Tuesday, April 1, 2008

CATO and data

http://www.cato.org/pressroom.php?display=comments&id=881

A couple weeks ago a friend forwarded me a CATO piece decrying the American Community Survey, which is presently under consideration for being cut.

Today, CATO is complaining about having states measure their dropout rates on a common scale so as to facilitate comparisons. CATO and I agree on the desirability of a highly decentralized education system (with more private provision than at present). However, we seem to disagree on what seems to me the rather obvious empirical point that data are a public good. It seems to me that, in fact, the federal Department of Education should do only two things. It should collect and distribute standardized data for use by parents and by state and local governments and it should undertake serious ("rigorous" to use the preferred term at the Institute for Education Sciences) evaluations of various education reforms, whose results can then be shared throughout the research and policy communities. Evaluation, like data, is a public good and in the classical liberal vision of the state which CATO supposedly shares with me, it is all about public goods, with perhaps some thoughtful redistribution around the edges.

CATO, however, seems to adopt the view one sometimes hears associated with stories about Hong Kong, namely that ignorance is bliss. In this view, if you give people data, they will just want more government. This seems to me both empirically wrong (it was data that did in communism, not ideology) and a fundamentally illiberal (in the classical meaning of that term). Having states produce comparable data allows the public and policymakers to actually learn from cross-state comparisons, and puts the brake on state-level politicians simply making things up in ways that suit their political agenda but do a disservice to the truth. Particularly in the current second- or third- or tenth-best world in which the federal government hands out large amounts of money to the states, the least that can be expected in return is accurate measurement of outcomes (not impacts!) on a common scale.

Not a good month for CATO.

No comments: