Preregistration (part 1): Sampling rule and data split for the study 'An LLM manual for coding teacher talk moves' (TalkMoves)
Year: 2026
Type: article
Abstract: Sampling rule for passages from the TalkMoves dataset (Suresh et al., 2022): development part (Dev-Train), validation part (Dev-Val), aggregator pool and locked test part, assigned by lesson series. The rule was approved and the draw carried out on 2026-10-28 with seed 20261028; the allocation files contain only sentence IDs, no transcript text. Hypotheses, conditions and the analysis plan follow in part 2 before the main runs.
Author: Filler, Simon
Institution: TU Dortmund University
Language: English
Subfield: General Social Sciences
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7224088106