POLI 30 Lecture Notes - Lecture 18: Observational Error, Test Statistic, Statistical Hypothesis Testing
Document Summary
If there is no relationship between x and y the real regression line will be flat, i. e. b=0 perfectly flat. However due to random sampling (and other) error, the line will never be. Our hypothesis test will ask: what"s the probability of seeing a pattern as strong as. The closer the points are to a line the more obvious the pattern. The more data the more obvious the pattern. Regression and inference relationship between x and y. In some contexts you could set it lower or higher, depending on what would be worse for you: a false positive or a false negative. Null: there is no relationship between x and y, so slope (b) is zero. Alternative: there is a relationship between x and y, so b does not equal zero. You can have more than one null and alternative if you have more than one independent variable.