International assessments, such as the PISA or TIMSS tests, are often used to compare and rank the quality of education across different nations. Martin Carnoy, the Vida Jack Professor of Education at the Stanford Graduate School of Education, joined GSE Dean Dan Schwartz and Senior Lecturer Denise Pope on School’s In to discuss the impact these types of assessments have on education policy.

Listen to the full episode at the link below and find more episodes at Stanford Radio. School’s In airs weekends on SiriusXM Insight channel 121.

Interview Highlights

Why countries participate in international tests

There's been pressure by the World Bank, and of course by the test givers, to get more and more countries to take these tests. The reasoning behind it is that you’ve got to know even if you score low; it would be good to know how your kids are doing vis-a-vis the rest of the world. [But] in fact, the framers of these original tests back in the 1960s really warned about not doing these country comparisons. They said, “Don't do this—what we're really trying to find out is why some educational systems are different from others, and we want to use these tests to try to figure out.” Well, that's been lost in all the publicity.

Comparing states instead of countries

States vary enormously in how they run educational systems. We know that, so it makes little sense to say, “Well, how does Finland compare with the United States?” Finland is a country of five million people compared with a country of 330 million. And we have this tremendous variation within the country.

What I suggested we should do is look at state performance and start to compare with states, because first of all, Americans are much more like Americans than they are like Finns. And secondly, state policies vary, so it's really useful to do that.… These kinds of [state] comparisons are actually useful to policy. Unfortunately, international tests are not very useful for giving us policy implications.

Interpreting data

Economists, sociologists and everybody else just love data, and one thing you can say is that these tests have produced tons of data over the years. There are two sides to the argument. We love the data but we know that there are real problems with it, and particularly we know the real problems are the way they're used in order to convince people that some things are good in education and other things are not. And they really jump to conclusions, which are not evidenced by their data.