COMPSCI C8 Lecture Notes - Lecture 1: Confounding, Scientific Control, Decimal Mark
19 Jan 2018
School
Department
Course
Professor
Document Summary
Associate does not mean causation: just because two things are happening at the same time does not mean there"s a direct relation. If take away all the factors to why eat ice cream or why earthquake happening then a connection may be drawn. In order to prove causation, need to get rid of confounding factors. Random controlled experiment: control the factors that go into experimental groups (can have confounding factor) Randomization: help get rid of the controlled factor even more. (limit the amount of selection between two groups, and have the same number of variants in both groups. Observational experiment: no interference, just observe and gather data. B: just answer the two questions w/o reading news. What can i answer: you can answer association/connection between the two factors, not cause and effect, you can"t answer whether ny times have an association with did you vote.
