Real-time assessment of users' cognitive states has practical importance, allowing organizations to infer user behaviors. Realizing its importance, prior studies - specifically those using mouse cursor movements - have applied various theories to answer a similar question, i.e., how does a high cognitive load influence the users' device usage behavior? While numerous activities can increase cognitive load, we argue that the mechanisms behind how humans process information can more holistically be explained using Dual Process Theory (DPT) (i.e., when cognitive load is either low or high) and can be applied under a broad range of usage contexts. Using a within-participant experiment and a simple typing task, we demonstrate that DPT is robust to work by examining DPT and mouse cursor movements. Specifically, users' typing speed and task execution are significantly slower when engaged in the task (System 2) and significantly faster when completing the task with lower cognitive effort and engagement (System 1).