01:220:102 Lecture Notes - Lecture 2: Selection Bias

• School: Rutgers University
• Department: Economic
• Course: 01:220:102
• Professor: Barbara Crockett

Document Summary

Scienti c method: developing models that explain some part of world; testing those models using data 2 see how closely model matches what we actually observe. 2 important features of models: they aren"t exact, they generate predictions that can be tested w/ data. Positive correlation: both change in same direction (directly related) Negative correlation: change in opposite directions (inversely related) Causation: when 1 thing directly affects another. Positive concave: increasing at a decreasing rate. Curves w/ maximum: relationship w a maximum. Curves w/ minimum: relationship w a minimum. Why isn"t correlation same thing as causation: Omitted variables: if we ignore something that contributes to cause and effect, then that something is an omitted variable. A correlation might not make sense until the omitted variable is added. Reverse causality: reverse causality is when there is cause and effect, but it goes in opposite direction as what we thought. Controlled experiment: subjects randomly put into treatment (something happens)