如何按分组变量匹配对应权重计算加权均值与weighted_se？

方法1：动态提取对应权重（Rowwise方式） #

这种方法无需重构数据，直接在每行根据variable的值匹配对应的权重列：

library(diagis)
library(dplyr)

# 先清理variable的重复因子水平（避免匹配异常）
table_selection$variable <- factor(table_selection$variable, levels = unique(levels(table_selection$variable)))

# 动态匹配权重并计算统计量
result <- table_selection %>%
  rowwise() %>%
  # 根据variable拼接权重列名，提取对应权重值
  mutate(matched_weight = get(paste0(variable, "_pop_weights"))) %>%
  ungroup() %>%
  group_by(year, variable) %>%
  summarize(
    weighted_mean = weighted.mean(value, w = matched_weight, na.rm = TRUE),
    weighted_se = weighted_se(value, w = matched_weight, na.rm = TRUE),
    .groups = "drop"
  )

print(result)

方法2：重构权重数据为长格式后合并 #

通过将宽格式的权重列转为长格式，再与原数据匹配，逻辑更直观：

library(diagis)
library(dplyr)
library(tidyr)
library(stringr)

# 清理variable的重复因子水平
table_selection$variable <- factor(table_selection$variable, levels = unique(levels(table_selection$variable)))

# 将权重列转换为长格式：year + variable + 对应权重
weight_long <- table_selection %>%
  select(year, ends_with("pop_weights")) %>%
  pivot_longer(
    cols = -year,
    names_to = "variable",
    values_to = "matched_weight",
    # 去掉权重列名的后缀，匹配原variable值
    names_transform = list(variable = ~str_remove(., "_pop_weights"))
  )

# 合并数据后分组计算
result <- table_selection %>%
  select(year, variable, value) %>%
  left_join(weight_long, by = c("year", "variable")) %>%
  group_by(year, variable) %>%
  summarize(
    weighted_mean = weighted.mean(value, w = matched_weight, na.rm = TRUE),
    weighted_se = weighted_se(value, w = matched_weight, na.rm = TRUE),
    .groups = "drop"
  )

print(result)

关键说明 #

• 原数据的variable因子存在重复水平（如两个C_ha），提前清理能避免匹配出错
• 两种方法都能实现按variable匹配对应权重的需求：
  • 方法1更简洁，适合数据结构简单的场景
  • 方法2逻辑清晰，适合后续需要扩展权重使用的场景

内容的提问来源于stack exchange，提问作者Tom