It is possible that these estimates of time on aid are biased if graduates or those who never
enroll in postsecondary programs have higher rates of attrition in the NLSY. This would reduce the
number of months they are in the sample and potentially lead to an underestimate of the number of
welfare months in total. To check this, I examine the number of NLSY interview years in which
each welfare recipient participates. Welfare recipients who attend college were present for an
average of 19.0 of the possible 20 interview years, or 95.0 percent of the potential interview months
covered by that time period. Welfare recipients who do not attend college were similarly present for
an average of 18.6 of the possible 20 interview years and 93.1 percent of the potential interview
months. Comparing graduates to non-graduates, I also find no discernible evidence of bias in
aggregate. Graduates were present for an average of 18.9 of 20 years and in 94.5 percent of potential
interview months. Non-graduates were present for an average of 19.0 years and in 95.2 percent of
potential interview months.