Validation and limits

This page records the methods, definitions, reference checks, and limits of the current public editor.

Supported inference

• Pairwise t-tests on Column and XY tables: unpaired Welch, unpaired Student (equal variance), and paired
• Difference is mean(B) − mean(A); 95% CI uses the matching standard error and t critical value
• Core Student/Welch values are locked by fixed-vector unit tests against documented formulas and high-precision reference values
• Significance brackets can be filled on Column graphs; XY can compute the test but does not place column-centered brackets

Descriptive statistics

• Mean, sample SD (n−1), SEM, and mean 95% CI via jstat
• Geometric mean and geometric SD on positive values only
• Median; median 95% CI uses the Altman/Campbell order-statistic approximation, which can differ from software that uses exact binomial ranks
• Kaplan–Meier summaries use Greenwood standard errors and log-log 95% CI

Not provided

• ANOVA, multiple-comparison procedures, chi-square, Fisher’s exact test, and log-rank
• Hypothesis tests for Grouped, Nested, Contingency, and Survival tables
• Dose-response / nonlinear regression as a finished analysis workflow

How to cite

Use a versioned URL and an access date. Do not invent a paper citation.

PivoGraph. https://pivograph.com/ (accessed YYYY-MM-DD).

For reproducibility, record the PivoGraph version, test design, variance assumption, and other analysis settings used for the figure.