Thursday, March 08, 2007

Aborigines and the Development of Public Policy - a Methodological Note

Neil, Ninglun, has had some interesting posts on his blog. I am struggling with time at present and have not contributed to Neil's dialogue in the way I should.

In one post, Neil said in part "I also think the sterile impasse between our two brands of denialist, left and right, has to be broken if we are to move forward on indigenous issues."

This quote captures what I am trying to do in my current series on the Aborigines. It's not easy.

At the moment I am bogged down in what statistics there are about Aboriginal conditions in NSW, trying to look at patterns across the state.

Part of my argument in this series has been that we need to recognise and understand variations in Aboriginal conditions across space, not just talk in averages, state or national. I am also concerned about the way in which particular issues come to twist and dominate the debate.

I am absolutely confident from my own experience that both these positions are correct. However, to show this to others I have to have evidence, not rely on assertion.

I think that the available evidence, while not good in statistical terms, does support my position. However, we then have to drill down further. Okay, so there are local and regional variations, what does this mean? To try to illustrate by example.

If we look at the regions in NSW as defined by the NSW Department of Aboriginal Affairs, we find some common patterns but also significant variations. Why, for example, is the pattern of pre-natal visits, one indicator of child care, higher in New England-North West than in many other parts of NSW, why does the incidence of imprisonment appear lower when on some other measures the Aborigines in New England North West are some of the worst off in the state?

What do all these differences mean? The regional stats such as they are are themselves averages. Has the Aboriginal experience in Armidale been the same as in Moree? Based on qualitative evidence I don't think so. Yes, there are commonalities, but there also appear to be considerable differences.

The problems associated with data imperfections increase as we drill down.

Consider measures of Aboriginal unemployment. The NSW Department is forced to rely on 2001 census data. This is too old to be a reliable indicator of today's conditions. What has happened since? Why? What does it all mean?

Again taking Armidale as an example.

Based on anecdotal evidence, I suspect that the 2006 census will show a major increase in Armidale's Aboriginal population because of in-migration. I also suspect that the data will show an increase in the rate of Aboriginal unemployment in Armidale because, and this is another assumption, the new arrivals have less education and less access to existing networks than those already there.

If all this is correct, then indicators drawn from the 2006 census may show a deterioration in the social condition of Armidale's Aborigines, but one due to in-migration.

In policy terms, these types of problems are usually handled by what I call point and counterpoint, constant switches between the general and the particular.

The aggregate data suggests x. Test this by looking at specific examples. Look at specific examples, what does this suggest about broader issues? How do we test this against the stats.

In all this, one of the core needs is to properly define the problem to be addressed. In my view, misspecification of problems is the single most common cause of policy failure.

Here one question that I constantly ask is is the issue in question an Aboriginal problem or a problem for our Aborigines?

By this I simply mean is the problem unique in some way to Aborigines (an Aboriginal problem) or one shared in some ways with other groups in the Australian community (a problem for our Aborigines)? If the second, are there aspects to the problem that are specific to the Aboriginal community?

Take the point that I have been making constantly about the failure to properly address economic development issues in New England.

This is a problem shared by all, so it is a problem for New England's Aborigines, but has to be addressed by broader policy measures. On the other hand, there are aspects to the problem such as poor Aboriginal education that may need specific corrective measures as well.

No comments: