01 / 20
X Y +

Combining
Random Variables

Learn the rules for finding the mean and variance when you add, subtract, or scale random variables: the foundation of inference in statistics.

AP Statistics

What Is a Random Variable?

A random variable assigns a numerical value to each outcome of a random process. It is described by its probability distribution, which tells us what values it can take and how likely each value is.

\(\mu\) : Mean
The expected (average) value: \(\mu_X = E(X)\)
\(\sigma^2\) : Variance
Measures spread: \(\sigma_X^2 = \operatorname{Var}(X)\)
\(\sigma\) : Std Dev
Square root of variance: \(\sigma_X = \sqrt{\operatorname{Var}(X)}\)

Why Combine Random Variables?

In real life, outcomes often depend on more than one random quantity. Combining them lets us predict the behavior of totals, differences, and scaled measurements.

Total Scores
SAT Math + SAT Verbal: what are the mean and spread of the total?
Profit/Loss
Revenue minus Cost: how variable is the difference?

Linear Transformations: \(a + bX\)

When you shift by a constant \(a\) and scale by a constant \(b\), the mean and variance change predictably:

Core Rules: Sums and Differences

1
Means always add (or subtract):
\(\mu_{X+Y} = \mu_X + \mu_Y\)    and    \(\mu_{X-Y} = \mu_X - \mu_Y\)
2
Variances always add (if \(X\) and \(Y\) are independent):
\(\sigma_{X+Y}^2 = \sigma_X^2 + \sigma_Y^2\)    and    \(\sigma_{X-Y}^2 = \sigma_X^2 + \sigma_Y^2\)
3
Standard deviations do NOT simply add.
\(\sigma_{X+Y} = \sqrt{\sigma_X^2 + \sigma_Y^2}\)

Variance Always Adds

\(X\)
\(\sigma^2 = 4\)
\(Y\)
\(\sigma^2 = 9\)
\(X+Y\)
\(\sigma^2 = 4+9=13\)
+ = The combined distribution is wider: variances add up.

The Independence Requirement

The variance addition rule only works when \(X\) and \(Y\) are independent. If they are not independent, a covariance term appears:

Independent
\(\operatorname{Cov}(X,Y) = 0\), so the simple rule applies.
Dependent
\(\operatorname{Cov}(X,Y) \neq 0\). You cannot simply add variances.

\(X_1 + X_2\) vs \(2X\)

These look similar but are fundamentally different. One involves two independent observations; the other scales a single observation.

\(X_1 + X_2\) \(2X\)
Meaning Sum of two independent draws Double one single draw
Mean \(\mu + \mu = 2\mu\) \(2\mu\)
Variance \(\sigma^2 + \sigma^2 = 2\sigma^2\) \(4\sigma^2\)
Std Dev \(\sigma\sqrt{2}\) \(2\sigma\)

A filling line that deposits one statistic for a whole tray

A filling line loads bottles into trays while one statistic is recorded for each tray.A production line can combine two independent measurements or scale one measurement. Those operations can share a mean while producing different variances.

Double one draw or add independent draws

Two outcome distributions contrast doubling the same 7-or-13 draw with adding two independent draws.Suppose X is equally likely to be 7 or 13. Then 2X is 14 or 26 and has standard deviation 6. For independent X₁ and X₂, their sum is 14, 20 or 26 with probabilities 1/4, 1/2 and 1/4 and standard deviation √18. Both means are 20.

How to Apply the Rules

1
Identify the random variables and whether each is independent.
2
Write the combination as a formula: e.g., \(T = X + Y\) or \(D = X - Y\).
3
Find the mean: add or subtract the individual means.
4
Find the variance: add the individual variances (if independent).
5
Take the square root of the variance to get the standard deviation.

Example 1: Adding Two Tests

A student's math score \(M\) has \(\mu_M = 520\), \(\sigma_M = 115\). Their verbal score \(V\) has \(\mu_V = 480\), \(\sigma_V = 110\). Scores are independent. Find the mean and standard deviation of \(T = M + V\).

Example 2: Difference of Variables

A factory produces bolts with diameter \(B\): \(\mu_B = 10\,\text{mm}\), \(\sigma_B = 0.3\,\text{mm}\), and nuts with hole diameter \(N\): \(\mu_N = 10.2\,\text{mm}\), \(\sigma_N = 0.2\,\text{mm}\). Let \(G = N - B\) (the gap). Find \(\mu_G\) and \(\sigma_G\).

Example 3: Scaling Then Combining

Let \(X\) have \(\mu_X = 50\), \(\sigma_X = 8\), and \(Y\) have \(\mu_Y = 30\), \(\sigma_Y = 6\). \(X\) and \(Y\) are independent. Find the mean and standard deviation of \(W = 3X - 2Y + 10\).

