summarize: descriptive statistics and missing values
Calculate counts, means, standard deviations, minima, and maxima for the full dataset or separately within groups.
Complete example
The synthetic data contain 24 rows. One value of write is deliberately missing so the output shows that summarize uses the nonmissing observations of each variable.
clear
import delimited "data/student_scores.csv"
summarize math read write
by race: summarize math
summarize math, detailOutput from stats.camp
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
math | 24 57.375 7.654567 43 72
read | 24 56.16667 6.349712 44 68
write | 23 57.65217 6.69877 46 70
write reports 23 observations while math and read report 24. Its mean and standard deviation are calculated from those 23 nonmissing values.
Summarize variables by group
Prefix the command with by race: to calculate the same statistics separately for each observed race value:
by race: summarize mathrace Variable Obs Mean Std. dev. Min Max 1 math 6 50.50 6.5345 43 60 2 math 6 60.50 5.1284 53 67 3 math 6 54.33 5.3914 47 62 4 math 6 64.17 6.0470 56 72
Each group contains six observations. Grouping changes the calculation sample for each row of output; it does not replace the dataset in memory.
Current behavior of detail
The parser currently accepts summarize math, detail, but the current stats.camp renderer still returns only the basic count, mean, standard deviation, minimum, and maximum row:
Variable | Obs Mean Std. dev. Min Max
-------------+---------------------------------------------------------
math | 24 57.375 7.654567 43 72
Percentiles, variance, skewness, and kurtosis from Stata's detailed display are not yet produced. The compatibility table marks this command partial until that output is implemented.
Checks and limits
- Compare counts across variables before combining them in a model.
- Inspect minima and maxima for impossible values or coding sentinels.
- Remember that a plausible mean can conceal outliers or a multimodal distribution.
- Use a supported table or graph command when the five-number output is not enough.
Use the t-test guide for mean comparisons or the rank-sum guide for an independent-groups rank test. See the current summarize entry before relying on option-level parity.
Related guides
Run this exact example
Clone the example repository in stats.camp and open the exact do-file shown on this page. No stats.camp account is required.
Run the grouped summary example