Title:Optimal Academic Plan Derived from Articulation Agreements: A Preliminary Experiment on Human-Generated and (Hypothetical) Algorithm-Generated Academic Plans

Abstract:Objective: Community college students typically submit transfer applications to multiple universities. However, each university may have differing lower-division major requirements in order to transfer. Accordingly, our study examined one pain point users may have with ASSIST, which is California's official statewide database of articulation agreements. That pain point is cross-referencing multiple articulation agreements to manually develop an optimal academic plan. Optimal is defined as the minimal set of community college courses that satisfy all the transfer requirements for the multiple universities a student is preparing to apply to. Methods: To address that pain point, we designed a low-fidelity prototype that lists the minimal set of community college courses that a hypothetical optimization algorithm would output based on the user's selected articulation agreements. 24 California college students were tasked with creating an optimal academic plan using either ASSIST (which requires manual optimization) or the optimization prototype (which already provides the minimal set of classes). Results: Experiment participants assigned to use the prototype had less optimality mistakes in their academic plan, were faster in creating their plan, and provided higher usability ratings compared to the ASSIST users. All differences were statistically significant (p < 0.05) and had large effect sizes (d > 0.8). Conclusions: Our preliminary experiment suggests manually developing optimal academic plans can be error prone and that algorithm-generated academic plans can potentially reduce unnecessary excess community college credits. However, future research needs to move beyond our proof of value of a hypothetical optimization algorithm and towards actually implementing an algorithm.