2 Some Tests Based on the Binomial Distribution

Definition 2.1 (Binomial distribution). A random variable \(X\) counting the occurrences of some outcome in a fixed number \(n\) of observations is Binomial, written \(X\thicksim B(n,P)\), if

*
each observation falls into one of just two categories, success or failure;
*
the \(n\) observations are independent;
*
the probability \(P\) of success is the same for every observation.

The tests of this chapter — the sign test, the binomial test and the quantile test — all reduce the data to a count of successes and so all rest on this one distribution. The reduction is what makes them distribution-free: whatever the shape of the population, the count of observations falling on one side of a hypothesised value is binomial, and nothing further need be assumed.

Questions on this section

Stuck on something here? Ask below and it stays attached to this topic.