An investigator wishes to determine whether either of two new treatments for hair loss are better than placebo, and if so, how much better.

He randomizes 100 men with the same type of hair loss problem to either treatment I, II, or III (placebo). He then determines the number of new hairs on each man's head by measuring the number of hairs in the area of balding before and two months after beginning treatment.

Even before collecting the data he asks two basic questions of statistics:

1. Estimation: What measurement should be used to summarize the data found in this investigation?

2. Inference: How do we use the study data to investigate whether one or more of the treatments is really better than a placebo for men in the population from which the investigation's sample was selected?