This study looks at the relationship between food stamp participation and historical earnings over periods of 10-15 years. Earlier research found that households eligible for the Food Stamp Program that had short-term income declines were less likely to participate than those that had sustained low incomes. This analysis expands on that research by using a data set that matched historical Social Security earnings records to the 1996 Survey of Income and Program Participation, allowing examination of the relationship between participation and earnings over a longer timeframe than available previously. The results show some evidence that historical annual earnings as far back as 5 years earlier are negatively and significantly associated with households’ decisions to participate in the Food Stamp Program; that future earnings, which may proxy for earnings expectations, are also negatively and significantly associated with participation; and that monthly income volatility plays an important role. However, because of weaknesses in the specification of the regression models, findings in this paper are suggestive rather than precise descriptions of the relationship between longer run income and participation.