Statistical Sciences 2035 Quiz: Include a dichotomous variable as a predictor

26 views3 pages

Document Summary

Smoking behavior is a dichotomous variable because (0=non-smoking, 1= smoking) To determine whether we should include this variable, first perform an independent samples t- test for testing the effect of smoking behavior on sbd (the new variable on the dependent variable): But it is possible to use regression analysis to test the significance of the difference between the 2 means: first: check scatterplot for the relationship between smoking behavior and blood pressure: Even though the scores of smokers (2nd column) are generally a little higher, there is still a lot of overlap. Second: draw best fitting straight line and produce regression equation: R^2=0. 06 6% explained variation by smoking yes/no. The line connects the 2 group means. Non-smoker group"s (x=0) best prediction is simply their sbd score: Regression coefficient = difference between the mean of group of smokers and mean of group of nonsmokers. Third: perform an analysis by regressing sbd on smoking behavior and compare results to previously performed t-test: