basic understanding of the text, which can then be synthesized and integrated with other knowledge during the integration phase. Thus, close reading could be an important component of reading comprehension instruction, but could certainly result in diminishing returns if overemphasized. Further there are likely multiple techniques to mapping out a coherent text base, as the research on reading strategies, reciprocal teaching, and questioning the author suggests.

There are some caveats. Even though Duke and colleagues (2011) indicate that Kintsch's model is the most complete model we have, the model does not suggest what exactly this means for practice (as others have pointed out in this strand. See also Shanahan's articles in the Reading Teacher and RRQ on the Science of Reading).

Dr. Dewitz suggested in this thread for those interested to read a historical overview of the field. I highly suggest Pearson and Cervetti's 2015 piece 'Fifty Years of Reading Comprehension Theory and Practice' in "Research-based Practices for Teaching Common Core Literacy, edited by P David Pearson and Elfireda H. Hiebert.

Linda Diamond
Apr 18, 2021 07:01 PM

Call me old. When I was taking Rhetoric at UC Berkeley a long time ago, close attention to text was what we lives and breathed. In addition. As a Jr. Great Books discussion leader. Paying close attention to text was what we did. That did not mean we also failed to teach some comprehension strategies. What frustrates me most is we seem always stuck in either or debates. I appreciate the pragmatic discussion Tim has here.

Apr 20, 2021 01:39 PM

Hi Tim,
Thanks for such an informative post and discussion. I’m interested also in knowing your thoughts on the question raised by Will above regarding the CCSS’s references to knowledge building.
Connected to Will’s question, I would like to know if the trend on knowledge rich curriculums is likely to have a beneficial effect on reading. If so, should all curriculums be knowledge rich?
Many thanks,

Timothy Shanahan
Apr 20, 2021 01:44 PM

Will and Mat--

Indeed, I think it wise to teach reading using texts that facilitate students learning valuable information about their social and natural worlds -- including high quality literature, as well as texts on social studies, the science, and the arts. In the long run, that is likely to have some modest but positive impacts on reading comprehension (for the same reason vocabulary instruction has a modest impact on reading comprehension -- the knowledge and words have to match with the text content to have a facilitative effect).



Anila Nayak
Jun 19, 2021 10:37 PM

Hi Patrick,
I checked out the links for the TAW lessons you provided. I am studying them in greater detail and was wondering if I could aso get the copy of texts that the students were reading.

Apr 11, 2021 02:34 PM

Morning, Tim. You wrote in a comment, “What isn't clearly supported is ... simply teaching more social studies and science with the idea that such instruction would generalize to better reading of texts on other topics.”

But the ELA CCSS note on range and content of student reading states,
"... By reading texts in history/social studies, science, and other disciplines, students build a foundation of knowledge in these fields that will also give them the background to be better readers in all content areas. Students can only gain this foundation when the curriculum is intentionally and coherently structured to develop rich content knowledge within and across grades. ..."

Can you help explain the difference in opinions? Thank you!

Leah Falkowski
Apr 08, 2021 05:47 PM

The vocabulary and other parts of written language can help with comprehension is debated with the research as it may benefit the student or may not.

Michael Grave’s early studies stated pre teaching vocabulary, prior knowledge and knowing key concepts can help with comprehension.
There is a discrepancy with the science of reading as far as public testing an reading
programs with close reading.

A comment that stuck out with me was about the rose colored glasses and it saying there was no real foundation in teaching the science of reading. I disagree as I think it important for students to get the basic instruction to learn how to read. .

There is a debate if prior knowledge for students helps with reading comprehension as it can help but not help depending on what the subject matter is.

Does close reading affect comprehension by taking the time to do explicit and direct instruction?

Timothy Shanahan
Apr 08, 2021 07:07 PM

I made no claim that there was not a science of reading; there is no justification for the claim that kids will make the best progress in learning to read if they are taught from text sets (multiple texts on a topic). Despite the claims made in the article I was responding to, there is no research evidence supporting that particular claim (and when one goes back and looks at documents from the past, it is possible to see the kind of mental gymnastics required to make such a claim).

There is no question that we use our knowledge when we read. There is no question that this can both facilitate our comprehension and mislead it. There is no question that teaching kids the content they are to read about will improve their comprehension of that particular text (if only because they will have gone through the same content twice). There is no question that reading can be taught successfully using texts drawn from social studies and science. There is no question that directly teaching students about how to handle particular text elements (like tables and charts) or how to use particular strategies (like self questioning) has a positive and generalizable payoff in terms of improved reading achievement. Those points are well supported in the research literature.

What isn't clearly supported is the idea of dropping reading instruction for older students (grades 2 and up), or simply teaching more social studies and science with the idea that such instruction would generalize to better reading of texts on other topics.



Leah Falkowski
Apr 10, 2021 04:51 PM

Thanks for your feedback.

