如何利用ggplot()绘制multcomp包glht()计算得到的置信区间？

If you need to compute simultaneous confidence intervals for pairwise treatment group comparisons in R, the glht() function from the multcomp package is a reliable tool—especially when using Tukey contrasts to adjust for multiple testing. Here's a practical walkthrough:

Step 1: Fit the Linear Model #

First, we start by fitting a linear model to our dataset using lm():

# Fit the base linear model
model <- lm(Years ~ Attr, data = MockJury)

Step 2: Compute Tukey Contrasts & Confidence Intervals #

Next, we use glht() to generate simultaneous confidence intervals with Tukey contrasts. This method adjusts the intervals to maintain a specified family-wise confidence level (we’re using 95% here):

# Load the multcomp package
library(multcomp)

# Calculate Tukey contrasts and corresponding confidence intervals
tukey_results <- glht(model, linfct = mcp(Attr = "Tukey"))

# View the formatted confidence interval output
confint(tukey_results)

Sample Output #

Running the above code will produce results like this:

Simultaneous Confidence Intervals
Multiple Comparisons of Means: Tukey Contrasts

Fit: lm(formula = Years ~ Attr, data = MockJury)
Quantile = 2.3749
95% family-wise confidence level

Linear Hypotheses:
Estimate lwr upr
Average - Beautiful == 0 -0.3596 -2.2968 1.5775
Unattractive - Beautiful == 0 1.4775 -0.4729 3.4278
Unattractive - Average == 0 1.8371 -0.1257 3.7999

Key Interpretations #

• Family-wise confidence level: The 95% value means we’re 95% confident that all these intervals capture the true difference in means for their respective group pairs.
• Linear Hypotheses breakdown: Each row represents a pairwise comparison of group means, with the null hypothesis being that the difference between the two groups is 0.
  • If the interval (spanning lwr to upr) includes 0, we can’t reject the null hypothesis—indicating no statistically significant difference between those two groups at the 95% family-wise level.
  • For example, the interval for Average - Beautiful includes 0, so we don’t have evidence of a meaningful difference between these groups. The interval for Unattractive - Average just barely includes 0, suggesting a marginally non-significant difference.

内容的提问来源于stack exchange，提问作者Khashir