Plotting the Veterinary Examples

1 Plotting Principles

The original book uses simple black-and-white plots. In R, ggplot2 can recreate the same scientific comparisons with clearer labels and reproducible code. The goal is not decorative plotting; it is to show design structure, grouping, and model implications.

2 Split-Plot Packed Cell Volume Data

if (requireNamespace("collapse", quietly = TRUE) && requireNamespace("ggplot2", quietly = TRUE)) {
  ex124_plot <-
    ex124 |>
    collapse::fmutate(herd = factor(herd))

  ggplot2::ggplot(
    ex124_plot,
    ggplot2::aes(x = herd, y = PCVdif, shape = dose, color = drug)
  ) +
    ggplot2::geom_point(size = 2.4, position = ggplot2::position_jitter(width = 0.08)) +
    ggplot2::facet_wrap(~ drug, nrow = 1) +
    ggplot2::labs(
      x = "Herd",
      y = "Change in packed cell volume",
      shape = "Dose",
      color = "Drug",
      title = "Packed cell volume response by herd, drug, and dose"
    ) +
    ggplot2::theme_minimal()
}

3 Region, Drug, and Dose Means

if (requireNamespace("ggplot2", quietly = TRUE)) {
  ggplot2::ggplot(
    ex125,
    ggplot2::aes(x = dose, y = Pcv, group = Drug, color = Drug)
  ) +
    ggplot2::stat_summary(fun = mean, geom = "line") +
    ggplot2::stat_summary(fun = mean, geom = "point", size = 2.6) +
    ggplot2::facet_wrap(~ Region) +
    ggplot2::labs(
      x = "Dose",
      y = "Mean packed cell volume",
      title = "Mean response by region, drug, and dose"
    ) +
    ggplot2::theme_minimal()
}

4 Longitudinal PCV Data

if (requireNamespace("ggplot2", quietly = TRUE)) {
  ggplot2::ggplot(
    ex33,
    ggplot2::aes(x = time, y = PCV, group = animal_id, color = breed)
  ) +
    ggplot2::geom_line(linewidth = 0.5) +
    ggplot2::geom_point(size = 1.4) +
    ggplot2::labs(
      x = "Time",
      y = "Packed cell volume",
      color = "Breed",
      title = "Longitudinal packed cell volume by animal"
    ) +
    ggplot2::theme_minimal()
}

5 Model-Based Lines for Example 3.1

The book displays fitted lines for Example 3.1 Model 3. The packaged ex31 data can be plotted with simple group-specific linear fits to inspect the same relationship between baseline PCV and one-month PCV.

if (requireNamespace("collapse", quietly = TRUE) && requireNamespace("ggplot2", quietly = TRUE)) {
  ex31_plot <-
    ex31 |>
    collapse::fmutate(
      treatment = ifelse(drug == "BERENIL", paste(drug, dose), as.character(drug))
    )

  ggplot2::ggplot(
    ex31_plot,
    ggplot2::aes(x = PCV1, y = PCV2, color = treatment)
  ) +
    ggplot2::geom_point(size = 2) +
    ggplot2::geom_smooth(method = "lm", se = FALSE, linewidth = 0.8) +
    ggplot2::labs(
      x = "PCV at treatment",
      y = "PCV one month later",
      color = "Treatment",
      title = "Relationship between baseline and follow-up PCV"
    ) +
    ggplot2::theme_minimal()
}