Content
Summary
The binomial distribution models the count of successes in a fixed number of independent Bernoulli trials, denoted as . It requires a constant probability of success and failure , with outcomes restricted to integers from 0 to . Key metrics include the mean and standard deviation .
Intuition
Think of a binomial experiment as a series of yes-or-no questions asked repeatedly under identical conditions. The random variable X counts how many 'yes' answers (successes) you get. The term 'success' is just a label for the outcome you are tracking, not necessarily a good thing. Because you are counting whole events, X must be an integer. The distribution's shape depends on n and p; if p is 0.5, it is symmetric, but if p is near 0 or 1, it skews. We use the mean to predict the average outcome over many trials and the standard deviation to understand the typical spread or variability around that average.
Key equations
Notation indicating X follows a binomial distribution with n trials and success probability p.
Probability of failure, complementary to the probability of success.
Probability mass function for exactly x successes in n trials.
Expected value or mean number of successes.
Standard deviation measuring the spread of the distribution.
Derivation
- Identify a sequence of n independent trials.
- Confirm each trial has exactly two outcomes: success (prob p) and failure (prob q=1-p).
- Define X as the total count of successes, which must be an integer between 0 and n.
- Calculate the mean as the product of trials and success probability: .
- Calculate the variance as , leading to standard deviation .
Worked example
In a survey, 41% of adult workers have only a high school diploma. If you sample 20 workers, what is the expected number and standard deviation of such workers?
- Identify parameters: n = 20, p = 0.41.
- Calculate q: q = 1 - 0.41 = 0.59.
- Calculate mean: .
- Calculate standard deviation: .
Answer Mean = 8.2, Standard Deviation approx 2.20
Common mistakes
- Treating 'success' as a positive value judgment rather than the specific outcome being measured.
- Assuming the probability of success changes between trials; binomial trials must have constant p.
- Calculating probabilities for non-integer values of X, which is impossible for discrete counts.
- Confusing the binomial coefficient nCx with simple multiplication n*x.