This study investigated the psychometric properties of the modified checklist for autism in toddlers (M-CHAT) in a diverse rural American low-socioeconomic status (SES) sample. Four hundred and forty-seven English (n = 335) and Spanish (n = 112) speaking caregivers completed the M-CHAT during their toddler's 18- or 24-month well visit in a Southwest Virginia pediatric clinic. The M-CHAT did not show acceptable internal consistency in groups with low maternal education or minority status. Caregivers reporting low maternal education and minority status were more likely to endorse items suggestive of autism. These results indicate that the M-CHAT may require modifications to be more internally consistent and accurate across ethnic and educational groups in rural areas with low levels of SES. Recommendations to increase the utility of the M-CHAT are discussed.