Check out my code 
# Packages ---------------------------------------------------------------------
pacman::p_load(
  dplyr,
  ggplot2
)

# Data -------------------------------------------------------------------------
data <- readRDS(here::here("analysis/data/data_long.RDS")) |>
  filter(wave == "2020") |>
  select(ID, status, Total_p)

To examine whether procrastination scores differed across cognitive status groups in 2020, we first tested the assumption of homogeneity of variances using Levene’s test, which was violated (\(p = 0.039\)). Consequently, we employed the Kruskal–Wallis test, a rank-based non-parametric alternative to one-way ANOVA, to evaluate whether procrastination scores differed significantly among participants with normative cognitive function, mild cognitive impairment (MCI), and dementia.

# Running test
kw_results <- kruskal.test(Total_p ~ status, data = data)

# Outputting result
broom::tidy(kw_results)
## # A tibble: 1 × 4
##   statistic  p.value parameter method                      
##       <dbl>    <dbl>     <int> <chr>                       
## 1      17.5 0.000155         2 Kruskal-Wallis rank sum test

The test revealed a significant effect of cognitive status on procrastination scores (\(\chi^2(2) = 17.54, \; p < 0.001\)), suggesting that at least two groups differ.

Post-hoc analysis

To identify where these differences lay, we conducted pairwise Wilcoxon rank-sum tests with Benjamini–Hochberg (BH) correction for multiple comparisons:

# Running multiple comparison
pairwise_res <- pairwise.wilcox.test(data$Total_p, data$status, p.adjust.method = "BH")

# Outputting result
broom::tidy(pairwise_res) |>
  mutate(across(c(group1, group2), ~ case_when(
  .x == "1" ~ "Normative Cognitive Function",
  .x == "2" ~ "MCI",
  .x == "3" ~ "Dementia"
  )))
## # A tibble: 3 × 3
##   group1   group2                       p.value
##   <chr>    <chr>                          <dbl>
## 1 MCI      Normative Cognitive Function 0.00394
## 2 Dementia Normative Cognitive Function 0.00500
## 3 Dementia MCI                          0.334

The results indicated that:

  • Normative cognitive function participants reported significantly lower procrastination scores than those with MCI (\(p = 0.004\)).
  • Normative cognitive function participants also reported significantly lower scores than those with dementia (\(p = 0.005\)).
  • No significant difference was found between the MCI and dementia groups (\(p = 0.334\)).

Visualisation

Figure fig-kruskal illustrates the distribution of procrastination scores across cognitive status groups. Individual scores are displayed as dotplots (left side of each box), and boxplots summarize the group distributions (right side). Significance bars represent the results of the pairwise comparisons described above.

Check out my code 
anova_fig <- data |>
  ggplot(aes(x = status, y = Total_p, fill = status, colour = status)) +
  gghalves::geom_half_point(
    side = "l", 
    transformation = ggbeeswarm::position_quasirandom(width = 0.15),
    size = 1.5, alpha = 0.5) +
  gghalves::geom_half_boxplot(side = "r", colour = "black") +
  ggsignif::geom_signif(
    comparisons = list(c("1", "2"), c("2", "3"), c("1", "3")),
    annotations = c("**", "NS", "**"),
    size = 1,
    textsize = 4.5,
    step_increase = 0.05,
    margin_top = 0.05,
    y_position = c(60, 65, 70),
    tip_length = 0.01,
    vjust = 0.01
  ) +
  ggokabeito::scale_fill_okabe_ito() +
  ggokabeito::scale_color_okabe_ito() +
  scale_x_discrete(labels = c(
    "1" = "Normative Cognitive Function", 
    "2" = "MCI", 
    "3" = "Dementia")) +
  scale_y_continuous(breaks = seq(0, 60, by = 10), 
                     expand = expansion(mult = 0.05, add = 1)) +
  labs(x = NULL, y = "Procrastination") +
  guides(fill = "none", colour = "none") +
  theme_minimal() +
  theme(axis.text.x = element_text(size = 10, face = "bold"))

anova_fig
Figure 1: Procrastination scores across cognitive status groups (2020).