This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 1 May 2, 2009 It is E370 Time!!! ‘ Announcements: Turn your prelab quiz in to your lab coach! ‘ Today’s Goals: You should be able to T discuss guidelines for interpreting your regression results. T explain how to test each assumption that must be checked for the linear regression. T describe the effect a violation of each assumption would have on the regression results. ‘ Next Time~~Ethics: Misdirection, Fudging and Downright Lies! ‘ Regression Diagnostics T The process of checking the accuracy and validity of your data and your regression. h Clean data h Assess validity of assumptions ‘ Before the regression, check the data. T Analyze each variable individually (ordered array, graph, descriptive statistics, etc.) T Note missing observations, if any. T Check all observations that calculate as outliers (> 2.5). h measurement accuracy • correct if possible • if not, live with it h applicability • delete if necessary T Confirm that the data set is complete and appropriate h eliminate all observations that do not have complete data. ‘ NOTE: retain eliminated observations for analysis if questions arise. ‘ ALWAYS report every change you make to your data. T Check the assumptions made for the model. T When basic assumptions are violated, regression results are not valid. 2 T The assumptions are: h the Y variable is related to the X variables linearly h X variables are not correlated to each other....
View
Full Document
 Winter '09
 STAFF
 Economics, Normal Distribution, Regression Analysis, Variance, multicollinearity, regression results

Click to edit the document details