Calculate Standard Deviation Between Two Random Variables

Calculate the mean, variance, and standard deviation of the sum or difference of two random variables using their expected values, standard deviations, and correlation.

Formula used:
Var(X + Y) = sigma_x^2 + sigma_y^2 + 2rho sigma_x sigma_y
Var(X – Y) = sigma_x^2 + sigma_y^2 – 2rho sigma_x sigma_y
SD(Z) = square root of Var(Z)

How to Calculate Standard Deviation Between Two Random Variables

When people ask how to calculate the standard deviation between two random variables, they are usually trying to measure the spread of a new variable formed by combining two others. In practice, that often means finding the standard deviation of X + Y or X – Y. This matters in finance, engineering, quality control, operations research, social science, and experimental design because combined uncertainty behaves differently from single-variable uncertainty.

The most common mistake is to assume that standard deviations simply add or subtract directly. They do not. Standard deviation is based on variance, and variance depends not only on the size of each variable’s spread, but also on the relationship between the variables. That relationship is captured by covariance or, more commonly in calculator settings, by correlation.

The Core Formula

If you know the standard deviation of random variable X, the standard deviation of random variable Y, and their correlation rho, then the variance of the combined variable is:

• For a sum: Var(X + Y) = sigma_x^2 + sigma_y^2 + 2rho sigma_x sigma_y
• For a difference: Var(X – Y) = sigma_x^2 + sigma_y^2 – 2rho sigma_x sigma_y

Once you have the variance, take the square root to get the standard deviation:

• SD(Z) = sqrt(Var(Z))

This is the central idea behind the calculator above. You enter the mean of each variable if you want the combined expected value, but for standard deviation itself, the key inputs are each variable’s standard deviation and their correlation.

Why Correlation Changes the Answer

Correlation measures the degree to which two random variables move together. It ranges from -1 to 1:

• rho = 1: perfect positive relationship
• rho = 0: no linear relationship
• rho = -1: perfect negative relationship

For sums, positive correlation increases variance because large values of X tend to occur with large values of Y. Negative correlation reduces variance because one variable tends to offset the other. For differences, the pattern reverses: positive correlation often reduces the spread of X – Y, while negative correlation can increase it.

If X and Y are independent, then correlation is 0 and covariance is 0. In that special case, the formulas simplify to Var(X + Y) = sigma_x^2 + sigma_y^2 and Var(X – Y) = sigma_x^2 + sigma_y^2.

Step-by-Step Example

Suppose you have two random variables:

• X has mean 50 and standard deviation 10
• Y has mean 30 and standard deviation 15
• The correlation between X and Y is 0.25

If you want the standard deviation of Z = X + Y, follow these steps:

1. Square each standard deviation: 10^2 = 100 and 15^2 = 225
2. Compute the covariance term from correlation: 2 x 0.25 x 10 x 15 = 75
3. Add the parts: 100 + 225 + 75 = 400
4. Take the square root: sqrt(400) = 20

So the standard deviation of X + Y is 20. The mean of X + Y is simply 50 + 30 = 80.

If instead you wanted Z = X – Y, the variance would be:

1. 100 + 225 – 75 = 250
2. sqrt(250) is approximately 15.81

This example shows why correlation matters so much. With the same two variables, the standard deviation of the sum and the standard deviation of the difference can be quite different.

Comparison Table: How Correlation Changes Combined Standard Deviation

The table below uses real calculations with sigma_x = 10 and sigma_y = 15. It shows how the result changes across different correlation assumptions.

Correlation rho	Var(X + Y)	SD(X + Y)	Var(X – Y)	SD(X – Y)
-1.00	25	5.00	625	25.00
-0.50	175	13.23	475	21.79
0.00	325	18.03	325	18.03
0.50	475	21.79	175	13.23
1.00	625	25.00	25	5.00

This comparison makes the logic concrete. When the variables move together strongly, adding them magnifies variability and subtracting them cancels variability. When they move in opposite directions, the opposite happens.

What If You Only Know Covariance?

