Friday, March 22, 2013

The Common Core is not "research-based" and its "efficacy" is unproven

In poking around looking at PARCC (the new testing associated with the Common Core) I stumbled onto an interesting document: the "publishers' criteria" set out by the two lead authors (Coleman and Pimentel) of the ELA Common Core "State" Standards. These criteria are intended to guide publishers in their development of curricular materials for teaching under the Common Core regime. (Although I don't actually know any teachers who use a publisher's curriculum materials, many must do so, since producing these materials seems to be a huge industry.)
Here's the conclusion--again, this is Coleman and Pimentel telling publishers how to create materials for teaching to the Common Core standards:

Curriculum materials must have a clear and documented research base. The most important evidence is that the curriculum accelerates student progress toward career and college readiness. It can be surprising which questions, tasks, and instructions provoke the most productive engagement with text, accelerate student growth, and deepen instructor facility with the materials. A great deal of the material designed for the standards will by necessity be new, but as much as possible the work should be based on research and developed and refined through actual testing in classrooms. Publishers should provide a clear research plan for how the efficacy of their materials will be assessed and improved over time. Revisions should be based on evidence of actual use and results with a wide range of students, including English language learners.

This paragraph may seem fairly reasonable on its face; nevertheless, I have two thoughts about it.

I. The Common Core does not, itself, "have a clear and documented research base"

My first thought is that what Coleman and Pimentel say about publishers' Common Core-aligned materials seems quite relevant for the Common Core itself.  It is as if Coleman and Pimentel had realized all the things that should have been done, but that they didn't do, in writing up educational standards for the entire country.

The Common Core is hardly "research-based"; the research base on which it rests is incredibly flimsy.  The authors of the Common Core make their case, such as it is, in an appendix purporting to offer, among other things, "research supporting key elements of the standards."  The relevant section of this appendix is only three pages long and offers "research" worthy of an undergraduate paper or a blog post, not of a major national endeavor.  Its three pages make roughly the following case:

1) College performance correlates with the ability to read and understand difficult texts, and especially expository texts. (This is probably true.)

2) Complexity of texts assigned in high school has declined over the last 50 years, and high school students read relatively little expository text. (This may be true.

3) Therefore, students need to be assigned more difficult texts, and more of those texts should be expository. (This third part neither follows logically nor is supported by empirical data.)

As I noted, the first part of this argument seems very likely, and the second is plausible, but the third part is very problematic, neither seeming logical nor being supported by the historical record or empirical research. The gap in logic is obvious: just because kids who are better at reading complex expository texts do better in college does not mean that most of the reading kids do in high school should be of complex expository texts, and just because the complexity of texts assigned in high school has declined somewhat over the past several decades does not mean that assigning more complex texts is the right remedy.  For these conclusions to be valid, there would have to be empirical support either of the historical or experimental variety.  There is neither.

First, the historical fallacy.  Robert Hass tells us that all poetry is about loss.  The same is true of the stories Ed reformers tell—we have lost the good old days when teachers taught more rigorously and even poor students could achieve like the rich kids.  The problem with this story is that it is by no means clear that students were more college-ready 50 years ago.  The only historical evidence the Common Core authors cite is what they call a "statistically significant" decline in adult reading proficiency.  What is the actual decline? From 15% in 1992 to 13% in 2007.  This hardly seems precipitous, it covers fifteen years, not fifty, and is belied by (slightly) rising reading scores on the NAEP.  The historical record does not provide a clear argument for making students read more difficult texts or having more of those texts be expository--and Ed reformers, unlike poets, don't have the excuse of poetic license. (For a longer analysis of this historical fallacy in the stories Ed reformers tell, see here.)

Second, there is a shocking lack of experimental data.  As usual in writing about education, the discussion in this Common Core appendix is a mish-mash of much unfounded assertion and some offhand citations of actual empirical research; also as usual, even when there are references, the articles cited often fail to support the assertion.  I'll discuss just one example--the first specific citation I looked at--but they are almost all equally embarrassing.

