Data and analysis code for: Interpreting recovered soil mass in forensic science: evidence from a full factorial experimental study

This repository contains the dataset and analysis code supporting the study "Interpreting recovered soil mass in forensic science: evidence from a full factorial experimental study" (Mendes, L.R.; de Oliveira, M.L.; da Veiga, M.A.M.S.), submitted to Forensic Science International (Elsevier). Study design: a full factorial experiment combining 5 soils of contrasting physical and pedological properties, 2 moisture regimes (dry, wet), 3 fabric types, 2 contact modalities, and 3 recovery methods (180 combinations, one observation per combination). The response variable is the dry mass of soil recovered from each textile surface. Contents: (1) data_repo.xlsx — the full dataset (180 records) and every statistical result reported in the article (GLM/ANOVA Type III main-effects and interaction models, adjusted means with Tukey HSD grouping, residual diagnostics, and the stratified models for fabric, contact modality, and recovery method), organized in sheets that map directly onto the article's Tables 3-7. A ReadMe sheet documents the correspondence between each sheet and the corresponding article table. (2) documentation.py — the Python script (pandas, statsmodels, scipy) that reproduces every result in the workbook from the raw dataset. Self-contained: license and requirements are documented in the file header. Author contributions (CRediT): L.R. Mendes — Conceptualization, Methodology, Investigation, Formal analysis, Data curation, Writing – original draft. M.L. de Oliveira — Visualization, Formal analysis, Writing – review & editing. M.A.M.S. da Veiga — Conceptualization, Methodology, Funding acquisition, Supervision, Writing – review & editing. Note: soil identification, collection context, and physicochemical characterization (article Tables 1 and 2) are not included in this repository; they are reported in the article's Materials and Methods and Supplementary Material.