Common Mistakes

✕ Adding Std Devs
You cannot add standard deviations directly. You must add variances, then take the square root.
✕ Subtracting Variances
For \(X - Y\), the variance is \(\sigma_X^2 + \sigma_Y^2\), not \(\sigma_X^2 - \sigma_Y^2\). Variances always add.
✕ Confusing \(X+X\) with \(2X\)
\(X_1 + X_2\) (two draws) gives variance \(2\sigma^2\). \(2X\) (one draw doubled) gives variance \(4\sigma^2\).

Summary of Rules

Operation Mean Variance (indep.)
\(a + bX\) \(a + b\mu_X\) \(b^2 \sigma_X^2\)
\(X + Y\) \(\mu_X + \mu_Y\) \(\sigma_X^2 + \sigma_Y^2\)
\(X - Y\) \(\mu_X - \mu_Y\) \(\sigma_X^2 + \sigma_Y^2\)
\(aX + bY\) \(a\mu_X + b\mu_Y\) \(a^2\sigma_X^2 + b^2\sigma_Y^2\)
\(X_1 + X_2\) (iid) \(2\mu_X\) \(2\sigma_X^2\)
\(2X\) \(2\mu_X\) \(4\sigma_X^2\)

Key Takeaways Before Practice

1
Means follow the operation: add means for sums, subtract means for differences.
2
Variances always add (for independent variables), regardless of whether you add or subtract.
3
Standard deviations do not add: compute variance first, then take the square root.
4
Constants added shift the mean but don't change variance. Constants multiplied are squared in the variance.

Worked practice: Practice Problems, Problem 4 — Applied

Work through this original lesson problem, then reveal the solution and illustrations.

A manufacturing process produces bolts with mean diameter μ_B = 10.00 mm and standard deviation σ_B = 0.05 mm, and nuts with mean inner diameter μ_N = 10.20 mm and standard deviation σ_N = 0.04 mm. Bolt and nut diameters are independent. An engineer defines the clearance as C = N − B (nut inner diameter minus bolt diameter). A bolt-nut pair is considered acceptable if the clearance is between 0.10 mm and 0.30 mm. (a) Find the mean and standard deviation of C. (b) Assuming C is approximately normal, find the probability that a randomly selected bolt-nut pair is acceptable. (c) A redesign reduces σ_B to 0.03 mm. Without recalculating the exact probability, explain how this change affects the distribution of C and the proportion of acceptable pairs.

Show answer and explanation

(a) [2 points] μ_C = μ_N − μ_B = 10.20 − 10.00 = 0.20 mm. Since B and N are independent, σ²_C = σ²_N + σ²_B = (0.04)² + (0.05)² = 0.0016 + 0.0025 = 0.0041. Thus σ_C = √0.0041 ≈ 0.0640 mm. (b) [2 points] Use σ_C = √0.0041 ≈ 0.064031242 mm without premature rounding. Then z_low = −0.1/√0.0041 ≈ −1.561737619 and z_high ≈ 1.561737619. Thus P(0.10 < C < 0.30) = 2Φ(1.561737619) − 1 ≈ 0.881650. About 88.2% of pairs are acceptable under the approximate normal model. (c) [2 points] Reducing σ_B decreases the total variance of C. The new σ²_C = (0.04)² + (0.03)² = 0.0025, giving σ_C = 0.05 mm. Since the mean clearance remains 0.20 mm (the center of the acceptable range), a narrower distribution concentrates more area within the acceptable interval, increasing the proportion of acceptable pairs.

Normal clearance model centered at 0.20 mm with standard deviation √(0.04²+0.05²) mm. The shaded acceptable interval 0.10 to 0.30 mm has probability about 88.2%.
Independent nut and bolt variation combine through added variances for their clearance, under the stated distribution assumptions.
Two normal clearance models centered at 0.20 mm, with standard deviations about 0.064 and 0.050 mm. The narrower model has a taller density and more probability in the acceptance interval.
Two normal clearance models centered at 0.20 mm, with standard deviations about 0.064 and 0.050 mm. The narrower model has a taller density and more probability in the acceptance interval.

Question 1 of 5

If \(X\) and \(Y\) are independent with \(\sigma_X = 3\) and \(\sigma_Y = 4\), what is \(\sigma_{X+Y}\)?

Question 2 of 5

Let \(X\) have \(\mu_X = 20\) and \(\sigma_X^2 = 9\). What are the mean and variance of \(Y = 5 + 3X\)?

Question 3 of 5

Independent random variables \(A\) and \(B\) have means \(\mu_A = 100\) and \(\mu_B = 60\), with variances \(\sigma_A^2 = 25\) and \(\sigma_B^2 = 16\). What is \(\operatorname{Var}(A - B)\)?

Question 4 of 5

A random variable \(X\) has \(\mu_X = 10\) and \(\sigma_X^2 = 5\). Two independent observations \(X_1\) and \(X_2\) are drawn. What is \(\operatorname{Var}(X_1 + X_2)\)?

Question 5 of 5

Which statement is always true for independent random variables \(X\) and \(Y\)?