12 Sep 2005 - 03:00

E. D. Hirsch on Cargo Cult Research

I read E.D. Hirsch's article on Classroom Research and Cargo Cults last night. It's a short easy read, and a worthwhile one, I think.

I believe, like practically everyone else, that ed research is crummy. I think that's undisputed; everyone is claiming that education research supports their point of view, because everyone can, because ed research results are all over the board. They add to the confusion, rather than clearing the fog. I thought that that was probably because the field is backward, in its scientific infancy, and that they don't understand things like statistics and random sampling and longitudinal studies and experimental design in general.

But Hirsch claims that the statistical methods used in education research these days are of good quality (and he credits a recent article, by Thomas Cook and Monique Payne, for bringing good experimental design into the mainstream of educational research practice). He believes that the problem with classroom research is that it cannot be sufficiently controlled to eliminate extraneous influences that throw the results off (one such influence, for example, would be the effect of the teacher's knowledge and personality, which is, as any parent knows, enormous). This problem, he thinks, results in classroom results being irreproducible and therefore unreliable for guiding education decisions. They are valid as far as they go, he feels, but we shouldn't regard them as science, or try to make policy from them.

He believes that better inferences for education policy can be made from highly controlled research on cognitive science. Cognitive science is actually converging on a consensus about how people learn, and what practices increase learning efficiency. These principles/practices are:

Prior knowledge is a prerequisite to effective learning. Bernie and I used to say, when Ben was little, that it was hard for him to learn things because, having had a condition that made it hard to attend to his environment, he had few 'hooks' on which to hang new knowledge. By hooks, we meant prior knowledge that we could draw analogies to. This principle is just that: that learning is improved when hooks in the form of prior knowledge are present. A novice will learn less than an expert from a new scenario, even though he is a beginner and has more to learn, because he has less context to base new learning on.

The right mix of generalization and example is critical. Good teaching goes from an illustrative set of examples to the general case. For example, you wouldn't just demonstrate the distributive property formula:

a(b+c) = ab+ac

to a bunch of first graders. Instead, we work for several years on examples, learn the multidigit multiplication formula, and so on, before introducing the distributive property in its full generality. We need to do this because the abstract concept and the specific example are inextricably linked in people's minds.

Attention determines learning. Surprisingly, motivation isn't a prerequisite to learning, but attention is. If attention is paid to something, and if we have a 'hook' to hang the knowledge on, we'll learn it, plain and simple.

Rehearsal is usually necessary for retention. How long something is remembered depends on how long it's been attended to. There is a "sweet spot" of practice past which things are permanently remembered, and practice that is spread out in time ("distributed practice") is much better than cramming ("massed practice").

Automaticity (through rehearsal) is essential to building higher skills. Our working memory (our mental scratch space) is extremely limited. Practicing a skill to complete automaticity frees up working memory.

Implicit instruction of beginners is usually less effective. Cognitive scientists actually give a complex answer to the question of whether explicit instruction (in early reading, for example, this would be 'phonics' instruction) or implicit instruction ("whole language") is more effective. The consensus is that both are required. In tennis coaching, for example, drills that isolate skills are desirable, but actual games must also be played. The mix should grow more implicit, and less explicit, as expertise grows.

What I want to know now is this: is there a good book yet on what cognitive science has to say about learning? Willingham's series of articles on cognitive science is excellent, but I'd like to know whether there's something more in-depth that talks about the specific studies that support these principles.

And: are cognitive science studies admissible as research evidence in support of NCLB's goals? If not, how can we get them admitted, and test the degree to which a given curriculum adheres to them?

Back to main page.

---

---

Comments

After entering a comment, users can login anonymously as KtmGuest (password: guest) when prompted.
Please consider registering as a regular user.
Look here for syntax help.

---

Fabulous post! Thank you for summarizing this great info for all of us. I'm sure I'll want to refer back to it when as my son grows (he's 2 1/2 now).

-- StephanieO - 12 Sep 2005

---

What a fantastic post!!!!!

Great, great work, Carolyn.

We need to have this linked on the sidebar, so newcomers can find it immediately.

I've written about cog sci for years, and majored in psychology with an emphasis on cognitive psychology as an undergrad, and all of this is true (as far as I understand, of course).

These are the principals I think about every day, whether I'm teaching Christopher or teaching myself.

-- CatherineJohnson - 12 Sep 2005

---

I'm positive--though I haven't fact-checked this--that cognitive science studies are germane to NCLB.

-- CatherineJohnson - 12 Sep 2005

---

The one statement I would quibble with here is the implication that educational research is now on par with research in other fields. I'm fairly certain that's not the case, given the fact that the government just had to put through a highly-protested bill saying that the Department of Ed could consider whether or not a study has a control group in making funding decisions.

-- CatherineJohnson - 12 Sep 2005

---

Also!

Offhand, I wouldn't go with Hirsch's view that ed research will never be reliable because you can't control for all variables.

That would rule out Caroline Hoxby's research, and anything done by economists, who, IMO, are doing some of the most interesting studies going these days.

-- CatherineJohnson - 12 Sep 2005

---

One surprise: he says motivation isn't a factor?

I'll have to read the article.

I do believe that motivation isn't a factor if you've got parents to serve as your frontal lobes--i.e., parents who make you sit down and do your homework.

This is something I'm constantly telling the parents around me, many of whom believe grade school children should be independent, which means they should organize themselves and be responsible for their own homework. (Stevenson & Stigler say that independence is a major American value--and it is!)

I'm always telling people, Kids don't have the frontal lobe capacity to organize themselves. That's your job.

-- CatherineJohnson - 12 Sep 2005

---

The point about working memory is incredibly useful in conversations with parents, teachers, etc.--because everyone knows this at some level. They just haven't put it into works, or thought about the implications.

Whenever I tell a parent that her child needs to be fast on his math facts because of the limitations of working memory, she instantly gets it, and believes it.

The working memory studies are a case where the findings of cognitive science validate and formalize folk psychology.

-- CatherineJohnson - 12 Sep 2005