This study provides welfare evaluations of income distributions based upon several distinct dimensions of distribution functions. By jointly considering poverty, inequality, and the level of income, the methodology avoids controversial judgments concerning trade-offs among these dimensions. Microdata from the 1980 decennial census of population are used to compare individual state income distribution functions to the U.S. average distribution. Recently developed statistical inference procedures that do not require a priori specification of a parametric form for the income distribution function are used to simultaneously test for a higher per capita income, less Lorenzian income inequality, and larger absolute incomes for the poor.