企业管理研究方法（英文版·第11版）唐纳德·库珀-Chap019.pptVIP

* Collinearity exists when two independent variables are highly correlated. Multicollinearity exists when more than two independent variables are highly correlated. If these conditions exist, the effects on multiple regression are damaging. The estimated regression coefficients may fluctuate widely. Correlations of .80 or greater should be dealt with using the techniques listed in the slide. Exhibit 19-2 contains a column called Collinearity Statistics that shows a variable inflation factor (VIF) index. This column is shown in the slide. The VIF is a measure of the effect of the other independent variables on a regression coefficient. Large values, usually 10.0 or more, suggest collinearity or multicollinearity. With the hybrid-mail example, multicollinearity is not a problem. * Exhibit 19-3 Discriminant analysis joins a nominally scaled criterion or dependent variable with one or more independent variables that are interval- or ratio-scaled. Once the discriminant equation is found, it can be used to predict the classification of a new observation. The is done by calculating a linear function of the form Use formula from page 578 Where Di is the score on discriminant function i. The dis are weighting coefficients; do is a constant. The Xs are the values of the discriminating variables used in the analysis. A single discriminant equation is required if the categorization calls for two groups. If three groups are involved, two discriminant equations are required. The most common use is to classify persons or objects into various groups or to analyze known groups to determine the relative importance of specific factors. Exhibit 19-3 provides an example based on the selection process for employees at KDL. The independent variables are years of prior work experience, GPA, and employment test scores. An algorithm determines how well these variables classify employees who have been judged as successful and unsuccessful. Exhibit 19-3 shows the standardized and unstandardi