Development and Validation of A Self-Regulated Learn-Ing Instrument For Distance Education In The Brazilian Context (eaa-Ead-Br): Integration of Computational Psychometrics and Traditional Methods

Distance Education in Brazil accounts for 50.75% of undergraduate enroll-ments (3.9 million students), yet it maintains high dropout rates (63.7%). Self-Regulated Learning (SRL) is recognized as a predictor of academic suc-cess; however, existing instruments lack specificity for Distance Education contexts and have not been validated in Brazil. This doctoral research pro-poses a hybrid methodology integrating a systematic literature review, gen-erative psychometrics (AI-GENIE, DynEGA), and traditional psychometric validation to develop a Self-Regulated Learning Scale for Distance Education in Brazil (EAA-EAD-Br). The study will systematically synthesize 200–300 validated items from the literature, generate three comparative item pools (literature-based, AI-generated, and hybrid) using large language models (GPT-4o, Gemma-2), and optimize semantic embeddings via Dynamic Ex-ploratory Graph Analysis. Empirical validation will apply Exploratory Factor Analysis (EFA) and Exploratory Graph Analysis (EGA) in parallel (N ≥ 200) for structural identification, followed by Confirmatory Factor Analysis (CFA) (N ≥ 500) to confirm a four-dimensional structure (metacognitive, motiva-tional, behavioral, contextual-technological), including invariance testing. Expected contributions include: the first SRL instrument for Distance Educa-tion validated in Brazil; empirical validation of the contextual-technological dimension; demonstration of convergence between in silico and empirical validation (r ≥ 0.70); and the establishment of a replicable methodological paradigm enabling the application of computational psychometrics in educa-tional technology