Most people do not struggle with the arithmetic of statistics. They struggle with the choices: which test fits this data, whether the assumptions hold, and what the output actually means. Stat Dojo is built around those choices.
Start with the method selector, or pick a topic from the map. Each page teaches one method completely, from the idea to the report sentence. 3 of 28 planned topics are live, and more are on the way.
Method selector
Which method should I use?
Answer a few questions about your data and your question. You will get the method, what to check first, the Excel and Minitab route, and where to learn more.
The selector needs JavaScript. The table below lists every method.
See every method in one table
| Method | What it does | Excel | Minitab |
|---|---|---|---|
| Two-sample t-test | Compares the means of two independent groups of measured values. Use the Welch version unless you have a strong reason to assume equal variances. | ||
| Mann-Whitney test | Compares two independent groups using ranks, so it does not need normal data. | ||
| Paired t-test | Compares two measurements on the same units, such as before and after, by testing whether the average difference is zero. | ||
| Wilcoxon signed-rank test | The rank-based alternative to the paired t-test. It tests whether the median difference is zero. | ||
| One-way ANOVA | Compares the means of three or more groups defined by one factor, with a single test that controls the false-alarm rate. | ||
| Kruskal-Wallis test | The rank-based alternative to one-way ANOVA for three or more groups. | ||
| Two-way ANOVA (factorial) | Studies two or more factors at once and shows whether the effect of one depends on the level of another (an interaction). | ||
| Tests for equal variances | Compares the spread of two or more groups. Levene's test is robust to non-normal data; the F test and Bartlett's test assume normality. | ||
| Correlation and simple regression | Measures how strongly one input moves with an output and fits a line to predict it. Correlation is not cause. | ||
| Multiple regression | Models an output from several inputs at once and shows each input's effect with the others held constant. | ||
| Binary logistic regression | Models the chance of a pass/fail outcome from continuous or categorical inputs. | ||
| Chi-square test | Tests whether two categorical variables are related, or whether counts across three or more groups differ from what you expect. | ||
| One-sample t-test | Tests whether the average of a sample differs from a target or standard value. | ||
| One-proportion test | Tests whether a defect rate or pass rate differs from a target. | ||
| Two-proportion test | Compares the rates of two groups, such as the defect rate before and after a change. | ||
| One-variance test | Tests whether the spread of a process differs from a target value. | ||
| Poisson rate test | Compares a rate of defects per unit, where one unit can have several defects. | ||
| Normality test and probability plot | Checks whether the data (or the residuals of a model) are close enough to a normal distribution for tests that assume it. | ||
| Sample size and power | Decide how many observations you need to detect a difference of a given size with a given confidence. | ||
| Describe and plot the data | Always look at the data before testing it: the center, the spread, the shape, and the outliers. | ||
| Process capability | Compares what the process delivers with what the customer allows, using Cp, Cpk, Pp, and Ppk. | ||
| Control charts | Separates normal variation from signals that the process has changed, so you act only when you should. |
The Dojo Map
Six groups, from the data itself to the statistics behind control charts and experiments. Pages marked coming soon are planned.
Foundations
What the data are, how to describe them, and the distributions behind every test.
- Data Types and ScalesComing soon
Continuous, discrete, ordinal, and nominal data, and why the type decides the method.
- Foundation
- Descriptive StatisticsComing soon
Mean, median, mode, range, variance, and standard deviation, and when each misleads.
- Foundation
- Distributions: Normal, Binomial, PoissonComing soon
The shapes behind process data and how to tell which one you have.
- Green Belt
- Sampling and the Central Limit TheoremComing soon
Why averages behave better than individuals, and what a standard error means.
- Green Belt
- Graphical AnalysisComing soon
Histograms, box plots, dot plots, and scatter plots: look before you test.
- Foundation
Estimation
How sure are you? Intervals, sample size, and the power to see a real effect.
- Confidence IntervalsComing soon
A range of plausible values for a mean, a proportion, or a difference.
- Green Belt
- Sample Size and PowerComing soon
How much data you need to see the difference that matters.
- Green Belt
Hypothesis Tests
Decide whether a difference is real or just noise: means, proportions, variation, and shape.
- P-Values and Error TypesComing soon
Type I and Type II error, alpha, beta, and what a p-value does and does not say.
- Green Belt
- One- and Two-Sample t-TestsComing soon
Compare a mean with a target, two means, or paired measurements.
- Green Belt
- Tests for ProportionsComing soon
Defect rates and pass rates: one proportion, two proportions, and the sample you need.
- Green Belt
- Tests for VariancesComing soon
Is one process more variable than another? F, Bartlett, and Levene tests.
- Black Belt
- Chi-Square TestsComing soon
Association between categories, and goodness of fit to a distribution.
- Green Belt
- Normality Tests and Probability PlotsComing soon
Check the assumption most tests lean on, and decide what to do when it fails.
- Green Belt
- Nonparametric TestsComing soon
Mann-Whitney, Wilcoxon, Kruskal-Wallis, and Mood's median when the data will not behave.
- Black Belt
ANOVA
Compare three or more groups, and then find out which ones differ.
