如何用ggplot2调整森林图中优势比的自上而下显示顺序

要让森林图的模型从上到下按你指定的顺序排列，核心是手动控制分类变量的显示顺序，以下是两种可靠方案：

方案1：直接在ggplot中指定轴顺序（无需修改原始数据） #

如果用了coord_flip()，翻转后原x轴会转为y轴，直接在代码中添加scale_x_discrete()并传入目标顺序即可：

# 适配你的森林图代码示例
ggplot(your_data, aes(x = model_name, y = OR, ymin = lower_CI, ymax = upper_CI)) +
  geom_pointrange(size = 0.8) +  # 你的森林图视觉元素
  geom_hline(yintercept = 1, linetype = "dashed", color = "gray50") +
  coord_flip() +
  # 关键：手动指定从上到下的模型顺序
  scale_x_discrete(limits = c(
    "Unadjusted",
    "Adjusted for family factors",
    "Adjusted for family and self-functioning factors",
    "Adjusted for family, self-functioning, and socio-demographic factors"
  ))

注意：要保证limits中的字符串和你数据里model_name列的取值完全一致（包括空格、标点），否则对应模型会被过滤。

方案2：将分类变量转为有序因子（适合多次复用顺序） #

如果需要重复使用这个顺序，建议先修改数据，把模型列转为带指定levels的因子：

# 假设你的数据框为df，模型名称列是model_name
df$model_name <- factor(df$model_name,
  levels = c(
    "Unadjusted",
    "Adjusted for family factors",
    "Adjusted for family and self-functioning factors",
    "Adjusted for family, self-functioning, and socio-demographic factors"
  ),
  ordered = TRUE
)

# 后续画图会自动沿用因子定义的顺序
ggplot(df, aes(x = model_name, y = OR, ymin = lower_CI, ymax = upper_CI)) +
  geom_pointrange(size = 0.8) +
  geom_hline(yintercept = 1, linetype = "dashed", color = "gray50") +
  coord_flip()

如果翻转后顺序仍不符合预期，只需在scale_x_discrete()中添加limits = rev(levels(df$model_name))即可快速反转显示顺序。

内容的提问来源于stack exchange，提问作者keherder