Willingham: Feds should leave ed policy to states
My guest is cognitive scientist Daniel Willingham, a professor at the University of Virginia and author of “Why Don’t Students Like School?”
By Daniel Willingham
Suppose you read a research study which concludes that a particular drug is effective and has minimal side effects. Would you be less confident in the conclusion if you learned that the researchers were full-time employees of the company that manufactures the drug?
Most of us would be. You needn’t assume that the researchers are dishonest. A great deal of research in social psychology over the last 30 years has shown that people often fail to be impartial in their judgments if they have a stake in the outcome, even if they want very much to be fair.
I am therefore troubled by the lack of independence between state and federal entities that create education policies and those that evaluate educational outcomes.
We have seen several examples in recent years of what might be unconscious bias, if not downright research shenanigans or political spin.
On several occasions the New York Department of Education claimed that policies in the system were leading to substantial gains, but intrepid blogger Jennifer Jennings (aka Eduwonkette) offered analyses of the data that many took to be more appropriate, which showed a less rosy picture.
At the federal level, Margaret Spellings, education secretary under president George W. Bush, offered data interpretations that indicated a positive impact of the No Child Left Behind Act, but again, other analysts argued that the effects were minimal or absent.
It simply doesn’t make sense to have the purveyors of policy telling us whether or not the policy is working.
I suggest that the federal government largely remove itself from education policy-making and put more effort into evaluation.
Simply trying to ensure greater independence between policy makers and the evaluators probably won’t work. There would still be doubt as to whether there is true independence.
In addition, it ought to be useful to compare student performance across states and that requires a national measure.
Whether it were a new version of the National Assessment of Educational Progress--sometimes called "the nation's report card"--or some other test, suppose that parents received by mail a report showing their child’s score on the test, along with the average score for their child’s school, the state average, and the national average. Would this not make it plain to parents how things were going with their child and at the school as a whole?
The state-wide information ought to be informative as well. If I lived in Louisiana or New Mexico, or another very low-scoring state, I think I’d ask my representative, “Why don’t we just try whatever it is Massachusetts does?”
But we need to think beyond the type of information that is measured by the NAEP. As I have emphasized before, it’s impossible to make progress on educational goals in the absence of measurement. Yet we measure outcomes relevant to only a few of the goals that many people would list for high school graduates.
We can measure factual knowledge and to a lesser extent we can measure students’ ability to solve problems. We don’t have good measures of whether kids like learning, whether they are resourceful about teaching themselves things, whether they are creative, or whether they can work in teams. In fact, name a goal other than “learn facts” and the odds are that we can’t measure it.
Factual knowledge is important, but it’s not the only thing we want our students to learn. Yet there is surprisingly little interest in developing measures for these other constructs. Of course such development is difficult. That’s why I’m arguing that it ought to be the responsibility of an organization with substantial resources--the federal government.
What are the likely objections to such a plan?
Might a test not create the problems associated with No Child Left Behind, namely curriculum narrowing and test-prep frenzy? As I envision it, the federal government would not attach sanctions or rewards to the outcome— in edulingo, the results are formative, not summative. If a state is going to use the test that way, well, that state would probably do the same thing with a state test.
Another objection might be that it’s not possible to create a single test for all fifty states. I don’t hear that criticism leveled at the NAEP all that often, but if a state finds that the national test doesn’t reflect their educational priorities, state legislators should not find it difficult to explain to their constituents why that is the case. For example, Texans might want to ignore the history results, as the test might well make some mention of Thomas Jefferson.
Finally, one might argue that we need less testing, not more. I argue that we need more tests that are well designed and sensibly interpreted, and fewer poor tests the results of which are twisted out of shape by policy-makers with an axe to grind.
If we are to know where things are going well and where they need improvement, we must measure outcomes. I think the federal government should evaluate, and leave policy to the states. I’d be happy with other solutions, so long as there is a wall between policy and evaluation.
Follow my blog all day, every day by bookmarking washingtonpost.com/answersheet And for admissions advice, college news and links to campus papers, please check out our new Higher Education page at washingtonpost.com/higher-ed Bookmark it!
| March 29, 2010; 11:00 AM ET
Categories: Daniel Willingham, Guest Bloggers, No Child Left Behind, Standardized Tests | Tags: Daniel Willingham, NCLB, guest bloggers, student assessment
Save & Share: Previous: Broaddus--The limits of proxies in college apps
Next: Arne Duncan the heartbreaker and Race to the Top
Posted by: SeaTigr | March 29, 2010 12:49 PM | Report abuse
The comments to this entry are closed.