CRIM 320 Lecture Notes - Lecture 9: Homicide, Lincoln Near-Earth Asteroid Research, Explained Variation
Document Summary
Conceptually, bivariate regression is very similar to correlation. Still concerned with associations but want to go beyond mere associations. Trying to build towards causality - specify a 1 dv and 1 or more ivs that explain our dv. Primary difference specification of dv and iv. But this subtle difference is critically important in terms of application and interpretation. Building a model that says gun ownership produces/ is related to "causes" (gun homicide rates) Use to make predictions and determine how good those predictions are. Idea regression line can be used to make predictions. We will look at ways of determining whether the predictions are "good" or not. The functional form of a predicted regression line. Y hat = b sub 0 + b1 x where: Slope is a measure of the effect of our iv on our dv. This tells you that any straight line that crosses our x/y axis (minimum y axis), can be represented by this formula.