Acknowledgments
Download 3.57 Mb.
|
Figure 6. Complete-Second Order Model 2.3. CATEGORICAL VARIABLES In all of the methods mentioned prior we have assumed that the variables being analyzed have all been numerical values or quantitative variables. However, it is possible that the chosen variables are categorical (qualitative). In this particular project, there are several categorical variables—sex, race, sport played, chosen major, and home state. Nevertheless, there are methods used to incorporate these categorical variables into the analysis. DICHOTOMOUS VARIABLE First consider the simple case. Suppose the categorical variable being examined has two possible categories, such as male or female. This type of variable with two categories is called a dichotomous variable. With dichotomous variables, one must assign a dummy or indicator variable x. This dummy variable has two possible values, zero or one, and indicates which category is applicable for any chosen variable [2]. This concept is best explained through an example. Example: Suppose that it has been discovered that annual salary is dependent on the number of years of experience and whether or not the employee has a college degree. Let the dependent variable y = the annual salary and the independent variable  the number of years of experience. Since the presence of a college degree is a categorical variable, let
Take for instance the liner regression model,  .
Thus, the mean value of the annual salary depends on whether the employee has a college degree: mean salary =  when  (doesn’t have a degree)
Download 3.57 Mb. Share with your friends:
|