SOC350H5 Lecture Notes - Lecture 3: Interval Ratio, Nominal Level, Categorical Variable
Document Summary
Normal dist has property that allows us to infer. You have to learn to diagnose problems to encounter with describing data. To be able to summarize data in an efficient way. Cant use ordinary sq root reg if your dependent variable is not interval ration. Agree/disagree (likehart scale) ordinal level data. Dichotomous can only take on two categories. Nominal level of measure when 0 is referent category (not ranked) As one variable goes up, y does something. Nominal go up and down the referent catagory is category that use to compare other category. Level of measurement matter because it matters what we can do with those variables. Cant find that avg of race can only find mode. These categorical data can be ranked just don"t know the exact difference. Eg working class, upper class, and middle class. Ranked and exact distance between two numbers is known. This is important because it has most importance.
