在ggplot散点图均值点添加横纵置信区间线的实现方法

首先需要明确：要实现不同条件下的置信区间方向差异，你的数据中需要有一个分组变量（比如命名为condition，包含两种条件标签，例如"A"和"B"）。下面是完整的实现步骤和代码：

1. 预先计算分组的均值与置信区间 #

先用dplyr按条件分组，计算每组的均值、标准误和95%置信区间：

library(dplyr)

# 计算分组统计量（替换condition为你实际的分组列名）
summary_stats <- table_all %>%
  group_by(condition) %>%
  summarise(
    post_mean = mean(post, na.rm = TRUE),
    pre_mean = mean(pre, na.rm = TRUE),
    # 计算标准误
    post_se = sd(post, na.rm = TRUE)/sqrt(n()),
    pre_se = sd(pre, na.rm = TRUE)/sqrt(n()),
    # 计算95%置信区间（大样本用1.96，小样本可换成qt(0.975, df = n()-1)）
    post_ci_low = post_mean - 1.96*post_se,
    post_ci_high = post_mean + 1.96*post_se,
    pre_ci_low = pre_mean - 1.96*pre_se,
    pre_ci_high = pre_mean + 1.96*pre_se
  )

2. 绘制带定向置信区间的散点图 #

在你的原有ggplot代码基础上，添加geom_segment分别绘制水平/垂直置信区间线：

library(ggplot2)

scatter_plot <- ggplot(table_all, aes(x = post, y = pre)) + 
  geom_point(color = "black", size = 12.5) +
  geom_point(size = 12, color = "#C0C0C0") +  
  labs(x = "Post", y = "Pre") +  
  theme_classic(base_size = 35) +
  lims(x = c(0, 3), y = c(0, 3)) +
  geom_abline(intercept = 0, linetype = "dashed", color = "#000000") +
  # 添加分组均值点（修正了原代码重复的color参数）
  geom_point(data = summary_stats, 
             aes(x = post_mean, y = pre_mean), 
             fill = "black", size = 8, shape = 21, color = "black") +
  # 为条件A添加水平置信区间线（y固定，x覆盖置信区间范围）
  geom_segment(data = filter(summary_stats, condition == "A"),
               aes(x = post_ci_low, xend = post_ci_high,
                   y = pre_mean, yend = pre_mean),
               color = "black", size = 1.5) +
  # 为条件B添加垂直置信区间线（x固定，y覆盖置信区间范围）
  geom_segment(data = filter(summary_stats, condition == "B"),
               aes(x = post_mean, xend = post_mean,
                   y = pre_ci_low, yend = pre_ci_high),
               color = "black", size = 1.5)  

print(scatter_plot)

关键说明 #

• 如果你的分组变量名称不是condition，或者分组标签不是"A"/"B"，请直接替换代码中对应的字符串；
• 若需要小样本的置信区间，将1.96替换为t分布临界值，比如qt(0.975, df = n()-1)（需在summarise中额外计算每组自由度）；
• geom_segment的size参数可调整置信区间线的粗细，按需修改即可。

内容的提问来源于stack exchange，提问作者Tk Shmk