Proceedings of the Annual Meeting of the Cognitive Science Society

Natural language processing (NLP) provides an innovative avenue to understand and explore human language content,yet minimal research has utilized it to classify students literacy, cognition, or motivation. This study investigated theassociation between grade 4-6 students (n = 143) writing and their cognitive-motivational profiles (CMPs) based on theirself-regulated learning, locus of control, writing self-efficacy, and goal-orientation. LPA (Mplus 7.4) results indicated atwo-class CMP solution with predominantly positive or negative CMPs. Using NLP, 404 lexico-syntactic writing featureswere extracted from students writing. Random forest with 10-fold cross-validation was implemented in Weka 3.8 (withSMOTE to equate class instances) to accurately (93%) classify students CMPs (class 1 True Positive Rate (TPR) = .942;class 2 TPR = .925) based on the NLP-processed lexico-syntactic writing features. These results highlight the potentialfor machine learning to analyze students writing and accurately classify learner profiles to provide formative feedback andcustomized interventions.