GPT-4o prompted only with SRL theory references generates a codebook accurately representing most key elements of both foundational SRL theories¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? A theory-only prompt with full references led GPT-4o to propose constructs capturing all three phases of Zimmerman's model and three of four phases of Winne & Hadwin's model. → Zambrano 2026

Evidence¶

Zambrano 2026¶

Zambrano, A.F., Wei, Z., Zhang, J., Baker, R.S., Ocumpaugh, J., Barany, A., Liu, X., Ginger, J., Paquette, L., Zhou, Y., & Borchers, C. (2026). Data Plus Theory Equals Codebook: Leveraging LLMs for Human-AI Codebook Development. Journal of Educational Data Mining, Volume 18, No 1. https://osf.io/g3z4x doi:10.35542/osf.io/4wyqc_v1

q2 · i? · design · r2

Human evaluation of Step 1 (Theory-only) codebooks generated three times by GPT-4o from references to Winne & Hadwin (1998) and Zimmerman (2000), consolidated by two researchers. The authors report the codebook "accurately represents most of the key elements of both SRL theories", though the task-definition phase was not directly represented.

"the codebook suggested by GPT based on the SRL foundational theory papers accurately represents most of the key elements of both SRL theories, indicating that the model's knowledge base includes appropriate and accurate representations of these theoretical frameworks."

Discussion¶

Related Claims¶

• Adding think-aloud data to theory prompts improves GPT-4o codebook completeness and alignment with the data — related
• In the interest-development context, naming the theory without full references produced the most practical and usable codebook, while supplying full papers enhanced theoretical alignment but reduced applicability — related
• Fully inductive LLM codebook development risks importing unexamined sensitizing concepts, such as folk theories and scientific misconceptions, from the model's training data — related
• Human review remains essential after GPT-based codebook generation, producing a final refined SRL codebook of eleven constructs — related
• Over one-third of AI-SRL studies specify no SRL theory; Zimmerman's model is the most frequently applied — related
• Empirical studies support Zimmerman's cyclical and multi-level SRL models, including a linear trend favoring training more SRL phases — related
• Prompting is a form of writing: task definition is a constitutive element of knowledge production, not a neutral procedural step — related
• Zimmerman's (2000) model was the most frequently cited SRL model (74%) and Pintrich's (2000) model followed (67.4%), with model choice not affecting the areas promoted — related