重复测量设计：long-format数据直方图分数提取与绘图语法咨询

Hey there! Let's tackle your R questions step by step—since you mentioned using melt() to switch to long format, I’ll assume you’re working with typical tidy data structures in R.

You can grab these scores using either base R subsetting or the dplyr package for cleaner, pipe-friendly syntax:

• Base R approach:

  # Replace "Score" with your actual score column name if it's different
  a_scores <- df[df$Tests == "A", "Score"]
  
  # If you want the full rows (including Participant IDs) instead of just scores:
  a_test_data <- df[df$Tests == "A", ]
  

• dplyr approach (more readable for complex operations):

  library(dplyr)
  
  # Pull just the scores as a numeric vector
  a_scores <- df %>% 
    filter(Tests == "A") %>% 
    pull(Score)
  
  # Or keep it as a data frame with Participant info
  a_test_data <- df %>% 
    filter(Tests == "A")
  

Pro tip: Double-check that your Tests column has "A" exactly (no typos like lowercase "a" or extra spaces)—if not, adjust the filter condition to match.

This depends on whether you’re using base R’s hist() or ggplot2. Let’s cover both:

• Base R hist(): The first argument is the numeric vector of values you want to plot. For Test A scores, it would look like this:

  hist(a_scores, main = "Histogram of Test A Scores", xlab = "Score")
  

  Here, a_scores replaces "???"—it’s the numeric vector of scores you extracted earlier.
• ggplot2: You’ll map your score column to the x-axis in the aes() function. Using the filtered long-format data:

  library(ggplot2)
  
  df %>% 
    filter(Tests == "A") %>% 
    ggplot(aes(x = Score)) + # "Score" is what replaces "???" here
    geom_histogram(bins = 10, fill = "lightblue", color = "black") +
    labs(title = "Histogram of Test A Scores", x = "Score", y = "Frequency")
  

  Replace "Score" with your actual score column name if it’s different (e.g., Value, Result).

In repeated measures, each participant has multiple scores (one per test/time point). Here are two common scenarios:

• Histogram for each test/time point: Just filter the long-format data for each test (like we did for Test A) and plot individually. For side-by-side comparisons, use ggplot2’s faceting:

  ggplot(df, aes(x = Score)) +
    geom_histogram(bins = 10, fill = "lightblue", color = "black") +
    facet_wrap(~Tests) + # Creates a separate histogram for each test
    labs(title = "Histograms by Test (Repeated Measures)", x = "Score", y = "Frequency")
  

• Histogram of change scores: If you want to visualize how scores changed between tests (e.g., Test B minus Test A), first calculate the difference by temporarily pivoting back to wide format:

  # Using tidyr's pivot_wider (reverse of melt())
  library(tidyr)
  
  change_data <- df %>% 
    pivot_wider(names_from = Tests, values_from = Score) %>% 
    mutate(Score_Change = B - A) # Calculate change from Test A to Test B
  
  # Plot histogram of change scores
  hist(change_data$Score_Change, main = "Histogram of Score Change (B - A)", xlab = "Score Difference")
  

  This gives you a distribution of how much each participant’s score changed between measures.

内容的提问来源于stack exchange，提问作者Rei