Thursday, August 28, 2008

Getting the evidence

Maybe we've been too hard on Kevin Rudd.

These days everyone is bashing him over the head about introducing policies without an "evidence base". I got in early.

It is true, he did promise to be guided by the evidence.

Here's what he said in November just before the election on The 7.30 Report:

"I'm a Labor moderniser. Always have been, always will be and what that's on about is good evidence based policy in terms of producing the best outcomes for this nation, carving out its future in a pretty uncertain century where things fundamentally are changing."

Rudd is in hot water now over his plan to fight school truancy by stripping welfare payments from parents whose children don't go to school.

"We've yet to see the evidence that this kind of approach will work," says the President of the Council of Social Service Lin Hatfield Dodds.

But here's my thought. Maybe that's what Rudd's plan is about - getting the evidence.

It'll run as a trial in specified suburbs and towns. It's possible to think of the welfare measure as a gigantic government-funded economics experiment using real people, real money and a real control group.

It'll produce evidence alright - on a scale researchers usually only dream of.


Matt C said...

Hi Peter,
I agree that pilot schemes like this provide a great evidentiary basis for a broader program. I'm a fan of the idea of having lots of social sciences 'experiments', and I know Andrew Leigh has advocated that view quite enthusiastically.

However, this particular pilot scheme is being pursued without appropriate mechanisms in place to evaluate the program. It is policy on the run. The federal, State and non-government agencies involved are insufficiently prepared to evaluate the outcomes of the policy, and as such the pilot is unlikely to yield meaningful evidence that can be used to guide its wider implementation.

Andos said...

With such a significant stake in the outcomes of the trial, should the Government be responsible for designing, implementing and analysing the trials?

Surely to get an unbiased outcome that can then be used to formulate effective policy these trials must be designed and run by a dedicated, non-partisan body.

But then that would require a whole rethink of the way Government formulates and implements policy to achieve certain outcomes...

A good feature on this issue (evidence based policy) in New Scientist recently:
(subscriber only)

Post a Comment