The current study analyzes variable subject pronoun expression (SPE) for first-person singular (1sg) and third-person subjects in a variety of Mexican Spanish spoken by first-generation Mexican immigrants in the state of Georgia, Southeastern U.S. Conversational data from sociolinguistic interviews are employed to examine tokens of 1sg and third-person variable SPE and their usage patterns, considering factors such as tense-mood-aspect (TMA), switch reference, polarity, and verb class by means of logistic regression analyses. Results suggest that all four factors influence 1sg variation, but that third-person variation is restricted to switch reference and TMA. In addition, a significant link between switch reference and TMA is found for third-person subjects, but not for 1sg. The findings lend further support to previous scholars advocating the importance of studying individual grammatical persons in SPE research as this can reveal previously obfuscated nuances in the patterns of subject variation.