October 21, 2005

A problem with research?

A perspective published in EdWeek.org argues that good, solid research isn't that helpful in the educational system nowadays:

I no longer believe that education research will turn our schools around. And it’s not likely to help us to fix our ailing schools for very specific reasons. First, research is not readily accessible—either physically or intellectually. The findings tend to be written for other researchers in academicspeak and appear in relatively obscure journals.

True enough. The general incomprehensibility of educational jargon has been widely mocked in the real world.

Second, even if research findings were more accessible, they wouldn’t be widely read. Teachers, principals, superintendents, and politicians are generally not consumers of research.

And whose fault is that? Don't the first three groups listed above generally hold at least a bachelor's degree? If they aren't consumers of research, perhaps the problem lies in their training to be part of the educational system.

The author continues with a few more reasons that research isn't used, or used well, in today's educational environment, before moving on to this:

Sixth, much education research is flawed because it relies so heavily on a flawed measure—standardized test scores.

Anyone surprised that this attitude shows up here? No? Let's move on.

Test scores may be the only “objective” data available, but they’re not necessarily a reliable measure of student learning.

Ah, no. "Reliability" has a very specific meaning when it comes to testing, and in an article that talks about how educators aren't generally consumers of research, it's ironic - or perhaps not - that the author appears to be misusing this technical term. Most, if not all, multiple-choice standardized tests that are developed for use are in fact highly reliable, because reliability in testing means consistency of measurement, or the extent to which the results are similar over different forms of the same test or occasions of test administration. Items that are objectively scored are more likely to produce similar test-retest results for examinees.

Exams composed of primarily multiple-choice items tend to have high reliability, and in fact most test developers shoot for a reliability of .90 for these types of exams. The addition of open-ended items, performance-based items - in fact, all the less-structured types of assessments that so many educators are fond of - usually lessen the reliability of these types of exams.

So I'm not sure why the author used low reliability as a criticism here, since that's probably where the least amount of error appears in this type of exam.

Nor do they measure many of the traits we hope schooling will produce in kids—like good habits of mind and behavior. They don’t measure Howard Gardner’s other intelligences, like artistic talent, athletic prowess, or social skills. After kids leave formal schooling, they’ll be judged for the rest of their lives on the quality of their work and their personal and professional behavior. Test scores are a poor proxy for those qualities and for a wide range of other skills and abilities.

Ah, the old "But my kid is so high in kinesthetic intelligence!" argument. What testing critics are hoping you don't notice with this type of criticism is the fact that, if you can't read and write and do basic calculations - skills for which test scores tend to be extremely good proxies - your chances of economic success in our society are extremely low, regardless of your academic or artistic abilities. Sure, there are kids with rock-bottom SATs who make big bucks on stage or on a playing field, but the percentage of Americans who make a living with those skills alone is pretty darn small.

What this type of testing critic wants you to conclude is that kids who do well on standardized tests have learned many literacy- and numeracy-related facts without really understanding them, that these skills are utterly separate from other mental and physical abilities, and that the development of skills that are measurable with tests always happens at the expense of other critical skills. I think that's nonsense. You want to teach your kids good habits of mind, good social skills, and some touch football or ballet as well? Then explain to them that, unless they're prodigies, they'll be supporting themselves with their minds, not their bodies, later on in life, and skills such as discipline and teamwork will serve them just as well later on life as they will on their upcoming exams.

Does research show that high test scores predict everything a kid will do later in life? Of course not. But I think there's sufficient research to show that low test scores are a sign of a real problem, and a strong indication that intervention is needed. Maybe if schools of education impressed this upon the would-be teachers and principals, educational research would have a bit more impact on education today.

Posted by kswygert at October 21, 2005 09:51 AM
Sitemeter