Two-Way ANOVA Calculator

Two-way analysis of variance examines one continuous outcome across combinations of two categorical factors. A 2 x 2 design has two levels of Factor A, two levels of Factor B, and four cells representing every possible combination. Unlike separate one-way tests, the factorial model estimates both main effects and an interaction in a single analysis. This calculator analyzes balanced designs with replication, meaning every cell must contain the same number of independent observations and at least two observations are needed per cell. The Factor A main effect compares the average outcome across A1 and A2 after averaging over both levels of Factor B. The Factor B main effect similarly compares B1 and B2 after averaging over Factor A. The interaction tests whether the effect of one factor changes depending on the level of the other factor. An interaction can be important even when neither main effect is large. For example, two treatments might have opposite effects in two age groups, producing equal overall treatment means but a strong crossover interaction. The calculator partitions variation into Factor A, Factor B, A-by-B interaction, and residual error. Each effect sum of squares has one degree of freedom in a 2 x 2 design. Residual variation is computed from differences between observations and their own cell means. Dividing each effect mean square by the residual mean square produces its F-statistic. Larger F values indicate that an effect is large compared with unexplained variation. Formal p-values require evaluating each statistic against an F distribution with one numerator degree of freedom and the displayed error degrees of freedom. Interpretation depends on standard ANOVA assumptions: independent observations, approximately normal residuals in every cell, and similar variances across cells. Balanced samples make the analysis more robust and keep the effects orthogonal, but they do not correct poor randomization, outliers, or dependent measurements. If the interaction is meaningful, interpret simple effects or cell comparisons before summarizing main effects, because averaging can conceal the pattern. Report cell means and variability alongside F statistics, degrees of freedom, p-values, and effect sizes. This tool is intended for transparent calculations, classroom examples, and quick verification of balanced replicated designs.