To assess differences, by health literacy status and behavioral condition, in participants' abilities to self-monitor behaviors accurately and recall key behavioral messages using data from a teach-back call.Cross-sectional.Rural, southwestern Virginia.Adults (n = 301). The majority were female (81.1%), 31.9% had a high school education or less, 66.1% earned < $25,000/y, and 32.9% were low health literate.First class session of 2 community-based behavioral interventions: SIPsmartER (reduce sugar-sweetened beverage intake) or MoveMore (increase physical activity).Reported accuracy of behavioral diary completion, proportion of behavioral messages recalled during the first round of teach-back, and rounds of teach-back.Descriptive statistics and generalized linear model.Low health literate participants were significantly less accurate in diary completion (P < .001), recalled fewer behavioral messages correctly (P < .001), and needed more rounds of teach-back (P < .001) than high health literate participants. Compared with SIPsmartER participants, MoveMore participants more accurately completed diaries (P = .001) but recalled a lower proportion of behavioral messages correctly (P < .001) and required more rounds of teach-back (P < .001).Health literacy status and behavioral target affect the ability to self-monitor and recall key concepts. Researchers should consider using teach-back early in the intervention to assess and reinforce participants' ability to self-monitor.