MASEMiner – AI-Assisted Data Extraction for Meta-Analytic Structural Equation Modeling
Abstract
Meta-analytic structural equation modeling (MASEM) synthesizes multivariate evidence across studies, allowing researchers to examine theory-driven path models, including mediated and predictive effects, and to evaluate the dimensional structure of psychological and behavioral measures. A major bottleneck is the labor-intensive extraction of statistical information from primary studies. Large language models and vision-language models offer new opportunities for extracting such heterogeneous statistics from textual and visual document content. Building on these capabilities, we introduce MASEMiner, an open-source application for AI-assisted extraction and verification of data required for MASEM that can be run locally or accessed through a hosted web interface. MASEMiner supports the extraction of directly reported correlations and related effect sizes as well as indirect information (i.e., factor loadings and factor correlations) used to reconstruct correlation matrices for MASEM. We evaluated MASEMiner in four use cases against human-coded reference databases from published meta-analyses, using separate study sets for prompt development and evaluation. Factor loadings were extracted with near-perfect accuracy (sensitivity ≥ 0.92, precision ≥ 0.98), whereas directly reported effect sizes were recovered completely (sensitivity = 1.00) but with lower precision (0.54–0.69) due to overextraction. In a verification study with four coders, automated extraction plus human verification reduced coding time on average by 42% relative to manual extraction at comparable accuracy. MASEMiner thus supports a human-in-the-loop workflow that reduces manual coding effort while preserving transparency and reproducibility in research synthesis.
Related
- Structural equation modelling--based meta-analytic model for diagnostic test accuracy studies
- Null Controllability of a Stochastic Parabolic Equation with a Space-Dependent Analytic Noise Coefficient
- Exact Analytic Solution for the Time-Fractional Hunter-Saxton Equation with Caputo derivative
- Primitive norm extraction and a signed descent for the cubic Fermat equation
- Dynamic Structural Equation Modeling in Blimp
BibTeX
@misc{schroeders2026maseminer,
title = {{MASEMiner – AI-Assisted Data Extraction for Meta-Analytic Structural Equation Modeling}},
author = {Ulrich Schroeders and Timo Gnambs and Johanna Einsiedler},
howpublished = {PsyArXiv},
year = {2026},
month = oct,
doi = {10.31234/osf.io/dceqs_v2},
url = {https://osf.io/preprints/psyarxiv/dceqs_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - MASEMiner – AI-Assisted Data Extraction for Meta-Analytic Structural Equation Modeling
AU - Ulrich Schroeders
AU - Timo Gnambs
AU - Johanna Einsiedler
PY - 2026
DA - 2026/10/25
PB - PsyArXiv
DO - 10.31234/osf.io/dceqs_v2
UR - https://osf.io/preprints/psyarxiv/dceqs_v2/
AB - Meta-analytic structural equation modeling (MASEM) synthesizes multivariate evidence across studies, allowing researchers to examine theory-driven path models, including mediated and predictive effects, and to evaluate the dimensional structure of psychological and behavioral measures. A major bottleneck is the labor-intensive extraction of statistical information from primary studies. Large language models and vision-language models offer new opportunities for extracting such heterogeneous statistics from textual and visual document content. Building on these capabilities, we introduce MASEMiner, an open-source application for AI-assisted extraction and verification of data required for MASEM that can be run locally or accessed through a hosted web interface. MASEMiner supports the extraction of directly reported correlations and related effect sizes as well as indirect information (i.e., factor loadings and factor correlations) used to reconstruct correlation matrices for MASEM. We evaluated MASEMiner in four use cases against human-coded reference databases from published meta-analyses, using separate study sets for prompt development and evaluation. Factor loadings were extracted with near-perfect accuracy (sensitivity ≥ 0.92, precision ≥ 0.98), whereas directly reported effect sizes were recovered completely (sensitivity = 1.00) but with lower precision (0.54–0.69) due to overextraction. In a verification study with four coders, automated extraction plus human verification reduced coding time on average by 42% relative to manual extraction at comparable accuracy. MASEMiner thus supports a human-in-the-loop workflow that reduces manual coding effort while preserving transparency and reproducibility in research synthesis.
ER -