SOAN 2120 Lecture Notes - Lecture 18: General Linear Model, Contingency Table, Categorical Variable
Document Summary
Today: recoding, anova with more than 2 categories, cross-tabulation with chi squared test. Recoding revisited: many of the variables in the dataset do not require recoding, however, when you use categorical variables. Limit your categories to three or four (i. e. , religion) The more categories you have the more you will need to interpret. Now that we have our recoded variable we can examine the relationship between martial status and earnings. These tables have to be at the bottom of the paper ***** very important. For your statistical analyses: first: decide whether it is necessary to recode or select cases, proceed with your analyses in two stages, descriptive statistics, statistical analysis involving two variables. Engage in a study where you are 50/50 you don"t know. Examine the relationship between gender and smoking: first: open spss, and open the dataset, second: pick the variables, gender: dvsex, whether respondent ever smoked: n2.