Monday, June 30, 2014

The sample sizes, they are a changing

From the BIT's blog:

"...the real take home message is not the specifics of the trial, but the fact that the trial is happening. This class, and this school, is one of thousands now taking part in controlled trials across the UK. There are now an incredible 2,400 schools, taking part in 87 different trials (74 or which are RCTs). That’s around 500,000 pupils involved, organised and funded by the independent Education Endowment Fund."

Fairly incredible when you think of past research which was important in terms of policy implications but which rested on experiments with much smaller sample sizes, such as the Perry Preschool Project (N=123 children) or Card & Kreuger's minimum wage natural experiment (N=410 restaurants). Support for evidence-based policy on this scale will mean more certainty about effect sizes and probably obviate the need for a lot of complex econometrics.

1 comment:

Liam Delaney said...

"... obviate the need for a lot of complex econometrics". Afraid not. Will increase the types of parameters that can be estimated which is good news but still a lot of work to do to get them.