It is not currently known how dietary assessment in older persons is affected by cognition. In a 1997-2000 study, the authors assessed the validity and reproducibility of a modified Harvard self-administered food frequency questionnaire (SFFQ) among 118 Black and 114 White randomly selected participants, aged 68-99 years, of the Chicago Health and Aging Project. Participants completed multiple 24-hour dietary recall interviews (mean=3.6) over 12 months and two SFFQs in the first and 12th months. The average energy-adjusted intraclass correlation coefficient for 15 nutrients was 0.59 for 1-year reproducibility in nutrient intake levels assessed by the SFFQ. The average energy-adjusted Pearson correlation coefficient was 0.46 for comparative validity between nutrient intake levels on the SFFQ and the dietary recalls. SFFQ reproducibility was higher among men, and comparative validity with the dietary recalls was higher among women. There were no remarkable differences in the correlations by age, race, educational level, presence of chronic conditions, or cognitive ability. The modified Harvard SFFQ is a reasonable method of dietary assessment even in a population of older persons, some of whom are at advanced age, have chronic health conditions, and have cognitive impairment.