Agricultural Experimental Designs in R • agrideshr

Agricultural Experimental Designs in R

A tutorial R package demonstrating the analysis of five classical experimental designs used in agricultural research. Each design includes a built-in dataset and a detailed vignette covering the full analysis pipeline.

Installation

# Install from GitHub
devtools::install_github("emantzoo/agrideshr", build_vignettes = TRUE)

Designs Covered

#	Design	Vignette	When to Use
1	Completely Randomized (CRD)	`vignette("01-crd")`	Homogeneous units, one factor
2	Randomized Complete Block (RCBD)	`vignette("02-rcbd")`	One known source of variability
3	Latin Square	`vignette("03-latin-square")`	Two sources of variability (rows + columns)
4	Factorial	`vignette("04-factorial")`	Two+ crossed factors, interest in interactions
5	Split-Plot	`vignette("05-split-plot")`	Two factors with restricted randomisation

Which Design Should I Use?

Start
 |
 +-- How many treatment factors?
      |
      +-- ONE factor
      |    |
      |    +-- Are experimental units homogeneous?
      |         |
      |         +-- YES --> CRD
      |         |
      |         +-- NO, one source of variability --> RCBD
      |         |
      |         +-- NO, two sources (grid layout) --> Latin Square
      |
      +-- TWO or more factors
           |
           +-- Can all factors be fully randomised?
                |
                +-- YES --> Factorial Design
                |
                +-- NO, one factor is harder to change --> Split-Plot

Each Vignette Covers

When to Use – decision criteria
The Design – model equation and layout
Data – package dataset with data()
Exploratory Visualization – ggplot2 plots
Model Fitting – aov() or lmer()
Assumption Checking – Q-Q plot, Shapiro-Wilk, Levene’s test
Post-hoc Comparisons – Tukey HSD and emmeans contrasts
Conclusion – summary and links to the next design

Datasets

data(crd_data)           # 3 squash varieties, 3 replicates
data(rcbd_data)          # 6 fertilizer levels, 4 blocks
data(lsd_data)           # 4x4 Latin square
data(factorial_data)     # 3 temperatures x 3 fertilizers
data(split_plot_data)    # 6 whole-plot x 3 sub-plot, 6 replicates

License

MIT