Calculator Library / Hypothesis Testing
Use this app to compare processes, validate improvements, and interpret statistical results in plain language. It supports Welch two-sample t-tests, one-way ANOVA, and chi-square contingency testing with manufacturing-specific presets.
Tester
Default decision rule: p < alpha indicates statistical significance
Use one line per group in the format `Group Name: value, value, value`.
Use CSV-style rows. The first row contains column labels, and the first column contains row labels.
Detail
Use the preset library to quickly test common manufacturing comparisons like before/after cycle time, line-to-line output, or defect distributions.
Preset Guidance
T-test: Compare two means, such as before vs. after improvement or Process A vs. Process B.
ANOVA: Compare three or more groups, such as shifts, lines, or suppliers.
Chi-square: Compare categorical count patterns, such as pass/fail by shift or defect type by supplier.
Instructions
Statistical significance answers whether the observed difference is unlikely to be due to random variation alone. Practical significance answers whether the difference is large enough to matter operationally.
This app is meant for fast decision support. For regulated or high-risk decisions, confirm the study design, data assumptions, and follow-up analysis before finalizing conclusions.
This tool helps teams compare process results using formal statistical tests instead of gut feel. It supports common comparisons like t-tests, ANOVA, and chi-square logic so engineers can ask whether a change is statistically meaningful, not just numerically different.
Use it for before/after trials, supplier comparisons, process experiments, audit findings, and project validation where a decision needs evidence rather than anecdote.
| Output | Meaning | Use |
|---|---|---|
| p-value | Probability of seeing the observed data if no real difference exists | Tests whether the result is statistically significant. |
| Confidence interval | Estimated range for the true effect | Shows magnitude and uncertainty together. |
| Effect size | Strength of the practical difference | Helps distinguish important improvement from trivial change. |
Suppose Process A averages 1.8% scrap and Process B averages 1.2% scrap over matched samples. A statistical test may show the difference is significant at 95% confidence, but the effect size and confidence interval still matter because the operational payoff may be small or large depending on volume and cost.
The tool helps tie those outputs together so the team does not stop at a p-value alone.
It tells you how compatible the data is with the no-difference assumption. It does not tell you how large or important the effect is.
Because a result can be statistically significant and still too small to matter operationally. Effect size keeps the conclusion tied to business reality.
Use ANOVA when comparing more than two groups or conditions. Repeating t-tests raises false-positive risk.
Using the wrong test for the data type or acting on significance without checking sample size, assumptions, and practical importance.
They show the plausible range of the true effect, which is often more decision-useful than a single yes-or-no significance call.
Use the workbook when the statistical test needs to sit next to capability, sigma, and defect metrics in one review package.
Use the guide to place statistical testing inside Measure, Analyze, and Control decisions instead of treating it as a standalone math step.