Multivariate regression with continuous outcome variable and categorical control variables

[Kalpa Padma]

I am new to Stata and am trying to run a logistic regression to test how, in a large sample population, regular bedtimes are associated with several categorical variables. These include verbal test restuls (in quartiles - thus categorical), ethnicity, academic attainment of parents, economic activity of parents, poverty level of the household and whether or not the parent is single (dummy variable). I have run univariate regressions with all these variables and regularity of bedtimes separately but when I put it together, I get "invalid variable" for the single parent dummy variable and ethnicity. I put an "i" in front of the categorical variables but not the 0/1 dummy variables.
Can anyone advise me? I must be making a basic error

[Carlo Lazzaro]

Kalpa.
welcome to the list.
Three remarks about your query:
- I'm not clear why you go -logistic. when you have a continuous dependent variable. But if you follow that road, please consider that logistic depvar allows two outcome only (yes/no; dead/alive and so on). If you have >2 outcomes, you should go -mlogit- (if hey are not ordered) or -ologit- (if they are ordered).
- when you write multivariate it should be read multiple (multivariate implies that you have more than one dependent variable, whereas multiple refers to more than one predictor. It's true that the first term is sometimes used as a synonym for the second one, but their technical meanings are different);
- as you might have read from the FAQ, the best way to get helpful replies is to post what you typed and what Stata gave you back (rather than descibing your procedure and results). Thanks.