A Model Context Protocol (MCP) server providing statistical tools to detect potential data fabrication or errors in research papers.
- Benford's Law Test: Detects anomalies in leading digit distributions.
- Terminal Digit Test: Checks for uniform distribution of last digits.
- GRIM Test: Verifies if a reported mean is mathematically possible for a given sample size.
- SPRITE Test: Reconstructs potential integer distributions to verify mean and standard deviation consistency.
- Integrity Score: Aggregates multiple test results into an overall verdict.
Add this to your MCP settings:
{
"mcpServers": {
"research-integrity": {
"command": "npx",
"args": ["-y", "github:kooltuoehias/research-integrity-mcp"]
}
}
}Tests whether a dataset follows Benford's Law. Use on large datasets of naturally occurring measurements (prices, populations, financial data). Requires minimum 30 values.
Tests whether the last digits of reported values are uniformly distributed (0-9). Real continuous measurements have uniform terminal digits. Fabricated data often clusters on round numbers (0, 5).
Granularity-Related Inconsistency of Means. Tests whether a reported mean is possible given integer-scale data and sample size (e.g., Likert scales).
Sample Parameter Reconstruction via Iterative TEchniques. Tests whether any integer distribution can produce the reported mean AND standard deviation. Automatically runs GRIM first.
Aggregates ToolResults into an overall verdict (PASS, SUSPICIOUS, FAIL, CRITICAL) with recommended actions.
These tools are based on established techniques in the field of research integrity:
- GRIM: Brown, N. J. L., & Heathers, J. A. J. (2017). The GRIM Test: A Simple Technique Detects Numerous Anomalies in the Reporting of Social Science Results.
- SPRITE: Heathers, J. A. J., et al. (2018). Recovering data from summary statistics: Sample Parameter Reconstruction via Iterative TEchniques (SPRITE).
Part of a research integrity toolkit. Image analysis tools are available in a separate Python-based skill.
ISC