Mixed ANOVA Calculator

Calculator Inputs

Between-Subject Factor Name

Within-Subject Factor Name

Within-Level Labels

Significance Level

Decimal Places

Delimiter

Upload CSV File

First row contains headers

Wide Dataset

Required format: first column subject ID, second column group label, remaining columns repeated measures. Equal subject counts per group are required.

Example Data Table

Subject	Group	Pre	Mid	Post
C1	Control	52	55	57
C2	Control	49	51	53
C3	Control	50	52	54
T1	Treatment	53	60	68
T2	Treatment	54	62	70

Formula Used

For a balanced split-plot design, the calculator partitions variability into between-group, within-subject, interaction, and their matching error terms.

Between-group sum of squares: SS_A = b × n × Σ(M_g.. − GM)²

Subjects within groups: SS_S/A = b × Σ(M_gs. − M_g..)²

Within-factor sum of squares: SS_B = a × n × Σ(M_.t − GM)²

Interaction sum of squares: SS_AB = n × Σ(M_gt. − M_g.. − M_.t + GM)²

Within-subject error: SS_B×S/A = Σ(Y_gst − M_gs. − M_gt. + M_g..)²

Mean squares are computed as MS = SS / df. F statistics are then formed as F_A = MS_A/MS_S/A, F_B = MS_B/MS_B×S/A, and F_AB = MS_AB/MS_B×S/A. Partial eta squared is reported as SS_effect / (SS_effect + SS_error).

How to Use This Calculator

Prepare a balanced wide-format dataset with subject ID, group, and repeated-measure columns.
Paste the data or upload a CSV file.
Enter factor names and optional within-level labels.
Choose alpha, decimals, and delimiter detection rules.
Run the analysis to view the ANOVA table, descriptive statistics, and interaction plot.
Download the output as CSV or PDF for reporting.

Frequently Asked Questions

1. What design does this calculator support?

It supports one between-subject factor and one repeated-measures factor in a balanced split-plot design. Every group must have the same number of subjects, and each subject must have values for all repeated levels.

2. What input format should I use?

Use wide format. Put subject IDs in column one, group names in column two, and repeated-measure values in the remaining columns. The sample dataset shows the expected structure.

3. Why does the calculator require balanced groups?

The formulas implemented here assume equal sample sizes across groups. Balanced data keeps the split-plot sums of squares and error terms straightforward and transparent for reporting.

4. What does the interaction effect mean?

The interaction tests whether change across repeated levels differs by group. A significant interaction suggests the pattern over time is not the same for every group.

5. Why are Greenhouse-Geisser and Huynh-Feldt values shown?

They adjust within-factor degrees of freedom when sphericity may be violated. Corrected p-values are especially useful when the repeated factor has three or more levels.

6. What effect size is reported?

The calculator reports partial eta squared for each tested effect. This helps you judge practical importance in addition to statistical significance.

7. Can I use missing values?

No. Each subject needs a complete set of repeated measurements. If your data contain missing values, clean or impute them before using this calculator.

8. What should I report from the output?

Report the F statistic, degrees of freedom, p-value, corrected p-values for repeated effects when relevant, partial eta squared, group means, and the interaction plot if useful.