Which ANOVA should you use?
- One-way ANOVA: compare a numeric outcome across the levels of one categorical treatment or group.
- Two-way ANOVA: examine two categorical factors and their interaction with a numeric outcome.
- Three-way ANOVA: examine three factors, their pairwise interactions, and their three-factor interaction. See the dedicated three-factor ANOVA calculator guide.
For factorial ANOVA, Statloft fits a full fixed-effects model. Each selected factor combination needs observations, and replicated observations are required to estimate residual error. Unequal cell sizes receive a caution in the results.
Prepare treatment data for analysis
Upload a CSV or XLSX table with one row per experimental unit or replicate. Include a numeric outcome column and a categorical column for each treatment or factor. If a file contains plant subsamples within a replicate, use the replicate mean when that matches your experimental design; do not count subsamples as independent replicates.
Statloft reviews column types and missing cells before analysis. Blank or nonnumeric outcomes are excluded with a visible count, while missing factor combinations and designs with no error degrees of freedom stop with a correction message.
Check assumptions before interpreting an F-test
ANOVA relies on independent observations, approximately normal model residuals, and similar residual variances. The standard Statloft models do not adjust for blocks, repeated measurements, or random effects. If the same plot, person, or experimental unit appears more than once, choose an analysis designed for that dependence or consult a statistician.
A statistically significant ANOVA indicates evidence that at least one modeled mean or effect differs; it does not identify every differing treatment pair. For treatment comparisons, review the model and consider a suitable post-hoc method such as DMRT where scientifically justified.
How to run an ANOVA in Statloft
- Open the free workspace and upload your CSV or Excel file.
- Confirm the header row, column types, treatment variable, and numeric outcome.
- Select one-way, two-way, or three-way ANOVA and choose the required factors.
- Read the sample size, model table, assumptions, and data checks before exporting.