     References that don't support what they are cited to support

I am skeptical of the idea that students need to read a lot of specifically expository text; my suspicion, based on my own and others' experience, is that the key thing is to simply to read a lot, and that a high volume of reading even of trashy airport thrillers will lead, with only a bit of specific practice, to skillful reading of expository text.  The Common Core document states that "students need sustained exposure to expository text to develop important reading strategies." Now, I doubt this very much, but while I only have anecdotal evidence (me, Malcolm Gladwell, and everyone else of our generation, who learned to read before reading strategy instruction was current) to support my skepticism, the Common Core folks seem to have more: they cite several articles to support their statement.  I looked at the first citation (Afflerbach, Pearson and Paris, 2008), a 2008 overview from "The Reading Teacher" about the difference between strategies and skills (unsurprisingly, the article suggests that skills are unconscious, strategies are conscious, and you need both; I agree, but I'm skeptical that these things can be explicitly taught or usefully assessed)--the article does not support the assertion it is cited to support.  

I have now read the article three times looking for evidence that "students need sustained exposure to expository text to develop important reading strategies," and I have so far failed to find any evidence at all.  Thinking maybe I'd missed it, I searched for "expository," for "exposure" and for "sustained".  None of those words appears in the text.  Neither does "non-fiction."  Looking through again, the closest I could come to anything supporting the claim that you need to read a lot of expository text in order to develop reading strategies is the following very general assertion about practicing strategies, which is completely untethered to any actual data:
"The scope and complexity of these strategies are large, and there is ample variety of text difficulty and genre variety to practice so that the skills become automatic. The general rule is, teach children many strategies, teach them early, reteach them often, and connect assessment with reteaching." 

First of all, we should note that this "general rule" is, like most general rules in writing about education, totally unproven and highly dubious. Many people, like me, like Ben Franklin, like Frederick Douglass, received virtually no formal instruction, and absolutely no assessment, in reading strategies, and yet learned to be highly skilled readers and writers.  Second, this brief, unfounded passage bears little relation to the Common Core's assertion that you need to read a lot of expository text in order to get better at it. If this-- "there is ample ... genre variety to practice so that the skills become automatic"--is supposed to mean this--"students need sustained exposure to expository text to develop important reading strategies"-- then either I am a bad reader, David Coleman is a bad reader, or he simply has no idea what a "clear and documented research base" would mean in a field that was, unlike education, actually scientific.

Again, the first citation offered by Coleman and Pimentel to support one of their central claims provides absolutely no support; far from reporting research or empirical data, the article never even mentions the matter at hand.  This is still amazing to me, despite my having found this to be the case with so many other supposedly "research-based" recommendations to teachers.

II. Books and (maybe) a teacher are all students really need

The other thing to notice about the Common Core’s recommendation to publishers is although the authors say that most of the materials "will by necessity be new", this is probably untrue. In fact, developing new materials may be unnecessary, since books alone would seem to fit most of their requirements.  Books support student "readiness for college."  Books have "a clear and documented research base."  Books have a long history of "actual use" with a "wide range of students."

The best ELA program for infants is simply a lot of natural adult talking and reading aloud, and the best college ELA curriculum is simply good books and an engaging professor, but somehow schools have fallen into a muddy puddle of "instruction" and "curriculum materials." Also, one of the major stated purposes of the whole Common Core/PARCC effort is to make sure young people are prepared for college.  I wonder why, then, the best private high schools and the best private colleges aren't using publisher-created curriculum.  Is it possible that Andover and Harvard are delivering a sub-par product.  Perhaps--but far more likely is that either their curriculum is better or that the curriculum really doesn't matter all that much. 

The Common Core anticipates that the questions, tasks and instructions used with readings will be created by publishing companies.  This begs the question of what role, if any, we teachers are supposed to play.  I suppose eventually we will be replaced by computer programs.  That might be okay...


Except that it's not okay. Children need human connection, and the best thing that we do every day is provide that connection.  If scholars and bureaucrats with zero teaching experience can tell us what we can do that will help us connect better, great!  But all too often--that is, almost all the time--their recommendations are shockingly unfounded on any empirical data.

On the one hand, many of the CC's specific recommendations—like the suggestion that students focus on close reading, or the observation that “it can be surprising which questions, tasks, and instructions provoke the most productive engagement with text”—seem reasonable, but the hypocrisy and hubris of the whole enterprise give off a very questionable smell

No comments:

Post a Comment