MAST10011 Chapter Notes - Chapter 1: Confounding, Cohort Study, Field Trial
12 Nov 2018
School
Department
Course
Professor
Document Summary
Chapter 1 epidemiological studies and experimental design. Confounding variable: the possible cause of the disease outcome, which is related to the exposure. Control: validity forms a baseline for comparison. Randomisation: validity - treatments allocated randomly by a computer. Blocking/stratification: precision - randomised units are carried out separately in each block; make comparisons between similar units. Replication: precision reduces the impact of natural variation. Blinding: validity e(cid:374)sures the e(cid:454)pe(cid:272)tatio(cid:374)s of the su(cid:271)je(cid:272)t a(cid:374)d/or e(cid:454)peri(cid:373)e(cid:374)ter does(cid:374)"t i(cid:374)flue(cid:374)(cid:272)e the out(cid:272)o(cid:373)e. Balance: precision same number of units in each treatment group. Lurking variable: a confounding variable that is not observed. Explanatory variables: variable related to the response variable e. g. age, gender, blood pressure. Clinical trial: studying treatments for a specific disease. Phase 1: first trial of the drug on healthy humans, observations made. Phase 2: given to people with the condition to observe any changes. Phase 3: same as phase two except larger sample size.