Sometimes a problem gives covariance instead of correlation. In that case, use:

• Var(X + Y) = Var(X) + Var(Y) + 2Cov(X, Y)
• Var(X – Y) = Var(X) + Var(Y) – 2Cov(X, Y)

Correlation and covariance are related by:

• Cov(X, Y) = rho sigma_x sigma_y

The calculator above uses correlation because many students and professionals know the relationship in standardized form, especially when comparing variables measured in different units.

Why Means Are Still Included

The mean does not directly affect the standard deviation of a sum or difference, but it does matter if you want the full distribution of the new variable. For example:

• E(X + Y) = E(X) + E(Y)
• E(X – Y) = E(X) – E(Y)

This is useful for planning and forecasting. You might not only want to know how much uncertainty exists, but also where the center of the combined outcome lies.

Normal Distribution Context

If the two variables are jointly normal, then the sum or difference is also normally distributed. In that case, standard deviation becomes especially useful because you can estimate the probability that the combined variable falls within certain ranges. The following percentages are standard statistical benchmarks used in normal models.

Distance from Mean	Approximate Share of Values	Interpretation
Within 1 standard deviation	68.27%	Most common outcomes fall here
Within 2 standard deviations	95.45%	Typical range for planning and quality checks
Within 3 standard deviations	99.73%	Extreme outcomes become rare

These percentages help translate a calculated standard deviation into practical decisions. If your combined variable is approximately normal, then the standard deviation tells you how wide a realistic operating range may be.

Common Use Cases

• Portfolio risk: The volatility of a two-asset portfolio depends on the standard deviations of both assets and their correlation.
• Measurement systems: Total error from two instruments depends on whether their errors are independent or related.
• Project schedules: Combined duration uncertainty depends on the relationship between tasks.
• Manufacturing: Differences between part dimensions require proper variance propagation, not direct subtraction of standard deviations.
• Testing and assessment: Composite scores often depend on variance and covariance structure across sections.

Frequent Errors to Avoid

1. Adding standard deviations directly. You should add variances and the covariance adjustment, then take a square root.
2. Ignoring correlation. Even moderate correlation can materially change the answer.
3. Confusing covariance with correlation. Correlation is unit-free; covariance is not.
4. Using an impossible correlation value. Correlation must stay between -1 and 1.
5. Forgetting the sign in X – Y. The covariance term flips sign for differences.

Interpreting the Result

A larger standard deviation means the combined variable is more spread out. A smaller standard deviation means the combined outcome is more stable or concentrated. By comparing the result across different assumptions about correlation, you can evaluate best-case, worst-case, and realistic scenarios.

For example, if two process measurements tend to move together, the spread of their sum may be much wider than expected under independence. On the other hand, if you are studying the difference between two highly correlated signals, the resulting standard deviation may be surprisingly small because the shared movement cancels out.

Expert Tip: Sensitivity Analysis Matters

If you are not certain about the exact correlation, calculate the result at several plausible values such as -0.5, 0, and 0.5. This provides a sensitivity band around the final standard deviation and often reveals whether your analysis is robust. In real-world work, uncertainty in correlation estimates can matter just as much as uncertainty in the original standard deviations.

Recommended References

For deeper statistical theory and validated formulas, review these authoritative sources:

• NIST Engineering Statistics Handbook
• Penn State STAT 414 Probability Theory
• University of California, Berkeley Statistics Resources

Bottom Line

To calculate the standard deviation between two random variables, first define the combined variable you care about, usually X + Y or X – Y. Then use each variable’s variance plus the covariance adjustment. If covariance is not given, use correlation and the product of the standard deviations. Finally, take the square root of the resulting variance.

The calculator on this page automates the arithmetic, but the key insight is conceptual: the spread of a combined variable depends on both individual variability and the degree to which the variables move together. Once you understand that, standard deviation calculations between two random variables become much easier to interpret and apply correctly.

Educational note: This calculator is intended for statistical estimation and learning. Always verify assumptions such as independence, linear correlation, and distribution shape before using the result in high-stakes decisions.