The ScottKnott Clustering Algorithm.

ScottKnott

ScottKnott is an R package that implements the Scott & Knott clustering algorithm as a multiple comparison method in the Analysis of Variance (ANOVA) context, for both balanced and unbalanced designs.

Key Features

• Performs the Scott & Knott clustering algorithm for balanced and unbalanced designs.
• Accepts input from formula, aov, lm, aovlist, and lmerMod objects.
• Supports single, factorial, split-plot, split-plot in time, and split-split-plot experiments.
• Adjusted means via Least-Squares Means (emmeans) for unbalanced data.
• Rich plot method with customisable dispersion bands (min–max, SD, CI, pooled CI).
• Reporting support with xtable.

Installation

Install from CRAN:

install.packages("ScottKnott")

Install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("ivanalaman/ScottKnott")

Quick Start

library(ScottKnott)

## Completely Randomized Design (CRD) — balanced
data(CRD1)

sk1 <- with(CRD1,
            SK(y ~ x,
               data = dfm,
               which = 'x'))
summary(sk1)
plot(sk1, 
     dispersion = 'sd',
     d.col = 'steelblue')

## Randomized Complete Block Design (RCBD)
data(RCBD)

sk2 <- with(RCBD,
            SK(y ~ blk + tra,
               data = dfm,
               which = 'tra'))
summary(sk2)
plot(sk2, 
     dispersion = 'ci', 
     d.col = 'red')

Project Layout

• /R: Core functions and S3 methods.
• /man: Reference documentation (.Rd files).
• /data: Example datasets (CRD, RCBD, LSD, FE, SPE, SPET, SSPE, sorghum).
• /demo: Runnable demos for each experimental design.
• /inst: Package citation file.

Contributing

Contributions are welcome. Open an issue or submit a pull request with:

• Bug fixes and performance improvements.
• Documentation and usability updates.
• New ideas for grouping procedures or graphical displays.

To check and build locally:

R CMD check ScottKnott
R CMD build ScottKnott
R CMD INSTALL ScottKnott_X.X-X.tar.gz

Roadmap

• Add automated tests (testthat) for all experimental designs.
• Expand vignettes covering balanced and unbalanced use cases.
• Keep documentation aligned with current S3 behaviour.

Developed by:
Faria, J. C.; Jelihovschi, E. G.; Allaman, I. B.
Universidade Estadual de Santa Cruz - UESC
Departamento de Ciencias Exatas - DCEX
Ilheus - Bahia - Brasil.