HSS 3101 Lecture Notes - Lecture 26: Columbia University Medical Center, Confounding, Randomized Controlled Trial
24 Nov 2018
School
Department
Course
Professor
Document Summary
What if you know that a variable like sex is a confounder of your exposure and disease. Randomization: make sure that you select people completely randomly, this should make sure that due to randomness your confounders are balanced across the people in your experiment. Sex is related to my exposure and my outcome, but i am not interested in it. I could enroll a large number of people in my study. Draw out a little example on board. I worked on studies at columbia university medical center in manhattan nyc. We were interested in comparing young and old adults to look at the relationship between education and cognitive function. We recruited from the university for young people and local community centers from old adults. These issues are related to sampling biases. If you match based on bmi, you have to be careful because it is possible that.
