This week at a stats meeting Stephen Haslett described advising small area measures against SDGs in a scattering of (non-OECD) countries in our region, and some nuances (similar from my recollection) to what Martin Travallion had spoken of with WB cross country poverty alleviation measures. This leads into vulnerability index territory - mapping regions within countries for aid allocation. But given the stark segmenting in Australia: all progress in top deciles, standing still in lower over the 2000-2010 decade, maybe a finer level examination is needed, What should progress look like?
Stephen, that project didn't really look at regional inequality. I use the Gini coefficient for Australia as a whole and for States, but not sure what is best for small areas. Rob Bray might advise.
Peter
Hi, Stephen,
Elena Delavega at the University of Memphis has developed some interesting and creative means of analyzing poverty measures. You might want to check out her work at http://www.memphis.edu/socialwork/research/poverty_fact_sheets.php
Dear Stepen, although I am not Peter.
You may look at these before Gini Coefficient. (i.e. Coefficient of Variation, Quartile coefficient of dispersion. Concern should also be to the "relative mean difference" which must be equal to twice the Gini coefficient.
thank you Atsul, Of course there are good reasons for using the Gini but I am never sure whether its statistical properties are taken into account. Income distributions are notoriously skewed and subject to outlier effects. The Gini is interpreted as an ordering measure on distributions that may in the event not be well defined. Further the dispersion of the Gini as an estimate is not often mentioned.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling