To make a statistical inference, one uses a *statistic* to infer what the value of the *parameter* from the population should be.

ex. To determine whether a filling machine in a cereal manufacturing plant is working properly, every hour a random sample of boxes of cereal that it filled are weighed. The boxes should have a filled weight of 20 oz. Twenty -five boxes are weighed and have an average of 19.6 oz.

Parameter = 20 oz. Statistic = 19.6 oz.

Suppose you took another sample of size 25 from the boxes that were just filled, would you expect that its average weight would be 19.6 oz?