Twenty-four-hour dietary recalls are used frequently to study homebound older adults' eating behaviors. However, the reliability and predictive validity of this method have not been established in this population. The purpose of this study was to examine whether homebound older adults provide reliable and valid measures of total energy intake in 24-hour dietary recalls. Two hundred thirty homebound older adults were interviewed in their homes using a questionnaire to assess eating behaviors and factors that could affect those behaviors. Participants completed three 24-hour dietary recalls at baseline and again at 6-month follow-up. Two subsamples were identified for analyses. For participants who were not hospitalized during the 6-month interval and had their weight measured at both assessments (n=52), sufficient test-retest reliability of energy intake was observed (r=0.59), but energy intake deficiencies relative to estimated energy requirements did not predict actual weight loss (r=0.08). When this sample was supplemented with 91 participants who experienced any adverse event (weight loss of 2.5% or more, hospitalization, institutionalization, or mortality) in the 6-month period (n=143), adverse events were more likely to occur for those with insufficient energy intake (odds ratio 3.49, P=0.009), and in white participants compared to African-American participants (odds ratio 3.13, P=0.016). Adequate test-retest reliability of the 24-hour dietary recall was demonstrated, but additional research with larger samples and longer follow-up intervals is needed to better evaluate the predictive validity of energy intake measures for this population. © 2010 American Dietetic Association.