# Video Resources

Short video resources will be added here as time permits.

**Lesson 6: Testing a Subset of Predictors in a Multiple Linear Regression Model.**(click fullscreen - lower right)

**Lesson 8: Additive model.**The Lesson8Ex1.txt dataset contains a response variable, Y, a quantitative predictor variable, X, and a categorical predictor variable, Cat, with three levels. We can code the information in Cat using two dummy indicator variables. A scatterplot of Y versus X with points marked by level of Cat suggests three parallel regression lines with different intercepts but common slopes. We can confirm this with an F-test of the two X by dummy indicator variable interaction terms, which results in a high, non-significant p-value. We then refit without the interaction terms and confirm using individual t-tests that the intercept for level 2 differs from the intercept for level 1 and that the intercept for level 3 differs from the intercept for level 1.

**Lesson 8: Interaction Model.**The Lesson8Ex2.txt dataset contains a response variable, Y, a quantitative predictor variable, X, and a categorical predictor variable, Cat, with three levels. We can code the information in Cat using two dummy indicator variables. A scatterplot of Y versus X with points marked by level of Cat suggests three non-parallel regression lines with different intercepts and different slopes. We can confirm this with an F-test of the two X by dummy indicator variable interaction terms, which results in a low, significant p-value.

**Lesson 12: Simpson's paradox.**An illustration of how a response variable can be positively associated with one predictor variable when ignoring a second predictor variable, but negatively associated with the the first predictor when controlling for the second predictor. The dataset used in the example is available is this file paradox.txt.