| The argument being made by some researchers is that drawing the poverty threshold relatively at fifty percent of the median household income for a given metropolitan area performs well in identifying areas suffering from the ill effects of high poverty concentration that are overlooked by the absolute federal measure. This study will comprehensively examine census tract data for 330 metropolitan areas in the United States to correlate census-tract level poverty concentration to nine separate indicators of social isolation and disadvantage using the absolute federal poverty index, the aforementioned relative poverty index, and an experimental expanded absolute poverty index. The intent of the study is to determine if the relative or expanded absolute poverty index clearly outperforms the incumbent federal poverty index in place today.;"Performance" in the context of this study, is measured by the strength of correlation between that particular poverty index and the instance of the negative outcomes that are associated in the literature with life in highly poor neighborhoods, as well as the percentage explained (variance) at the census tract level. This study posits that a poverty measure would be better suited for use in setting and evaluating poverty policy if it were shown to be more highly correlated (Pearson r) with the negative outcomes associated with poverty than the other measures, and if it is shown to explain a significantly larger percentage of the variance seen in these tract-level outcome data (r 2).;This study explores in detail the methodological and statistical pitfalls associated with all three poverty measures, with the intent on identifying the poverty measure that is most methodologically sound. |
