HLTB15H3 Lecture Notes - Lecture 8: Confounding, Hawthorne Effect
Document Summary
Theoretical bias - people want to be in the trial intervention - last week inclusion and exclusion. Because they mentally adapt to get the drug/counselling (whatever they want) so they might drop out due to expectations. If it is not randomized, then you cannot say causality. Therefore having a control group before and after non randomized won"t be effective. Also pre experimental uncontrolled non randomized study changes the outcome. Because there is no before and after, there is no background. If it is randomized, then it avoids confounding and say causality. In quantitative rct you don"t analyse until everything about the study design is closed because you may interact differently because you know things you shouldn"t. Sometimes you have to test feasibility because every study is different. In drug trials --> using triple blind is most appropriate. Triple blinding is the most valid and least causality. There are a lot of benefits of blinding.