Working calculators for the questions that come up in regulated labs: do two methods agree, how many samples you need, what a spot-check proves, and how long a product stays in specification.
Working calculators, running entirely in your browser. Nothing you enter leaves this page. Each one is a small version of analysis TrueGate Bio does on real programs. For decisions that matter, have the analysis reviewed against your protocol and SOPs.
Bridging two methods or two instruments? Passing-Bablok regression separates constant bias (intercept) from proportional bias (slope) and tolerates error in both methods. Bland-Altman shows agreement and its limits. Paste paired results as two columns (method A, method B), one pair per line.
How many samples does a comparison need before a null result means anything? Set the difference you need to detect and your tolerance for false positives and false negatives.
You reviewed a sample of records (redactions, batch records, data entries) and found some defects. What can you honestly claim about the whole set? Uses the exact binomial (Clopper-Pearson) upper bound; with zero defects it reduces to the rule of three.
Linear regression of an attribute over time with a one-sided 95% confidence bound on the mean, following the approach in ICH Q1E for a single batch. Shelf life is where the bound meets the specification. Paste time and value pairs, one per line.