Free online book · a live book, updated as it grows

Practical Statistics

The methods you meet in published papers and need for your thesis analysis, beyond the t-test and chi-square test. Each chapter explains the idea with diagrams first, lets you change the numbers yourself, and then reads a real open-access article that used the method.

21 chapters45 related TNMGRMU questions

New to statistics? Start with the Research Methodology book. Each chapter here lists the Research Methodology chapters to read first.

Getting the data ready

  1. 01Checking assumptionsRead nowIs your data normal enough for the test?
  2. 02Missing dataRead nowWhat do you do about missing values?

Comparing groups beyond the t-test

  1. 03Multiple comparisonsRead nowAfter ANOVA, which groups differ?
  2. 04Two-way ANOVA and interactionRead nowDo two factors act together?
  3. 05Repeated measures and mixed modelsRead nowWhat if the same patients are measured again and again?
  4. 06ANCOVA and baseline adjustmentRead nowHow do you allow for the baseline value?
  5. 07Equivalence and non-inferiority testsRead nowHow do you show that two treatments are about the same?

Regression models

  1. 08Multiple linear regressionRead nowHow do several factors together predict an outcome?
  2. 09Multicollinearity and model buildingRead nowWhen predictors overlap, which one matters?
  3. 10Logistic regression in depthRead nowHow do you model a yes/no outcome with several factors?
  4. 11Poisson and log-linear modelsRead nowHow do you model counts and rates?
  5. 12Cox regressionRead nowWhich factors change the time to an event?
  6. 13Stratification and propensity scoresRead nowHow can an observational study imitate randomisation?

Measurement, agreement and diagnosis

  1. 14Intraclass correlation and Bland-AltmanRead nowDo two methods or raters give the same measurement?
  2. 15Kappa and agreement on categoriesRead nowDo two observers agree on categories?
  3. 16Reliability and validity of scalesRead nowCan you trust a questionnaire score?
  4. 17Laboratory quality control and reference intervalsRead nowIs the laboratory result in control, and is it normal?
  5. 18Bayes' rule and combining testsRead nowHow does a test result change what you already suspected?

Many variables, validation and pitfalls

  1. 19Multivariate methodsRead nowWhat if there are many outcomes or many variables?
  2. 20Bootstrap and model validationRead nowWill your model work on new patients?
  3. 21Statistical fallaciesRead nowWhy do correct numbers still mislead?

Worked examples use made-up numbers for teaching unless a published source is named. Every open-access article is linked to its DOI and free full text, and every statistic quoted from it was checked against the article.