如何修改plot_function()指标及让plot_model()显示均值与标准差？

一、修改plot_function()中的指标 #

针对自定义的plot_function()，根据需求调整核心逻辑：

• 更换指标变量：如果要切换展示的变量，直接修改函数内ggplot()的aes()映射变量名。示例：

  # 原函数（示例）
  plot_function <- function(data) {
    ggplot(data, aes(x = Condition, y = Old_Var)) + geom_line()
  }
  # 修改后
  plot_function <- function(data) {
    ggplot(data, aes(x = Condition, y = New_Var)) + geom_line()
  }
  

• 更换统计量：如果要替换展示的统计指标（如从均值换成中位数/标准差），修改统计量计算代码。示例：

  # 原函数计算均值
  plot_function <- function(data) {
    sum_data <- data %>% group_by(Condition) %>% summarise(stat = mean(Var))
    ggplot(sum_data, aes(x = Condition, y = stat)) + geom_col()
  }
  # 修改为计算标准差
  plot_function <- function(data) {
    sum_data <- data %>% group_by(Condition) %>% summarise(stat = sd(Var))
    ggplot(sum_data, aes(x = Condition, y = stat)) + geom_col()
  }
  

二、让plot_model()展示均值与标准差 #

sjPlot::plot_model(type = "pred")默认展示模型预测值的置信区间，要改为展示原始数据的均值±标准差，可按以下两种方法操作：

方法1：在现有plot_model绘图基础上叠加标准差 #

1. 先从原始数据中按分组计算均值和标准差：

   library(dplyr)
   # 替换"Response_Var"为你的模型实际响应变量名
   summary_data <- Focus_full %>%
     group_by(Condition, Group) %>%
     summarise(
       mean_y = mean(Response_Var, na.rm = TRUE),
       sd_y = sd(Response_Var, na.rm = TRUE),
       .groups = "drop"
     )
   

2. 在原有绘图代码后叠加误差线图层：

   plot_Focus <- plot_model(fit, type = "pred", terms = c("Condition", "Group")) +
     coord_cartesian(xlim = c(0.5, NA), ylim = c(0,5), clip = "off") + 
     theme_tq() + scale_colour_tq() + scale_fill_tq(light) + 
     labs(title = "", y = "", x = "") +  
     ggplot2::labs(colour = "Group") + scale_color_manual(values=c('Red','Black')) +
     # 新增：叠加均值±标准差的误差线
     geom_pointrange(data = summary_data,
                     aes(x = Condition, y = mean_y, 
                         ymin = mean_y - sd_y, ymax = mean_y + sd_y,
                         color = Group),
                     position = position_dodge(width = 0.2)) +
     theme_apa()
   

方法2：直接用原始统计数据绘图（更直接） #

如果不需要模型预测线，可直接基于计算好的均值和标准差绘图：

library(ggplot2)
library(dplyr)
library(ggthemes) # 用于theme_tq、theme_apa

summary_data <- Focus_full %>%
  group_by(Condition, Group) %>%
  summarise(
    mean_y = mean(Response_Var, na.rm = TRUE),
    sd_y = sd(Response_Var, na.rm = TRUE),
    .groups = "drop"
  )

plot_Focus <- ggplot(summary_data, aes(x = Condition, y = mean_y, color = Group)) +
  geom_line(position = position_dodge(width = 0.2)) +
  geom_pointrange(aes(ymin = mean_y - sd_y, ymax = mean_y + sd_y),
                  position = position_dodge(width = 0.2)) +
  coord_cartesian(xlim = c(0.5, NA), ylim = c(0,5), clip = "off") +
  theme_tq() +
  scale_color_manual(values=c('Red','Black')) +
  labs(colour = "Group", y = "", x = "") +
  theme_apa()

内容的提问来源于stack exchange，提问作者user19785597