PSYC09H3 Study Guide - Final Guide: Simple Linear Regression, Linear Regression, Mahalanobis Distance
Document Summary
Check off saved standardized values as variables . This creates a new column in your data set of standardized values. Each data point is recognized as a x,y point. Multivariate outlier is given a value that is the distance away from the other cases in the data. Distances are interpreted using a p<0. 001 statistical significance level. With the mahalanobis distance value, have to find the corresponding x2 value with degrees of freedom. Spss saves the squared mahalanobis distance for each case. Distance can then be compared to a chi-square distribution with degrees of freedom (df) P-value is then computed for the right tail area: if your outliers are due to data entry errors, delete them and re-run the whole analysis. Look for changes in significance in your model: if you think there is no reason why you can"t have that value, then leave it (just make note of it) Value close to 2 suggests an independence of observations.
