STAT 100 Lecture Notes - Lecture 6: Penicillin, Dementia, Random Assignment

• School: Simon Fraser University
• Department: Statistics
• Course: STAT 100
• Professor: Gamage Perera

Document Summary

A block is a group of subjects (set up at the beginning of the experiment) such that subjects within the block are similar. The random assignment of treatments to subjects is done separately within each block. Note: matched pairs designs are simply block designs with two subjects in each block (or pair ). Block designs are a way to control for outside variables (besides the treatments) by accounting for those variables in the experiment (when we determine the blocks). Block designs and matched pairs designs allow us to make more precise conclusions about differences among treatments than we can with completely randomized designs. These advanced designs let us account for outside variables, and thus differences across blocks are separated from the differences among the treatments. We can also examine the differences among treatments separately within each block. This is similar to using stratified samples. Example 1: a new surgical method is proposed to benefit parkinson"s patients.