Hey there! Let's work through this variable label issue together—this is a common gotcha when moving between Stata and R, so you’re not alone.

First: Ditch foreign for haven #

The foreign package is pretty outdated and doesn’t support importing Stata variable labels at all—that’s why labels(df) only gave you variable names, and you couldn’t find labels via str(). It just wasn’t designed to carry over that metadata.

Instead, use the haven package—it’s built specifically for reading/writing Stata, SAS, and SPSS files while preserving all the metadata like variable labels, value labels, and data types.

Step 1: Correctly import your .dta file with haven #

Make sure you’re using the right function (read_dta() with an underscore, not read.dta which is from foreign):

# Install haven if you haven't already
install.packages("haven")
library(haven)

# Import the data
df <- read_dta("your_file_path.dta")

Step 2: View variable labels #

Here are the easiest ways to check your labels:

For a single variable

Use haven’s dedicated var_label() function, or directly access the "label" attribute:

# Using haven's helper function
var_label(df$your_variable_name)

# Directly accessing the attribute
attr(df$your_variable_name, "label")

For all variables at once

Call var_label() on the entire data frame—it’ll return a named list where each entry is the label for the corresponding variable:

var_label(df)

Quick overview with labels included

Both str() and dplyr::glimpse() will show variable labels when using haven-imported data (since haven returns a tibble that preserves this info):

# Base R option
str(df)

# Tidyverse option (install dplyr if needed)
library(dplyr)
glimpse(df)

Why weren’t you seeing labels with haven? #

If you tried haven but still couldn’t find labels, here are a few possible fixes:

• You used the wrong import function: Double-check you ran read_dta() (not read.dta from foreign).
• Outdated haven version: Old versions might have bugs with metadata import—update it with update.packages("haven").
• You missed the attribute: When you ran attributes(df$var), the label should be under the "label" key. If it’s not there, re-import with the correct haven function.

A quick note on foreign #

Just to clarify: foreign::read.dta() doesn’t import variable labels at all. The labels() function in base R (or foreign) refers to value labels (labels for categorical variable levels), not the descriptive variable labels you’re looking for. So that’s why it only returned variable names for you.

Hope this gets you those variable labels you need!

内容的提问来源于stack exchange，提问作者laura bmw