# Ch. 3 Notes - STAT 200 (Lang Wu) - Lecture notes Lang Wu...

• Notes
• 15
• 100% (2) 2 out of 2 people found this document helpful

This preview shows page 1 - 3 out of 15 pages.

Lecture notes, Lang Wu, UBC 1 Chapter 3. Study Designs In the previous chapters, we worked with data that were already available for analysis. In this chapter, we discuss how to collect good data . This is important because poorly collected data often lead to biased and misleading results in analysis. 3.1. Some Basic Concepts Before we discuss how to design a study to collect reliable data, we will first introduce several crucial and fundamental concepts of statistics. We use a simple example to illustrate these concepts. Suppose a doctor wishes to evaluate the effectiveness of a new drug that reduces back pain. The general approach for this is to design a study to collect data, analyze the data using statistical methods, and then draw conclusions about the effectiveness of the drug. Since it is impossible for the doctor to collect data on all patients with back pain, a more realistic plan is to randomly select a small group of patients (e.g., 40) with back pain, and then collect data on only these patients. The doctor then randomly assigns the patients to one of two groups (this random assignment prevents unconscious bias that may arise if patients are aware of which group they are assigned to): a treatment group : patients in this group receive the new drug, a control group : patients in this group receive a placebo (a dummy pill that looks and tastes like the drug but has no active ingredient). Following this, the doctor measures the reduction in back pain in both groups and records the results. Finally, a statistican (or biostatistician) analyzes the data using statistical methods to determine if any differences between the two groups are due to random variations or if they are due to the effectiveness of the drug, in which case the differences are deemed statistically significant . In the above example, first, 40 patients are randomly selected . Then, they are randomly assigned to one of two groups. The first step helps ensure the 40 patients are representative of all patients. The second step is called randomization , and it eliminates possible sorting biases so that observed differences between the two groups can be at-
Lecture notes, Lang Wu, UBC 2 tributed to the effects of the drug, rather than other factors (such as gender or age). Note that randomization is not about how the sample is obtained, but about how the subjects in an already obtained sample are randomly assigned to different treatments or conditions. Important concepts A number of important statistical concepts are illustrated in the above exam- ple. These concepts are relevant to all statistical research studies, so it is important to understand them well: population : all individuals of interest in the study. In the example, this would be all patients with back pain.