Go to Main Menu
Go to Power Search
Go to Topics Listing
Go to Methods Listing

Story Name: Alcohol and Tobacco
Story Topics: Consumer , Health
Datafile Name: Alcohol and Tobacco
Methods: Correlation , Dummy variable , Outlier , Regression , Scatterplot
Abstract: Data from a British government survey of household spending may be used to examine the relationship between household spending on tobacco products and alcholic beverages. A scatterplot of spending on alcohol vs. spending on tobacco in the 11 regions of Great Britain shows an overall positive linear relationship with Northern Ireland as an outlier. Northern Ireland's influence is illustrated by the fact that the correlation between alcohol and tobacco spending jumps from .224 to .784 when Northern Ireland is eliminated from the dataset.

This dataset may be used to illustrate the effect of a single influential observation on regression results. In a simple regression of alcohol spending on tobacco spending, tobacco spending does not appear to be a significant predictor of tobacco spending. However, including a dummy variable that takes the value 1 for Northern Ireland and 0 for all other regions results in significant coefficients for both tobacco spending and the dummy variable, and a high R-squared.

Image: Scatterplot of Alcohol vs. Tobacco, with Northern Ireland marked with a blue X.