Statistical Sciences 2141A/B Lecture Notes - Lecture 26: Scientific Control, Statistical Hypothesis Testing, Probiotic
Document Summary
In this chapter we extend the concepts learned in the last chapter (confidence intervals and hypothesis testing) to making comparisons between two population distributions. The sample sizes m and n may not be equal, but experiments are often designed with equal sample sizes particular if a b. The experimenter wishes to determine if there is a difference between the two populations, often in. If a = b this may be sufficient for the experimenter to conclude that the populations are the same, although variances may also differ. Experimenters are often interested in comparing the effects of two different treatments or the response after treatment with the response after no treatment (treatment vs. control) A control group provides a standard against which a new treatment can be measured. Patients in a control group are often given a placebo. A randomized controlled experiment results when experimenters assign subjects to the two treatments in a random fashion.