When you find a significant difference between groups using an ANOVA, you still don't know exactly which groups differ from one another. Use Tukey HSD post-hoc to compare every possible pair of groups. This helps you pinpoint exactly where your data differs while keeping your results reliable and accurate.
Understanding Pairwise Comparisons
An ANOVA tells you if there is a difference somewhere in your data, but it acts like a broad sweep. It flags that at least one group is unique, but it does not tell you which one. Tukey HSD post-hoc bridges this gap by breaking down the data into every possible pair of groups.
By systematically evaluating every pair, you gain clarity on your experiment's specific interactions. Lattice automates this calculation to show you the difference in means for every comparison and whether that difference is statistically significant.
Controlling for Family-Wise Errors
Whenever you perform many comparisons simultaneously, the chance of seeing an accidental difference increases. Tukey HSD post-hoc addresses this by controlling the 'family-wise error rate.' This ensures that your findings are not just lucky snapshots but represent genuine differences in your dataset.
The tool provides adjusted p-values for every pair. You can use these to confidently decide which groups stand out, knowing the risk of false positives has been mathematically mitigated.
Interpreting Your Results
Lattice presents your Tukey HSD post-hoc results as a structured list of pairs. For each pair, you will see the calculated mean difference, the confidence interval for that difference, and a simple indicator of whether the difference is significant.
If the confidence interval for a pair does not include zero, you can generally interpret that as a meaningful difference between those two groups. This allows you to quickly isolate which conditions or categories are driving the trends in your analysis.
1 · Intent → method
An LLM picks svt_posthoc_tukey from a fixed catalog.
2 · Method → numbers
Deterministic Python engine runs the math. Same input → same output.
3 · Numbers → plain language
A second LLM translates the result into your domain’s vocabulary.
Why use Tukey HSD post-hoc instead of running multiple t-tests?
Running multiple t-tests increases the likelihood of finding a false positive by chance. Tukey HSD post-hoc adjusts your analysis to account for all these comparisons at once, ensuring that your conclusions about which groups are different remain statistically sound.
What happens if my group sizes are different?
Lattice uses the Tukey-Kramer variation for this method, which is specifically designed to handle unbalanced data. You can proceed with confidence even if the number of observations in each group is not the same.
Tool input schema
Schema for svt_posthoc_tukey not exported yet (run pnpm export:registry).