Ok this might be a lame question but aren’t they both produce functional results, i.e. if X then 1 and if Y then 0? I also don’t understand what is “Measuring Independent Variables with Error” under the model misspecification section… Thanks!
True, they’re both dichotomous (i.e. 1/0) variables. The obvious answer is that in QDVs, you’re using them as dependent variables and in DVs, they’re INDEPENDENT variables. With QDVs, you can’t use OLS regressions for a number of reasons - for one, you get predicted values that are outside the (0,1) interval. There are others, but I can;t recall my econometrics. Hence, yo have to use a probit or logit model. On the other hand, with Dummy variables, you are using the dummy variable as an INDEPENDENT variable (i.e. a predictor). In essence, the coefficient on the DV captures the difference in mean value of your dependent variable between the two categories (i.e. for when X=1 vs. when X=0), whole correcting for the influence of the other X variables. So, DVS are on the right hand side and QDVs are on the left hand side.
Thanks busprof very insightful feedback…