tufte-data-forensics

You can detect p-hacking by running a forensic audit that checks for data substitution, unreasonable precision, and contradictory physics. This verifies that the analysis relies on primary measurements rather than manipulated surrogates.

The best way to verify dashboard analytics is to conduct a forensic audit using Tufte's principles of direct observation and provenance tracking. This identifies measurement errors and model-specification failures in the underlying data.

Yes, forensic auditing applies to institutional performance reporting by applying diagnostic tests for data substitution and batch effects. It traces metrics back to primary measurements to prevent the propagation of systemic errors and falsified claims.

You need a forensic audit when reviewing new research claims or analytical models to ensure they are free of measurement errors and bias. It is essential before using data to inform high-stakes decisions and prevent error propagation.

Diagnostic tests for data falsification include screening for data substitution, unreasonable precision, and contradictory physics. These tests identify measurement failures and verify that data provenance relies on primary observation rather than manipulated surrogates.