A binomial random variable X counts the number of successes in a fixed number n of independent trials, where each trial has exactly two outcomes (success or failure) and the same probability of success p. Its mean is np and its standard deviation is √(np(1-p)).
A binomial random variable is what you get when you run the same two-outcome experiment a set number of times and count how many successes you see. Think of flipping a coin 20 times and counting heads, or sampling 15 voters and counting how many support a candidate. Each individual trial is a Bernoulli trial, and the binomial variable is just the running tally of successes across all n of them.
For a setting to be binomial, you need four things, often remembered as BINS: Binary outcomes (success/failure), Independent trials, a fixed Number of trials, and the Same probability of success p on every trial. When those hold, the probability of exactly x successes is P(X = x) = (n choose x) · pˣ(1-p)ⁿ⁻ˣ, which the CED calls the binomial probability function. The parameters are clean too. The mean is μ = np and the standard deviation is σ = √(np(1-p)). You can also estimate binomial probabilities with a simulation using a random number generator, which is exactly how the AP exam often frames it.
Binomial random variables live in Unit 4 (Probability, Random Variables, and Probability Distributions), specifically Topics 4.10 and 4.11. The CED expects you to estimate binomial probabilities from simulation data (4.10.A), calculate exact probabilities with the binomial formula (4.10.B), calculate the parameters μ = np and σ = √(np(1-p)) (4.11.A), and interpret those values in context with units (4.11.B). This isn't a one-unit concept, though. The binomial setting is the foundation for everything you do with proportions later. The sampling distribution of a sample proportion in Unit 5, and all the proportion inference in Units 6, is built on counting successes in a sample, which is a binomial idea wearing different clothes.
Keep studying AP Statistics Unit 4
Bernoulli Trial (Unit 4)
A Bernoulli trial is one single success-or-failure attempt. A binomial random variable is just n Bernoulli trials stacked up and counted. If you understand one coin flip, the binomial variable is 20 of them with a tally sheet.
Geometric Distribution (Unit 4)
Same setup, different question. Binomial fixes the number of trials and counts successes. Geometric keeps trying until the first success and counts how many trials it took. Topic 4.12 sits right next to 4.10 because the exam loves making you pick the right one.
10% Condition (Units 4-6)
Sampling without replacement technically breaks independence, so trials aren't perfectly binomial. The 10% condition says that if your sample is less than 10% of the population, the trials are close enough to independent that the binomial model still works. This check follows you all the way into inference.
Expected Value (Unit 4)
The binomial mean np is a special case of expected value with a shortcut formula. Instead of summing x·P(x) over every outcome, you multiply the number of trials by the success probability. Expect 20(0.4) = 8 successes in 20 trials with p = 0.4.
Multiple-choice questions hit binomial random variables from several angles. You might read a simulation table (say, 800 repetitions of a binomial setting with n = 20 and p = 0.25) and estimate a probability from the relative frequencies, or critique a simulation design, like whether 'generated value ≤ 0.7 means success' correctly models p = 0.7. Other stems ask you to compute the mean and standard deviation of a transformed variable like Y = 2X - 10, which combines binomial parameters with the linear transformation rules from earlier in Unit 4. On FRQs, the classic moves are checking the binomial conditions (BINS), calculating a probability like P(X ≥ 3) using the formula or complement rule, and interpreting μ = np in context with units. A bare number with no context loses points, so always say something like 'on average, about 8 of the 20 customers would choose the new flavor.'
Both involve independent two-outcome trials with the same success probability p, which is why they get mixed up. The difference is what's fixed and what's counted. A binomial random variable has a fixed number of trials n and counts how many successes occur (possible values 0 to n). A geometric random variable has no fixed n; it counts which trial produces the first success (possible values 1, 2, 3, ... with no upper limit). Quick test: if the problem says 'until' or 'first,' think geometric. If it gives you a set number of trials, think binomial.
A binomial random variable counts the number of successes in n independent trials that each have the same success probability p and only two possible outcomes.
Check the BINS conditions before using the binomial model: Binary outcomes, Independent trials, fixed Number of trials, and Same probability of success.
The probability of exactly x successes is P(X = x) = (n choose x) · pˣ(1-p)ⁿ⁻ˣ, the binomial probability function.
The mean of a binomial random variable is μ = np and the standard deviation is σ = √(np(1-p)), and both must be interpreted in context on FRQs.
Binomial counts successes in a fixed number of trials, while geometric counts the trial number of the first success; the words 'until' or 'first' signal geometric.
You can estimate binomial probabilities with a simulation, where the relative frequency of an outcome across many repetitions approximates its true probability.
It's a random variable X that counts the number of successes in n independent trials, where each trial has two outcomes and the same success probability p. For example, the number of heads in 20 coin flips is binomial with n = 20 and p = 0.5.
Binomial fixes the number of trials and counts successes (values 0 through n). Geometric counts which trial gives the first success, so it has no upper limit. Their formulas differ too: binomial means are np, while geometric means are 1/p.
Yes, independence is one of the required conditions. When you sample without replacement, trials aren't truly independent, but the 10% condition (sample less than 10% of the population) lets you treat them as approximately independent and still use the binomial model.
The mean is μ = np and the standard deviation is σ = √(np(1-p)). So for X ~ Bin(20, 0.4), the mean is 8 successes and the standard deviation is √(20 · 0.4 · 0.6) ≈ 2.19 successes.
Yes, and the AP exam tests this directly (it's learning objective 4.10.A). Run many repetitions with a random number generator, then use the relative frequency of an outcome as your probability estimate. If P(X = 6) came up in 110 of 500 trials, your estimate is 0.22.