xtsum equivalents – ESSSSDA 2026 2J: Time Series and Collections of Them

This has been implemented in a host of user libraries, e.g. this

R Equivalents to Stata’s `xtsum`

There’s no single built-in R function that perfectly replicates xtsum’s overall / between / within decomposition, but here are the main options:

1. `plm` package (closest equivalent)

library(plm)

pdata <- pdata.frame(df, index = c("id", "time"))

# Panel dimensions (like xtdescribe)
pdim(pdata)

# Check between/within variation (flags invariant variables)
pvar(pdata)

# Summary of a panel variable
summary(pdata$y)

pvar() will warn you if a variable has no within or between variation — useful for the same diagnostics xtsum provides.

2. Custom `xtsum` function (most faithful replica)

This is the go-to approach shared widely in the R/Stata community:

xtsum <- function(data, var, id, time) {
  library(dplyr)
  
  data <- data %>% rename(y = {{var}}, i = {{id}}, t = {{time}})
  
  overall <- data %>%
    summarise(mean = mean(y, na.rm=T), sd = sd(y, na.rm=T),
              min = min(y, na.rm=T), max = max(y, na.rm=T), n = n())
  
  between <- data %>%
    group_by(i) %>%
    summarise(y_bar = mean(y, na.rm=T)) %>%
    ungroup() %>%
    summarise(mean = mean(y_bar), sd = sd(y_bar),
              min = min(y_bar), max = max(y_bar), n = n())
  
  within <- data %>%
    group_by(i) %>%
    mutate(y_within = y - mean(y, na.rm=T) + mean(data$y, na.rm=T)) %>%
    ungroup() %>%
    summarise(mean = mean(y_within, na.rm=T), sd = sd(y_within, na.rm=T),
              min = min(y_within, na.rm=T), max = max(y_within, na.rm=T), n = n())
  
  bind_rows(overall, between, within) %>%
    mutate(component = c("overall", "between", "within")) %>%
    select(component, everything())
}

xtsum(df, var = income, id = person_id, time = year)

3. `panelr` package

library(panelr)

panel_df <- panel_data(df, id = person_id, wave = year)
summary(panel_df)

panel_data objects give structured summaries with some between/within awareness, though less detailed than xtsum.

4. Quick manual approach with `dplyr`

library(dplyr)

# Between variation (person means)
df %>% group_by(id) %>% summarise(y_mean = mean(y))

# Within variation (demeaned)
df %>% group_by(id) %>% mutate(y_within = y - mean(y))

Summary

Goal	R approach
Panel dimensions	`plm::pdim()`
Flag invariant vars	`plm::pvar()`
Full between/within decomposition	Custom `xtsum` function above
Panel-aware data structure	`panelr::panel_data()`

The custom function is the most commonly recommended route when you need output that maps directly onto Stata’s xtsum table.

R Equivalents to Stata’s xtsum

1. plm package (closest equivalent)

2. Custom xtsum function (most faithful replica)

3. panelr package