An event that is technically a count, but the scale of the process means we can treat it as continuous
E.g. immigration rate, unemployment rate, etc.

Outcome could have happened at most N times, our data measures how many times it did happen
E.g. “how many days per week…”, etc.

An event that has no (theoretical) upper limit, but tends to happen at a relatively low rate
E.g. individual fertility, number of friends, grocery stores in a neighborhood, etc.


(See also: geometric distribution for ‘time until occurrence’ types of models)

λ=5.0

λ=20.0

λ=80.0

Poisson with a large λ is closely approximated by a normal distribution