• quantitative methods;
  • variation;
  • Spanish;
  • subject expression;
  • advanced learners

This study constitutes the first statistical analysis to employ a Bayesian multinomial probit model in the investigation of subject expression in first and second language (L2) Spanish. The study analyzes the use of third-person subject-expression forms and demonstrates that the following variables are important for subject expression: perseveration, switch reference, number of the verb, specificity, verbal tense/mood/aspect, object pronoun, referent cohesiveness, the interaction of switch reference and referent cohesiveness, and the interaction of native language and four independent variables (number, specificity, tense/mood/aspect, and object pronoun). However, only certain parameters of these independent variables predicted use. The analysis highlights that, for advanced L2 speakers and native speakers, all forms of subject expression (i.e., lexical noun phrases, null subjects, personal pronouns, and other pronouns) allow variation and should be examined for a thorough understanding of subject expression in Spanish. The study offers theoretical and empirical evidence for the need to conduct cross-disciplinary research in second language acquisition.