September 20, 2007

Why IQ really matters in education

Wes Becker makes the following point regarding IQ and a student's ability to learn in his book Applied Psychology For Teachers:

The notion that IQ measures the ability to learn is challenged by the cited logical arguments (on level scores and gain scores) and the actual data. In reviewing this issue, Cronbach (1970) concluded that the "ability to learn" is not a satisfactory construct. More relevant is the question "Learning what from what instruction?" There are many kinds of learning tasks, having different prerequisite skills, that make the new learning possible or impossible. And, there are many kinds of "instruction." He concludes: "There is a suspicion afoot that education calls for analytic ability just because the materials are capable of being put into meaningful relationships and the instruction has either failed to display the relationships or has given an explanation that is hard to follow. Then one has to use his brain!" (Cronbach 1970)

Cronbach is suggesting that the high-level analytic ability measured by IQ tests may be affecting learning in situations where the teaching is bad, in addition to situations where untaught vocabulary is required. Under poor teaching conditions, only intelligent students have a chance.

I think Becker is exactly right (even if he is understating the IQ and "ability to learn") relationship). IQ matters very much when the instruction is poor; not as much when the instruction is well designed and presented by a skilled teacher. You'd still expect the higher IQ kids to perform above the lower IQ kids, but the difference in performance should be less.

Becker concludes:

What any person can be taught depends on what he or she has already learned, as well as the methods used in instruction. If the method of instruction is lecture, and the same assignments are given [to] thirty students of highly different backgrounds, it is very likely that some will fail because they lack the preskills assumed by the teacher. In this situation, an IQ test could, in part, predict who would succeed and who would fail. If the method of instruction systematically builds upon the skills each individual brings to the educational setting and uses good motivating procedures, each student will learn to the degree that he or she is effectively taught.

The fact that human beings are products of their genetic histories does not limit what they can learn or guarantee that they will be taught. Genetic histories probably do have some influence on how well students learn, but since there is no measure of or control over the variables involved, the potential influences have no practical implications for the teacher.

Today, IQ is a brutal predictor of academic success and will remain as such (at least on the K-12 level) as long as teaching methods remain at their current primitive level.

Before education was made compulsory, it was the province of the wealthy and/or bright. Instructional techniques were developed over the millennia to service this group with the dimmer members self-selecting their way out when their ability to learn did not measure up to the instructional presentation. Only having to educate the bright, relieved educators from the burden of developing instructional techniques that were adequate for educating the average and dim.

When educating was made compulsory for everyone, it was soon noticed that educators were incapable of educating the unwashed masses with existing instructional techniques. It did not help matters that the job of education was taken over by government, an entity not known for its ability to innovate. And, innovation was dearly needed if education was to reach the unwashed masses.

Instead, IQ tests were developed during this period as a tool for government to determine which students were likely were to have trouble in school so that they could be given special help. Special help that schools did not know how to deliver and which would generally be ineffective. Thus, IQ tests grew out of the problems teachers were having after education became compulsory.

IQ tests remain a valid predictor of academic success today as they did when they were developed a hundred years ago. That's because the educational conditions in our public schools haven't improved since then. And, it doesn't look like there is the political will needed to effect the drastic change needed to improve education conditions, NCLB notwithstanding.


Anonymous said...

What do you make of these David Brooks musings?

KDeRosa said...

Here's a good point by point take down

Anonymous said...


Anonymous said...

Nice to see Wes Becker cited. A great and underappreciated man.

Some other studies that will interest you:
Engelmann, S. (1970). The effectiveness of Direct Instruction on IQ performance and achievement in reading and arithmetic. In J. Hellmuth (Ed), Disadvantaged child: Vol 3, Compensatory education: A national debate. New York: Brunner/Mazel
This is a very technical article with tables etc but shows how the program at the Engelmann-Bereiter preschool raised the IQs of EVERY child to over 100 (some had started very low indeed -- all were from the most deprived tenements in the area. This study predated Project Follow Through and is not mentioned in any of the FT data). These kids grew up to vastly outperform their likely prospects based on their entry-level IQ and SES (one girl became an engineer, another a bank manager, etc.)

Another book to check out is Arthur Whimbey's classic, "Intelligence Can Be Taught". It is out of print, but you can likely locate a copy used through abebooks or For a more detailed account of the Bereiter-Engelmann preschool program, get the (also out of print) book by Bereiter and Engelmann titled, "Teaching Disadvantaged Children In The Preschool."

One of the salient points made by Engelmann is that disadvantaged/low-IQ children bring a very limited language repertoire to the task of learning. It is more than merely not having been in a richly stimulating environment -- language is a tool for thinking, and the restricted language input (see also Hart and Risley's classic "Meaningful Differences" re differences in preschoolers language exposure) cripples their cognitive growth. The Bereiter-Engelmann program specifically targeted developing language/reasoning skills (also numerical reasoning, but through language rather than use of "manipulatives"). This is still true today -- in my own school most kindergartners enter at below the 20th percentile in language, social skills and other developmental milestones (the district does some kind of comprehensive survey on a rotating basis).

The Bereiter-Engelmann results were replicated elsewhere but never "caught on." The educational world was not looking for a better mousetrap then, and still isn't. Thus the predictability of IQ is much like the "developmental stages" literature -- this is what happens if you don't intervene effectively. Nature takes its course, more or less. If you DO intervene effectively, all bets are off.