John Young
Apr 10, 2021 05:35 PM

I was interested in the comment “Even when researchers have looked at comprehension instruction, the studies have almost always been so brief (6-8 weeks) it would be malpractice to base 13 years of comprehension instruction on those alone (something Dewitz and Graves properly complain about). Just because two months of a particular approach to instruction improved learning in a study, one should not conclude that several years of such teaching would continue to be either efficient or effective. Again, hard to criticize anyone for not adhering slavishly to that research”:
Quite a few years ago I did a literature review on sustained silent reading (unpublished). My view was that many studies were in my view too brief too expect them to make a difference in students achievement.

Apr 10, 2021 06:57 PM

There's a lot to think about, and these thoughts are front and center as I teach grade level text to my third graders. You say:

"But there are interpretive benefits to treating a text as being self-contained object, separate from historical or cultural contexts. It makes great sense to teach students to try to understand a text through consideration of the words and structures in the text, as opposed to doing so with the external information that teachers may provide."

This dance of the "do I tell them or do I let them learn about it as they read" is so important. Yesterday, we read an article about Stephen Curry's efforts to combat malaria. Here's what I didn't explain to my students before reading the article because this knowledge could be gained through the text: malaria, ambassador, refugee camp, albino.

Where does the need for prior knowledge stop and practice reading to learn by analyzing text begin?

Timothy Shanahan
Apr 10, 2021 07:31 PM

That is an interesting contrast. For the most part, those 6-8 week studies of reading comprehension strategy instruction have resulted in significant improvements in reading achievement, while short studies of encouraging reading have not (even though they have usually been longer in duration). That underlines a point I have long made -- kids make more certain gains from reading instruction than from independent reading. There are many other problems with the encouraging reading literature as well. Not only are many of the studies too short to assume one would see any kind of gain from that weak an intervention, but for the most part they make no effort to determine whether or how much the efforts have increased the amount of reading the students do. If providing books and time to read doesn't actually increase the amount that students are reading, then it is foolish to expect it to improve achievement.


Timothy Shanahan
Apr 10, 2021 07:36 PM


You raise one of the big problems with prior knowledge and some of the simplistic conclusions being drawn. I often read text for which I lack much prior knowledge. I definitely have to work harder at those texts than on those for which I have greater background. Nevertheless, I can usually make sense of them -- though, admittedly, there are times when gaining some supplemental knowledge is necessary. We simply don't know what are the long term effects of providing substantial prior knowledge for each text the students read. One of the ideas of close reading is to attempt to take on at least some texts without that kind of support.



Patrick Manyak
Apr 11, 2021 12:33 AM

Thanks to Tim and Dewitz & Graves for raising so many important questions and for pointing out where the extant research falls short in answering them. Our recent approach to addressing comprehension/textual analysis has been to construct a dense (in terms of instruction), feasible, multifaceted routine that supports students as they read and analyze challenging texts. We first worked on a routine for reading challenging children's fiction (we published a practical piece on this: and are now testing a routine for informational texts in science and social studies. These current studies are simply exploratory, aiming to develop, implement, and research the routines in a very local fashion. However, it does appear that the approaches have promise, and we hope to test them more thoroughly. Here is my point in bringing this project up: We aren't worried too much about what to "call" what we do... Fluency instruction? Yes, some of that. Close reading? Often, I think. Text structure instruction? Yes, when called for. Some explicit strategy instruction? Probably a little less of that than other things, but yes, when the text calls for it. Vocabulary? That, always. In essence, we are following a common blueprint: 1) very brief instruction/practice reading long words, 2) brief but rich vocabulary instruction, often explicit word-meaning teaching but, when relevant, also some context cue instruction/practice, 3) assisted/repeated reading to support fluency (over several days), 4) instruction/discussion in textual analysis, 5) instruction/application in response writing. Now, I think that each of these areas is vitally important, but #4 is closely related to the discussion here. And, with regard to it, we basically follow the approach of "what does the text provide for or require" in terms of comprehension/textual analysis. Many times, I would probably say that what it provides for would fit under "close reading." In a number of cases, a significant part of this involves recognizing and using text structure to facilitate analysis. However, at other times, what the text requires would probably be called comprehension strategy instruction (inference, visualize). From my reading of comprehension research, prompting students to develop this kind of flexible thinking about text and a variety of specific strategies and analytic tools for doing so, all while really focusing on the engaging and valuable content of the text at hand (we have used a variety of articles on ecology, climate change, and now Indigenous Americans) makes a lot of sense. If anyone is interested in checking out what these "Textual Analysis and Writing" lessons look like, here are a couple links to some examples that appear at different places in the sequence of lessons: (Links to an external site.) (Links to an external site.)

What Are your thoughts?

Leave me a comment and I would like to have a discussion with you!

Comment *

Does Close Reading Reject the Science of Reading?


One of the world’s premier literacy educators.

He studies reading and writing across all ages and abilities. Feel free to contact him.