Brief review of estimation and hypothesis testing

This chapter started with an introduction to the concepts of probability and random variable distributions. The role of probability is to assist in our ability to make statistical inferences. Test statistics are the numeric results of an experiment or study. The yardstick by which a test statistic is measured is how extreme it is. The term extreme in Statistics is used in relation to a value that would have been expected if there was no effect, that is, the value that would be expected by random chance alone. Confidence intervals provide an interval estimate for a population parameter of interest. Confidence intervals of (1 — a)% can also be used to test hypotheses, as seen in Chapter 8. 

The process of hypothesis testing is carried out using the following steps, which will be highlighted in subsequent chapters  

Drug lowered SBP Hq Drug had no effect on SBP Ha. Drug increased SBP 

Statistical inference is one way to use data to make a decision in the presence of uncertainty. The resulting decisions are not perfect. The commission of either a type 1 or a type II error can have significant impacts on drug companies, study participants, patients, and public health. Therefore, minimizing the probability that each might occur is an important part of the study design, including the manner in which data are analyzed and interpreted. 

---