Coherent mental models for successful comprehension require inferences that establish semantic "bridges" between discourse constituents and "elaborations" that incorporate relevant background knowledge. While it is established that individual differences in the extent to which postsecondary students engage in these processes are correlated with reading outcomes, there is little research exploring if first-year students and first-year students enrolled in developmental education (DE) differ in the extent that they engage in these processes. In this manuscript, authors report on the implementation of the Reading Strategy Assessment Tool (RSAT) with first-year students and first-year students enrolled in DE employing a think aloud protocol. RSAT is a computer-based system that collects typed "think aloud" protocols and analyses of the protocols for inference processes that support mental model construction. In this study, RSAT scoring was compared to human coding of the protocols. There was convergence between both approaches for scoring the protocols. Controlling for comprehension proficiency, both groups bridged to a comparable extent. There was evidence of less elaborative processes in the first-year students enrolled in DE, suggesting differences in relevant background knowledge across both groups. This study illustrates the utility of RSAT in the study of underprepared college students. [This paper will be published in "Journal of College Literacy and Learning" v46 May 2020.]