The relative effectiveness of the US Food Stamp Program in reducing th
e incidence and depth of poverty is compared at a number of poverty li
nes across time. The analysis is based upon an extension of dominance
methods for ordering poverty, recently developed statistical inference
procedures, and a merged data-set that permits an unusually comprehen
sive measure of income. The comparative effectiveness of the Food Stam
p Program in reducing headcount poverty over time is contingent on the
poverty line selected. However, the reduction in poverty gaps is grea
ter in 1990 than in 1982, regardless of the poverty line chosen. The s
ensitivity of the results to the income concept, equivalence scale and
time period used for measuring income are also considered.