- One-Way ANOVA
Compare the means of three or more groups with one factor.
- Green Belt
- Two-Way ANOVA and Interactions
Two factors at once, and what an interaction means.
- Black Belt
- Post-Hoc Comparisons and ANOVA Assumptions
Tukey, Fisher, and Dunnett: which groups differ, and how to check the model.
- Black Belt
Relationships
Correlation and regression: how inputs drive outputs, and how to check the model.
- CorrelationComing soon
Strength and direction of a linear relationship, and why it is not causation.
- Green Belt
- Simple Linear RegressionComing soon
Fit a line, read the slope and R-squared, and predict with care.
- Green Belt
- Multiple RegressionComing soon
Several inputs at once: adjusted R-squared, VIF, and model selection.
- Black Belt
- Binary Logistic RegressionComing soon
Model a pass/fail outcome from continuous or categorical inputs.
- Black Belt
- Residual Analysis and Model CheckingComing soon
Read residual plots, find outliers and leverage, and fix a poor model.
- Black Belt
Quality Statistics
Control charts, capability, reliability, and experiments, with the statistics underneath.
- Control Chart TheoryComing soon
Why 3-sigma limits work, how constants are derived, and the false-alarm rate.
- Green Belt
- Capability StatisticsComing soon
Cp, Cpk, Pp, Ppk, confidence intervals on capability, and non-normal data.
- Black Belt
- Reliability and Weibull AnalysisComing soon
Life data, censoring, the bathtub curve, and the Weibull shape parameter.
- Black Belt
- Analyzing Designed ExperimentsComing soon
Effects, interactions, ANOVA tables, and residuals for factorial designs.
- Black Belt
- Statistics Formula SheetComing soon
Every formula used in the dojo on one printable page.
- Foundation
- Statistics GlossaryComing soon
Plain-language definitions of the terms in the dojo.
- Foundation
How Every Page Is Built
Every topic page follows the same structure, so you always know where to look.
- The idea in plain language, with a figure, before any formula.
- When to use it, and what to use instead when it does not fit.
- A worked example by hand, so you can see where every number comes from.
- Excel and Minitab, step by step, with the output shown and the key numbers marked.
- How to read and report the result, including a sentence you can use.
- Assumptions, mistakes, and a practice problem with the answer.
Levels
| Level | Meaning |
|---|---|
| Foundation | Core ideas that every other page relies on |
| Green Belt | Methods a Green Belt uses on a DMAIC project |
| Black Belt | Methods that need more statistical depth or judgement |
Already on This Site
These guides, tools, and short definitions cover related ground and stay available as the dojo grows.
Guides
- Hypothesis TestingChoosing and reading tests
- Process CapabilityCp, Cpk, Pp, Ppk
- Measurement System AnalysisGage R&R and agreement studies
- Design of ExperimentsFactorial designs and effects
- SPC Control ChartsMonitoring variation
- Sigma Level, DPMO, and RTYConverting defects to sigma
- Black Belt Pocket GuideStatistics in the project context
- Green Belt Pocket GuideThe essentials
Tools
- One-Way ANOVA CalculatorCompare groups and run Tukey-Kramer
- Hypothesis Testing Quick TesterTest a difference
- Sample Size and Confidence CalculatorPlan the sample
- Standard Deviation CalculatorSummarize data
- Process Capability HelperCapability indices
- Control Chart SelectorChoose a chart
- DOE Quick PlannerPlan experiments
- Gauge R&R CalculatorMeasurement system study
Stat Dojo: Frequently Asked Questions
Who is Stat Dojo for?
Lean Six Sigma Green Belts and Black Belts, quality engineers, and anyone who has to decide from data whether a change made a difference. Each page is tagged Foundation, Green Belt, or Black Belt so you can start at the right level.
Do I need Minitab?
No. Every topic page shows how to do the analysis by hand and in Excel. Minitab steps are included because Minitab is the most widely used statistical software in quality work, and the page shows what its output looks like and how to read it.
How do I know which test to use?
Use the method selector at the top of this page. It asks about your question, the type of data, and the number of groups, and it points you to the method, the Excel and Minitab route, and the page that explains it.
Are the examples based on real data?
The examples use made-up data that was designed to be realistic and to give clean, checkable answers. The calculations are done with statistical software and the numbers on each page match each other.
Which topics are available now?
The topic map below shows which pages are live and which are planned. New pages are added regularly, and the method selector links to them as they appear.
Sources and Further Reading
- NIST/SEMATECH, e-Handbook of Statistical Methods (itl.nist.gov/div898/handbook).
- Douglas C. Montgomery, Introduction to Statistical Quality Control and Design and Analysis of Experiments, Wiley.
- David S. Moore, George P. McCabe, and Bruce A. Craig, Introduction to the Practice of Statistics, Freeman.
- Minitab Support documentation (support.minitab.com) and Microsoft Support documentation for the Excel Analysis ToolPak (support.microsoft.com).
This content is educational. Worked examples use made-up data. Menu names for Minitab follow recent versions of Minitab Statistical Software and can differ slightly in older releases; Excel steps use Microsoft 365 and the Analysis ToolPak.