How to Do a Meta-Analysis: Step-by-Step Guide
A practical workflow for planning, calculating, checking, interpreting, and reporting a defensible meta-analysis.
Read guide →Statistical Methods Guides and Decision Tools
This collection helps researchers move from a study question to an appropriate statistical method. Use the decision guide first when the analysis is unclear, then open the focused guides for confidence intervals, effect sizes, assumptions, and reporting.
Core answer
Start with the research question, outcome type, study design, and dependence structure. Then choose a method that estimates the quantity you need and report its uncertainty and effect size.
Multiple Comparison Adjustment: Which Method Should You Use?
Choose a multiple-testing correction by defining the hypothesis family, deciding whether to control FWER or FDR, and checking dependence assumptions.
Read guide →Missing Data in Statistical Analysis: Delete, Impute, or Model?
The right missing-data strategy depends on why values are missing, how much information is lost, the analysis model, and the assumptions you can defend.
Read guide →How to Interpret Logistic Regression: Odds Ratios, CIs, and Examples
Logistic regression models log odds, while odds ratios express multiplicative changes in odds. Predicted probabilities often communicate results more clearly.
Read guide →Eta Squared vs Partial Eta Squared: Which Effect Size Should You Report?
Eta squared and partial eta squared use different denominators, so the same ANOVA can produce different values.
Read guide →Confidence Intervals for Proportions: Wald, Wilson, and Exact Methods
The familiar Wald interval can perform poorly with small samples or proportions near zero or one.
Read guide →Which Statistical Test Should I Use? A Practical Decision Guide
Start with the research question and study design, then use variable type and data structure to narrow the method.
Read guide →How to use this collection
Begin with the guide closest to your immediate decision. Record the assumptions, settings, and definitions you use, then open the linked DataStatPro workflow to apply the method. Educational examples should be adapted to the actual study protocol and data structure.