High Heterogeneity in AI Anxiety Research: Meta-Analytic and Bibliometric Evidence Consistent with a Pre-paradigmatic Stage

Description

This deposit contains all materials required to reproduce the analyses reported in the manuscript "High Heterogeneity in AI Anxiety Research: Meta-Analytic and Bibliometric Evidence Consistent with a Pre-paradigmatic Stage" (Frontiers in Psychology, Manuscript ID 1924329). Contents of the upload pack: (1) Data.xlsx — the study-level meta-analytic dataset (11 studies, N = 6,156). Effect sizes are standardized deviations of sample means from scale midpoints (Hedges' g), computed with the one-sample sampling-variance formula. Sheets: Study Data, Summary, Subgroup, Leave-One-Out. (2) R Code.R — the complete meta-analysis code (version 10.1), implementing DerSimonian–Laird random-effects pooling with REML and Hartung–Knapp robustness checks, subgroup tests, exploratory meta-regression, PET-PEESE publication-bias assessment, influence diagnostics, and leave-one-out analyses. (3) Bibliometric Analysis Code.py — the bibliometric pipeline (term extraction with a concept dictionary, NetworkX co-occurrence network, Louvain clustering, Kleinberg two-state burst detection). (4) Bibliometric Document List.xlsx — the manually rescreened bibliometric corpus (n = 337 documents) with per-record screening outcomes.

Steps to reproduce

1. Meta-analysis: open "R Code.R" (version 10.1) in R with the metafor package installed and run the script top to bottom. It reproduces all meta-analytic results reported in the manuscript and Supplementary Materials: the overall pooled effect, heterogeneity statistics (Q, I-squared, tau-squared), prediction interval, subgroup comparisons, meta-regression, PET-PEESE, influence diagnostics, and leave-one-out analyses. The script reads the embedded study-level vectors that exactly match "Data.xlsx". 2. Bibliometric analysis: run "Bibliometric Analysis Code.py" with Python 3 (requires pandas and networkx). It reproduces the keyword co-occurrence network, thematic evolution analysis, Louvain clustering, and Kleinberg burst detection from the 337-document corpus in "Bibliometric Document List.xlsx".

Institutions

• Guangzhou Huashang College

  Guangdong, Guangzhou

Categories

Licence