Income Disparity and the Mississippi River

a clear association between being a Mormon (Utah has the second lowest state mortality rate) and lower mortality rates, it does not follow that if I compel people to become Mormons, mortality rates will fall.

Critique of the Study

Despite careful efforts by the authors, there are some glaring problems with the Harvard study.

1. In order to estimate the statistical significance — the measure of whether any association is random or real — the authors assumed that the values for all the states are statistically independent. The clustering of mortality and income disparity values of geographically adjacent areas like North and South Dakota, Minnesota, Iowa, Kansas and Colorado suggests that each state is not statistically independent. The quoted statistical significance is probably far lower than quoted.

2. Although, the authors attempted to correct for absolute poverty, they assumed a linear relationship between poverty and mortality. Any errors in estimating this relationship or any remaining nonlinearity will leak in the residuals and increase the apparent correlation with income disparity.

3. The two states that are most conspicuous in this study are Mississippi and Louisiana, with both the highest income disparity and the highest mortality. Those states have a disproportionate level of rural rather than urban poverty. This generally results in less access to medical care than in urban areas. We may be seeing the effects of different types of poverty rather than income disparity. Moreover, these two states have a far higher proportion of chemical and petrochemical plants. This may partially account for the increase in mortality for these two states.

4. The authors were careful to correct for age in their mortality figures. However, they did not correct for age in their income disparity statistics. If some portion of income disparity is associated with a wide age distribution (older people tend to earn more than younger people) the analysis is further muddied.

Conclusion

The real problem is not the particular study by Kennedy et al. It is an important and intellectually honest effort to address a difficult question. Although significant, the critiques listed above are difficult to address with available data.

The real problem is not willful or deliberate deceit but rather the sloppiness of passionate advocates in a desperate search for "scientific" evidence to support their views. The more people cavalierly cite academic studies, the more the public will believe that scientific studies can not be objective. When citations do not ring true,

Go To Page: 1 2 3 4

Articles in this Topic    Discussions in this Topic