How To Calculate Mean Of Two Variables In Stata

Use this interactive calculator to compute the average of two variables, compare row means versus separate variable means, and generate the exact Stata code you can use in your own dataset. Enter two equal length series, choose a method, and instantly view results with a chart.

Expert Guide: How to Calculate Mean of Two Variables in Stata

Calculating the mean of two variables in Stata is a common task in data cleaning, descriptive analysis, and model preparation. Researchers often need to combine two test scores, average repeated measures, summarize a pair of survey items, or create a composite indicator from two related metrics. While the idea is simple, there are several valid ways to calculate the mean depending on what you actually want to measure. In practice, many mistakes come from confusing the average of each variable, the row wise mean across variables, and the grand mean of all values pooled together.

In Stata, the phrase mean of two variables can refer to at least three different operations. First, you may want the mean of variable A and the mean of variable B separately. Second, you may want a new variable for each observation that equals the average of A and B for that row. Third, you may want one single number that combines all observations from both variables. The correct method depends on the analytical question, the structure of your dataset, and how you want to handle missing values.

A reliable rule is this: if you are averaging two measurements for each case, use a row mean. If you are summarizing each variable across the sample, use Stata summary commands. If you need one combined benchmark across both variables, calculate a pooled mean after deciding how to treat missing observations.

What the mean represents in this context

The arithmetic mean is the sum of values divided by the number of values. If you have two variables named x and y, and each row corresponds to the same person, household, firm, or time period, then the row mean for observation i is:

(x_i + y_i) / 2

This is useful when the two variables measure the same construct on the same scale. For example, you might average pre test and post test scores only if your design specifically requires a combined score. You might also average morning and afternoon measurements, or two raters’ scores, if your methodology supports that decision.

Most common Stata methods

• summarize x y gives the mean of each variable separately.
• generate avg = (x + y)/2 creates a new variable with a simple row wise average, but missing values can cause missing output.
• egen avg = rowmean(x y) creates a row mean and is often preferred because it handles missing values more flexibly.
• egen stacked_mean = mean(z) can be used after reshaping or stacking data when you need a pooled average in a long format workflow.

Method 1: Calculate the mean of each variable separately

If your goal is simply to know the average of each variable across all observations, the easiest command is:

summarize x y

Stata will report the number of observations, mean, standard deviation, minimum, and maximum for both variables. This is ideal when you want to compare central tendency between two variables but do not need to combine them into one new measure. For example, if x is income in year 1 and y is income in year 2, separate means show how the sample average changed over time.

One advantage of using summarize is that Stata automatically handles missing values variable by variable. If some observations are missing on x but not on y, the reported means still use all available data for each variable. This is statistically efficient for descriptive work, though it means the sample size can differ across variables.

Method 2: Calculate a row wise mean of two variables

If you want each row to have a new average based on the two variables, there are two classic options. The first is a direct arithmetic formula:

generate avg_xy = (x + y) / 2

This works perfectly when both variables are nonmissing. However, if either x or y is missing for a case, Stata returns a missing value for avg_xy. That behavior is sometimes desirable and sometimes not. If you want the row mean to use available values instead, use egen:

egen avg_xy = rowmean(x y)

The rowmean() function calculates the average across nonmissing values in the listed variables. With only two variables, that means if one variable is present and the other is missing, Stata returns the nonmissing value as the row mean. If both are missing, the result is missing.

This is one of the most important practical distinctions in Stata. Analysts often write generate avg = (x+y)/2 without thinking about missing data, then wonder why many new values are missing. In applied research, egen rowmean() is often the safer tool for survey scales, repeated measurements, and administrative data where some partial information exists.

Method 3: Calculate one combined mean across both variables

Sometimes you need a single overall mean using all values from both variables. Conceptually, this is the average of the stacked dataset where x and y are pooled into one long vector. If both variables have the same number of valid observations and no missing values, the combined mean equals the average of the two variable means. If missing values differ, the result can change.

One transparent approach is to reshape the data from wide to long format, then run a standard mean calculation. Another approach is to use a generated expression after counting valid observations carefully. For many users, the cleanest logic is:

1. Reshape or stack the two variables into one column.
2. Run summarize or mean on that stacked variable.
3. Interpret the result as the pooled average across all nonmissing values.

