The Human Suffering Index was created by the Population Crisis Committee (Washington, D.C.) to compare living conditions among countries. Table 8.11 shows results of this index in 1987 for countries classified by their continent. Though inference is not relevant here, one could use a model to describe how suffering varies by region. Treating region as a nominal predictor, attempt to fit a proportional odds model, using North America (NA) as the baseline category for the dummy variables. The region estimates are infinite, because all observations for NA fall in an extreme category of the response. To obtain finite estimates, add

(a) .001,

(b) .01,

(c) .1, to the moderate category for NA. Note that the estimates for odds ratio comparisons of other regions with NA depend strongly on the choice of added constant, but the comparisons for other pairs of regions are similar for each such choice and similar to those obtained by fitting the model to all but the NA data. Use the parameter estimates from

(a) to rank the regions on degree of human suffering and to compare Africa and Asia on this index.

Transcribed Image Text:

Table 8.11 Region Africa Asia Europe Latin America North America Human Suffering Index Minimal Moderate High Extreme 0 1 16 24 2 10 16 20 7 0 0 1 2 10 12 0 0 0 0 Source: H. F. Spirer and L. Spirer, Data Analysis for Monitoring Human Rights. 1993, American Association for the Advancement of Science. Reprinted with permission.