Friday, January 15, 2010

What has $100 billion bought us?

Apparently, very little according to Cato's Andrew Coulson and his review of an HHS study on the impact of Head Start.

HHS has finally released the second installment of its series of studies on the persistence of Head Start effects. Its finding (see page xiv): virtually all academic effects disappear by the end of 1st grade. There is only one positive statistically significant finding out of eleven academic outcomes measured, the size of that effect is minuscule by recognized standards (it’s half way between zero and what most social scientists consider “small”), and the confidence in the finding is low by recognized standards. (Many authors would categorize it as “insignificant” rather than “significant” — it’s only significant at a 90% confidence interval, not the more common 95% confidence interval).

We have spent more than $100 billion on the program to date (ballpark estimate from Table 375 here) and HHS’s own research shows that its results diminish to essentially nothing by the end of the first grade.
(emphasis added)



Tim Higgins said...


It should be of no surprise to anyone that a government program has no measurable benefit to those it serves. The surprise would be if as a result of the release of this data, the program and its funding were canceled.

While we're at it though and in these days of constant complaint against the funding of our efforts in Iraq and Afghanistan, how's that "War on Poverty" going?

Roman said...

It is not important to liberals that a good idea works. They just have to have the right intentions.

...if only the right people would try this, it would succeed, if we would just throw more money at it, etc, there would be success.

Too much, "been there, done that" is ignored. Many people cannot bring themselves to believe that their good, even great, ideas might not work in practice.

A good friend of mine, while testing a new part on his drag race car said: "If you have a theory, test it and it fails the test, believe the test and get a new theory."

Google Analytics Alternative