Comparison table: common Stata choices

Goal	Recommended command	Missing value behavior	Best use case
Separate mean for each variable	summarize x y	Uses available data within each variable	Descriptive statistics and comparisons
Simple average of x and y for each row	generate avg = (x + y)/2	Any missing input gives missing result	Clean data with complete observations
Row mean using available values	egen avg = rowmean(x y)	Averages nonmissing values only	Scales, partial item response, repeated measures
Pooled mean across all values	reshape long or stack, then summarize	Depends on pooled nonmissing values	Grand average across variables

Worked example with real numbers

Suppose you have six paired observations from two classroom assessments:

Observation	Variable x	Variable y	Row mean
1	10	14	12.0
2	12	11	11.5
3	9	13	11.0
4	15	17	16.0
5	18	16	17.0
6	20	19	19.5

In this example, the mean of x is 14.0 and the mean of y is 15.0. The pooled mean of all 12 values is 14.5. The mean of the row means is also 14.5 because both variables have complete data and equal weight. This equality often leads users to assume all methods are identical, but they are not. Once missing values appear, these quantities can diverge.

Missing data example

Now consider a second case where one observation is missing on y:

Statistic	Value	Interpretation
Mean of x	14.0	Based on 6 valid x values
Mean of y	14.2	Based on 5 valid y values if one y is missing
Mean using generate (x+y)/2	13.6	Only rows with both values contribute
Mean using egen rowmean(x y)	14.0	Rows with one nonmissing value still contribute

This comparison shows why command choice matters. A simple arithmetic formula effectively performs listwise exclusion at the row level. By contrast, rowmean() uses all available row information. Neither method is universally correct. The best choice depends on your research design and whether a one item response is enough to represent the intended construct.

Recommended Stata workflow step by step

1. Inspect the variables with summarize x y and codebook x y.
2. Check missingness with misstable summarize x y.
3. Decide whether your mean should be separate, row wise, or pooled.
4. If building a composite score, choose between generate and egen rowmean() based on missing value rules.
5. Validate the new variable using summarize avg_xy and a few manual spot checks.
6. Document your choice in comments or a do file so that your analytic decisions are reproducible.

Useful Stata code patterns

* Separate means summarize x y * Strict row mean requiring both values generate avg_xy = (x + y)/2 * Flexible row mean with missing value support egen avg_xy2 = rowmean(x y) * Quick inspection list x y avg_xy avg_xy2 in 1/10 * Check missingness misstable summarize x y avg_xy avg_xy2

When to use generate versus egen

Use generate when your formula is mathematically explicit and you want complete control over its logic. It is fast, transparent, and ideal when complete data are required. Use egen when the calculation needs row wise functions, grouped summaries, or more tolerant handling of incomplete records. In many social science and health datasets, egen rowmean() is preferred because it aligns better with how composite scales are often scored.

Interpretation tips

• If the two variables are on different scales, standardize them before averaging.
• If one variable is more reliable or conceptually important, a weighted mean may be more appropriate than a simple mean.
• If your two variables represent different time points, averaging may hide meaningful change.
• If the variables are categorical codes rather than continuous measurements, the mean may not be meaningful.

Common mistakes to avoid

• Assuming (x+y)/2 and rowmean(x y) are identical under missing data.
• Combining variables that use incompatible units or coding directions.
• Forgetting to reverse code one variable before averaging survey items.
• Reporting the mean of a new average variable without explaining how missing cases were treated.
• Using averages where a sum score or indexed scale is more standard in the field.

How this calculator helps

The calculator above lets you paste two numeric series and instantly see several outputs that mirror the way analysts think in Stata. It reports the mean of each variable, computes row means for paired observations, and gives a pooled overall mean across all values. It also generates suggested Stata code using your chosen variable names. This is useful for checking data manually before running a do file, teaching introductory methods, or validating expected results.

Authoritative references for statistical practice

For deeper background on averages, data summaries, and official statistical methods, consult these trusted resources:

• National Institute of Standards and Technology handbook on engineering statistics
• UCLA Statistical Methods and Data Analytics Stata learning resources
• U.S. Census Bureau guidance on statistical interpretation

Final takeaway

To calculate the mean of two variables in Stata, first define the question precisely. If you want the average of each variable, run summarize x y. If you want a new average for each observation, use generate avg = (x + y)/2 when complete data are required, or egen avg = rowmean(x y) when available values should still count. If you need one pooled mean across both variables, combine the values conceptually or structurally and then summarize the pooled result. This distinction is simple, but it prevents many common analytical errors.

In professional workflows, clarity about missing values, comparability of scales, and reproducibility of code matters just as much as the arithmetic itself. Once those decisions are made, Stata provides straightforward tools to calculate the mean accurately and efficiently.

Tip: If you are preparing a publication or report, always state exactly how the average was constructed, including whether missing values were excluded listwise or handled with a row mean function.