Topic 4.8 Mean and Standard Deviation of Random Variables: calculate and interpret the mean (expected value), variance, and standard deviation of a discrete random variable from its probability distribution.

What this topic is asking

The College Board (Topic 4.8) wants you to calculate and interpret the mean (expected value), variance, and standard deviation of a discrete random variable from its probability distribution, and to read the mean as a long-run average.

The mean as a weighted average

Expected value generalizes the ordinary average. An ordinary mean weights every data value equally; the expected value weights each possible value by how likely it is. So values that occur more often pull the mean toward them more strongly. The result is the value the long-run average of the random variable settles on, by the law of large numbers, exactly the "fair price" of a game or the average defects per item over a long production run.

Variance and standard deviation

The structure mirrors Unit 1's standard deviation but with probabilities as weights. You find the mean, take each value's deviation from it, square the deviation, weight by the probability, sum to get the variance, and square-root to get the standard deviation. The squaring (as in Unit 1) prevents positive and negative deviations from cancelling and emphasizes larger departures; the square root returns the spread to the original units so it is interpretable.

Interpreting expected value correctly

The most tested idea is what the expected value means. It is a long-run average, not a prediction of any single outcome, and it need not be an attainable value: a random variable that takes only whole numbers can have a fractional mean like $0.7$ defects, which simply says that over many items the defects average $0.7$ each. In a game of chance, the expected payout is the average you would win per play over a very long run, which is why it is the right notion of a "fair" stake: a game is fair if the expected net gain is $0$. Misreading the expected value as "the most likely outcome" or "what will happen next" is a classic error, the expected value of a single die roll is $3.5$, which never appears on any single roll. Holding the long-run-average meaning firmly lets you interpret expected value in any context, payouts, defects, waiting times, and it is the meaning the binomial mean $np$ (Topic 4.11) and all later expected values inherit.

Why these parameters matter downstream

The mean and standard deviation of a random variable are the parameters that the rest of the course revolves around. Combining random variables (Topic 4.9) is entirely a set of rules for how means and standard deviations behave when variables are added or scaled. The binomial and geometric distributions come with shortcut formulas for exactly these two parameters. And in the sampling-distribution unit, a sample mean and a sample proportion are themselves random variables whose mean and standard deviation (the standard error) drive every confidence interval and significance test. So the calculations here, weighted-average mean, weighted squared-deviation variance, are not isolated arithmetic but the foundation of the inferential machinery to come. Mastering them now, and especially interpreting the expected value as a long-run mean, pays dividends across half the course.

Try this

Q1. $X$ has $P(1) = 0.4$, $P(2) = 0.4$, $P(5) = 0.2$. Find $E(X)$. [2 points]

• Cue. $E(X) = 1(0.4) + 2(0.4) + 5(0.2) = 0.4 + 0.8 + 1.0 = 2.2$.

Q2. Explain why an expected value of $2.2$ can be valid even though $X$ never equals $2.2$. [1 point]

• Cue. The expected value is a long-run average over many trials, not a single outcome, so it need not be a value $X$ can actually take.