Purpose. The association between short sleep and obesity risk is well established. However, we explore a new pathway between short sleep and obesity: whether short sleep is linked to more time spent in secondary eating or drinking, that is, eating or drinking (beverages other than water, such as sugar-sweetened beverages) while primarily engaged in another activity, such as television watching. Design. This pooled cross-sectional study uses data from the American Time Use Survey (ATUS) from 2006 to 2008. Setting. The study takes place in the United States. Subjects. Subjects are 28,150 adults (55.8% female) aged 21 to 65 who were surveyed in the ATUS. Measures. Outcomes are time spent on (1) secondary eating and drinking and (2) primary eating and drinking. Our main predictor variable is sleep duration. Analysis. Controlling for demographic and socioeconomic characteristics of the respondents, we estimate multivariate regression-analysis models for the full sample, as well as by weekday/weekend status, race, and gender subgroups. Results. In multivariate models, compared to respondents reporting normal sleep, short sleep was associated with additional 8.7 (SE = 2.1) minutes per day of secondary eating (p <.01) and additional 28.6 (SE = 4.2) and 31.28 (SE = 5.0) minutes per day of secondary drinking on weekdays and weekends, respectively (p <.01). Conclusions. We find that short sleep is associated with more time spent in secondary eating and, in particular, secondary drinking. This potentially suggests a pathway from short sleep to increased caloric intake in the form of beverages and distracted eating and thus potential increased obesity risk, although more research is needed.