Education not only impacts expected future earnings, it may also impact income risk. This pilot will examine the impact of education on income volatility. When using standard cross-sectional data sets to estimate the impact of education on earnings, it is difficult to differentiate risk from heterogeneity. This paper overcomes this problem by exploiting the panel feature of the Panel Study of Income Dynamics (PSID) to estimate income volatility directly for many individuals. Differences in earnings volatility across people may simply result from unobserved heterogeneity and not the causal impact of education on income volatility. To address this problem, I plan to collect data on the location of colleges in order to construct a measure of a person's distance to a college. I will use this measure as an instrument for college attendance.