Topic 4.11 Parameters for a Binomial Distribution: calculate and interpret the mean and standard deviation of a binomial random variable using the shortcut formulas, and describe how the distribution's shape depends on n and p.

What this topic is asking

The College Board (Topic 4.11) wants you to calculate and interpret the mean and standard deviation of a binomial random variable using the shortcut formulas, and to describe how the distribution's shape depends on $n$ and $p$.

The shortcut formulas

The mean formula is intuitive: if each of $n$ trials succeeds with probability $p$, you expect about $np$ successes, just as $50$ free throws at $70\%$ should produce about $35$ makes. The standard deviation formula is less obvious but follows from adding the variances of $n$ independent trials (Topic 4.9): each trial contributes variance $p(1-p)$, so $n$ of them give $np(1-p)$, and the square root is the standard deviation.

Why the formulas work

Seeing the binomial as a sum of $n$ independent one-trial variables connects this topic back to Topic 4.9 and explains both shortcuts at once: means add to give $np$, and (because the trials are independent) variances add to give $np(1-p)$. This is also why you must not add standard deviations: the $\sqrt{\cdot}$ comes from adding variances and rooting at the end. Understanding the derivation, rather than just memorizing the formulas, makes them stick and clarifies why independence is required.

The shape of a binomial distribution

The third part of the topic is shape. With $p = 0.5$ the distribution is perfectly symmetric, because successes and failures are equally likely. With $p < 0.5$ successes are rarer, so the distribution is right-skewed (piled up at low counts with a tail toward high counts); with $p > 0.5$ it is left-skewed. But as the number of trials $n$ increases, the distribution becomes more bell-shaped and symmetric regardless of $p$, a preview of the normal approximation. The standard rule of thumb (the large-counts condition) is that the binomial is approximately normal when both $np \ge 10$ and $n(1-p) \ge 10$, that is, when you expect at least about $10$ successes and $10$ failures. When that holds, you can use the mean $np$ and standard deviation $\sqrt{np(1-p)}$ with the normal model to estimate probabilities, exactly the bridge that makes Unit 5's sampling distribution of a proportion approximately normal. So the shape discussion is not decorative: it tells you when the normal machinery of Topic 1.10 may be reused on a binomial count.

Using the parameters to judge surprise

A powerful exam move combines the binomial parameters with the z-score idea from Topic 1.10 to decide whether an observed count is surprising. If the number of successes you saw is several standard deviations from the mean $np$, it is far out in the tail and unlikely under the assumed $p$, evidence that $p$ may not be what was claimed. In the defects example, $20$ defective parts is about $3.2$ standard deviations above the expected $10$, which is extreme, so a $5\%$ defect rate looks doubtful. This is the same surprise-measuring logic from Topic 4.1, now made quantitative with $\mu = np$ and $\sigma = \sqrt{np(1-p)}$, and it is the direct ancestor of the significance test for a proportion in Unit 6. Being able to compute the two parameters, interpret the mean as a long-run expected count, and use the standard deviation to gauge how unusual an outcome is, is the full skill set Topic 4.11 builds.

Try this

Q1. For $X \sim B(100, 0.3)$, find the mean and standard deviation. [2 points]

• Cue. $\mu = np = 30$; $\sigma = \sqrt{np(1-p)} = \sqrt{100(0.3)(0.7)} = \sqrt{21} \approx 4.58$.

Q2. State the condition under which a binomial distribution is approximately normal. [1 point]

• Cue. The large-counts condition: both $np \ge 10$ and $n(1-p) \ge 10$ (at least about $10$ expected successes and $10$ expected failures).