Monday, February 23, 2009

Do Days in School Matter?


Oregon lawmakers' cavalier attitude toward early school closures suggests that the marginal impact of days in school on student learning is small. But where are the facts? I have yet to hear or read a quote from one single lawmaker about the marginal impact of school closures, so I decided to go find out for myself.

Understanding the marginal impact of a day in school on student performance is complicated by the fact that school closures are not usually random. For example, if we looked at the impact on Oregon students of the 2003 shortened school year we would have trouble disentangling the effects of the short school year with other effects of the economic downturn in the state like decreased provision of social services, increased unemployment and other disruptions that could hamper student performance. Therefore we could not plausibly attribute any decreased performance to school closings alone.

Two researchers in Maryland, however, use a clever strategy to isolate the effects of school instruction days on student performance by looking at weather forced closures of schools - snow days - on assessment scores. The strategy works because tests are administered in April, before the districts can make up snow days by tacking them on at the end. The fact that snow days are random events and are different across Maryland's school districts allows for a plausible counter-factual: the performance of kids not affected by snow days. But this is not all, the researchers use a panel of data - multiple years variation in school days that allow them to control for differences across schools as well. In other words, they are looking at year over year differences in a specific school's performance based on snow days.

The results are presented in this paper (and some similar results by one of the authors is presented in another related paper). Unfortunately, the measures are a bit crude: they see only the percent of children that pass or fail statewide assessment tests in math and reading. They find that for each lost day of school, the pass rate for 3rd grade math and reading assessment tests fall by more than half a percent. So a loss of five school days (like what is being contemplated in Oregon) would lead to a three percent increase in the failure rate. This is shocking in and of itself, but consider the reduced performance for all students, including those well above the failure line and the implications for the overall performance loss - the total loss of learning - is truly profound.

Now there is a caveat: schools that know they are loosing days can strategize to maximize learning at the end of school, but remember that Oregon already has one of the shortest school years in the nation (and in the industrialized world) - so we are already trying to squeeze as much learning out of the short school year as it is, it is therefore not clear to me that more can be squeezed out of an already dry stone.

It is simply astounding to me that the Governor can talk about positioning Oregon to become a leader in green technology while at the same time disinvesting in Oregon's human capital. I have news for him - Oregon will be a leader in green technology not because of a few selected tax breaks, but because we lead in technological innovation. And we will only lead in technological innovation if we have smart people and excellent schools and universities, this is the lesson of Silicon Valley.

It is time to have a real discussion of the costs to the state of expediency in budget cuts. The obvious alternative to a shorter school year is increased class size. So what is the impact of class size? There is some debate: Hoxby (2000) finds no measurable effect, Hanushek (1998) agrees, but Rivkin, Hanushek and Kain (2005) (in a much better test) find positive, but quite small increases in student performance from reduced class size. The conclusion is inescapable - a shorter school year imposes a much greater cost on Oregon's children than increased class size.

It is time for the Governor and lawmakers in Salem to wake up and start making smart, informed decisions rather than off-the-cuff ones.


Mike said...

I interpret the Governors proposal as a sign that he considers education a low priority. Granted, he has many difficult decisions to make given our terrible economic situation, but our K-12 system can't really take another substantial cut.

The Chalkboard Project has a paper on the benefits of extended learning (i.e., add more learning time!).

I personally find it astounding how low a priority K-12 education is in Oregon. It's sad and will likely come back to haunt us.

Hall Monitor said...

Check out for all the other problems facing our state of education.