The binomial model describes a random variable that counts the number of successes in a fixed number (n) of independent trials, where each trial has the same probability of success (p). On AP Stats, its mean is np and its standard deviation is √(np(1-p)).
The binomial model is what you use when you're counting successes in a repeated yes/no situation. Think of it as running the same coin-flip-style experiment n times and asking, "How many times did I get a success?" Each individual repetition is a Bernoulli trial, meaning it has exactly two outcomes (success or failure). Stack n of those trials together, keep them independent, keep the success probability p the same each time, and the count of successes follows a binomial distribution.
A quick way to check whether the model applies is BINS: Binary outcomes, Independent trials, fixed Number of trials, Same probability of success on every trial. If all four hold, the model is fully defined by just two parameters, n and p. From those you get everything else. The mean is μ = np and the standard deviation is σ = √(np(1-p)), straight from the CED (UNC-3.C.1). So if 200 items each have a 0.03 chance of being defective, you expect np = 6 defectives, with a standard deviation of √(200 × 0.03 × 0.97) ≈ 2.41 defectives.
The binomial model lives in Topic 4.11 (Parameters for a Binomial Distribution) in Unit 4: Probability, Random Variables, and Probability Distributions. It directly supports two learning objectives. AP Stats 4.11.A has you calculate the parameters (μ = np and σ = √(np(1-p))), and AP Stats 4.11.B has you interpret those values in context with units (UNC-3.D.1). That second part trips people up. Writing "μ = 6" earns less than writing "we expect about 6 defective items per sample of 200." The binomial model also sets up everything you do with proportions later. The sampling distribution of a sample proportion in Unit 5 is really just a binomial count divided by n, so getting this model down now pays off all the way through inference.
Bernoulli Trial (Unit 4)
A Bernoulli trial is one single success-or-failure attempt. The binomial model is just n Bernoulli trials stacked together, with the random variable counting how many succeeded. If you can't describe one trial cleanly, you can't set up the binomial.
10% Condition (Units 4-5)
When you sample without replacement, the trials aren't technically independent (each draw changes the pool). The 10% condition says that if your sample is less than 10% of the population, the dependence is small enough to ignore, so you can still use the binomial model. This same check reappears constantly in Unit 5 and in inference.
Sampling Distribution of a Sample Proportion (Unit 5)
A sample proportion p-hat is a binomial count divided by n. That's why the formulas for the mean and standard deviation of p-hat look like the binomial formulas with an n divided out. Unit 5 is built on top of this model.
Expected Value (Unit 4)
The binomial mean μ = np is just a shortcut version of expected value. Instead of summing every value times its probability, the binomial structure lets you multiply trials by success probability and get the same answer.
Multiple-choice questions test the binomial model in a few predictable ways. One classic gives you the mean and standard deviation and asks you to back-solve for n and p (for example, μ = 12 and σ = 3 means np = 12 and np(1-p) = 9, so 1-p = 0.75, p = 0.25, n = 48). Another classic describes a scenario and asks when the binomial formula would be inappropriate, which means checking the BINS conditions, especially independence and the 10% condition when sampling without replacement. On FRQs, binomial setups show up inside probability questions. You need to (1) verify the conditions, naming them explicitly, (2) calculate μ = np and σ = √(np(1-p)), and (3) interpret the results in context with units. A bare number with no context loses points. And if the conditions fail, say so and don't use the binomial formula anyway. Recognizing when a model does NOT apply is itself a tested skill.
Both models are built on independent Bernoulli trials with the same success probability p, but they count different things. The binomial model fixes the number of trials (n) and counts how many successes occur. The geometric model has no fixed n; it counts how many trials it takes to get the FIRST success. Quick test for the exam: if the problem says "out of 20 attempts, how many..." it's binomial. If it says "how many attempts until the first..." it's geometric. Also note the parameters differ. Binomial needs both n and p, while geometric only needs p.
The binomial model counts successes in a fixed number of independent trials where each trial has the same probability of success, and it is fully defined by two parameters, n and p.
Check the BINS conditions before using the model: Binary outcomes, Independent trials, fixed Number of trials, and the Same probability of success on every trial.
The mean of a binomial random variable is μ = np and the standard deviation is σ = √(np(1-p)); memorize both, since the formula sheet won't save you if you can't recognize when to apply them.
When sampling without replacement, trials aren't truly independent, so you must verify the 10% condition (sample size less than 10% of the population) before treating the situation as binomial.
Always interpret μ and σ in context with units, like "we expect about 6 defective items per batch of 200," because interpretation in context is its own learning objective (4.11.B).
If the binomial conditions fail, do not use the binomial formula; identifying when a model is inappropriate is a tested skill on its own.
It's the probability model for counting successes in a fixed number n of independent trials, each with the same success probability p. Its mean is np and its standard deviation is √(np(1-p)), which is exactly what Topic 4.11 tests.
Binomial fixes the number of trials and counts successes ("how many out of 20?"). Geometric counts trials until the first success ("how many tries until I make one?"). If the number of trials is set in advance, it's binomial.
Yes, independence is one of the required conditions. But if you're sampling without replacement, the 10% condition rescues you. As long as the sample is less than 10% of the population, you can treat the trials as approximately independent and still use the binomial model.
No. A Bernoulli trial is one single success-or-failure attempt. A binomial random variable is the count of successes across n Bernoulli trials. A Bernoulli trial is actually just a binomial with n = 1.
Set up two equations, np = mean and np(1-p) = variance (the standard deviation squared), then divide the second by the first to solve for 1-p. For example, if μ = 12 and σ = 3, then 1-p = 9/12 = 0.75, so p = 0.25 and n = 48. This exact setup appears in multiple-choice questions.
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.