Qualitative Coding_HigherEducationPolicies
Description
Purpose: The dataset supports a content analysis paper that maps how higher education institutions have responded through published policy to the emergence of generative AI since late 2023. The paper is under consideration for the 2nd National Conference on Pedagogy for Higher Education (NCPHE-2026). By publishing the underlying coded corpus openly, this dataset enables (a) verification of the paper's empirical claims, (b) methodological transparency about the inductive coding process, and (c) reuse of the codebook or corpus in subsequent studies. Method of construction: Institutions were identified through purposive convenience sampling on the criterion that they had published an AI policy in the public domain, in English (as original text or as officially translated by the institution). Each policy document was read in full, and clauses expressing a stance on AI use, a locus of authority, a condition of use, or a wider policy dimension were extracted verbatim. Coding was conducted inductively: initial codes were generated with the assistance of the Anthropic Claude LLM (Claude.ai) and then reviewed, consolidated, and systematised by the author into a codebook of 22 child codes organised under four parent categories. Every code is anchored to a specific text span within its clause; the workbook presents these spans in colour-coded form so that each coding decision is inspectable and independently challengeable.
Files
Steps to reproduce
ID — clause identifier (e.g., C1, CL2, T6) Verbatim (highlighted where coded) — the original policy text, with coded spans visually highlighted Child Codes — the codebook categories assigned to each span Source Institution — the institution that published the policy Document Title — full title of the source policy document URL — direct link to the published policy document Section within Document — the specific section or subsection of the source document Access Date — the month in which the policy document was accessed