Upload a CSV, pick your outcome and two groups, set a margin — and get the full TOST equivalence analysis: the difference plotted against your margin band, both one-sided tests, and the honest contrast with the standard t-test. Free.
Free analyses run on up to 10,000 rows. Larger files are randomly sampled to that size — sign up to analyze your full dataset.
Running the two one-sided tests...
Sent to — the difference-vs-margin plot, both one-sided tests with the combined verdict, group statistics, full methods disclosure, R code, and AI insights.
Analyze another fileThe analysis runs the classical two one-sided tests (TOST) on the Welch t statistic: one test rules out the difference being below the lower margin, the other rules out it being above the upper margin, each at alpha 0.05. Equivalence is concluded only when both reject — exactly when the 90% confidence interval for the mean difference lies entirely inside the margin band. The 95% interval, mean difference with standard error, and the standard Welch t-test are reported alongside, with a computed explanation of why a non-significant t-test alone never demonstrates equivalence. When only a deficit matters, a one-sided non-inferiority variant tests that the new group is not worse than the reference by more than the margin.
Use it when the interesting claim is sameness: validating a substitute, showing a change caused no meaningful harm, or demonstrating parity between two arms — with a margin that encodes how big a difference would actually matter.
Not for detecting whether groups differ (use the group comparison tool), for paired before/after data on the same subjects, or when no defensible margin can be stated and defaulting to 0.2 x pooled SD would be misleading.
Built for: Analysts, engineers, and researchers who need to demonstrate parity or no-harm rather than difference
Typical data source: Any spreadsheet or CSV with a numeric column and a two-level group column
One numeric outcome plus a two-level group column. For example, fill volumes from two production lines:
Minimum 10 rows · Best with 50-50,000 rows with two groups of comparable size
Standard-library analysis: are these two things practically the SAME within a margin? A standard t-test can never prove sameness — 'not significantly different' may just mean too little data. This tool runs the classical two one-sided tests (TOST): the mean difference with its 90% (TOST-consistent) and 95% confidence intervals plotted against your equivalence margin band, both one-sided tests spelled out, the standard t-test alongside for contrast, and a one-sided non-inferiority variant when only a deficit would matter. Supply a margin in your outcome's units, or let the tool use a clearly-labeled default of 0.2 x pooled SD.
The mean difference with its 90% and 95% confidence intervals against the margin band — equivalence is the 90% interval sitting entirely inside the band.
Both one-sided tests, the combined TOST verdict, and the standard t-test for contrast — the table that separates 'no evidence of difference' from 'evidence of sameness'.
Each group's n, mean, spread, and median, with the raw gap sized against the margin.
Full method and margin disclosure: the TOST rule, the 90% interval logic, and where the margin came from.
Plain-English interpretation — what the numbers mean, what's significant, and what to do next.
Prove the change did NOT make things worse
Map your metric as the outcome and old-vs-new as the group, and set a margin for how much degradation you could tolerate. A non-significant t-test cannot make this case — the TOST can, by actively ruling out any difference beyond your margin.
See our FAQ for details on pricing, data privacy, and how the analysis works. Every report includes a Methodology section showing the statistical test, assumptions checked, and diagnostics run.
Run any analysis on your own data — validated R analyses, interactive reports, AI insights, and PDF export.
Try Free — No Credit CardTell us what went wrong, in your own words. We capture the page you're on automatically, so no need to describe where you are.