The illustration below is related to the illustration in the Concentration Index sub-page of the Measuring Health Disparities page.It is intended to show that the Gini Coefficient is not a useful measure of a disparity because, like other measures (including the Concentration Index), the Gini Coefficient is affected by the overall prevalence of an outcome.
Table 1 below is based on the data in Table 1 of the BSPS 2006 presentation,[i] which show failure and success rates for advantaged (AG) and disadvantaged (DG) groups at various points defined by the AG failure rate.The table shows not only that the Gini Coefficient changes as the prevalence of an outcome changes, but that that coefficients based on the favorable and the adverse outcome change in the opposite directions.
Table 1:Gini Coefficients for Falling Below and Above Each Point Defined by the Advantaged Group’s Failure Rate in Table 1 of BSPS 2006.
8512 e 5 limit gini to key points
Ref Point
AGFailRate
DGFailRate
FailGini
PassGini
99%
98.98%
99.76%
0.00
0.31
97%
97.00%
99.13%
0.01
0.28
95%
94.95%
98.38%
0.01
0.26
90%
89.97%
96.25%
0.02
0.23
80%
79.96%
90.99%
0.03
0.19
70%
69.85%
84.61%
0.05
0.16
60%
59.87%
77.34%
0.06
0.14
50%
50.00%
68.79%
0.08
0.12
50%
50.00%
69.15%
0.08
0.12
40%
40.13%
59.48%
0.10
0.10
30%
30.15%
49.20%
0.12
0.08
20%
20.05%
36.69%
0.15
0.06
10%
10.03%
21.77%
0.18
0.03
5%
5.05%
12.71%
0.22
0.02
3%
3.01%
8.38%
0.24
0.01
1.00%
1.02%
3.44%
0.27
0.01
[i]The Misinterpretation of Health Inequalities in the United Kingdom, presented at the British Society for Populations Studies Conference 2006, Southampton, England, Sept. 18-20, 2006: http://www.jpscanlan.com/images/BSPS_2006_Complete_Paper.pdf.
[ii]Comparing the size of inequalities in dichotomous measures in light of the standard correlations between such measures and the prevalence of an outcome.Journal Review Jan. 14, 2008 (responding to Boström G, Rosén M.Measuring social inequalities in health – politics or science?Scan J Public Health 2003;31:211-215):