ANOVA/ANCOVA or multiple regression ? I'm lost :(

[yasmine yasmine]

Hello,

I have :

- balanced panel
- 1 continuous dependent variable Y
- 5 explanatory variables among which 2 are continuous and 3 are binary
- 5 control variables among which 4 are continuous and 1 is binary
which gives in totality 4 binary variables and 6 continuous

I have a problem of heteroscedasticity (so the Chow test of stability coefficients will reject H0) and no autocorrelation problem

My question is :

in this case i have to use a multiple regression ?
or
ANOVA to test the effect of the 4 binary variables on Y and ANCOVA to adjust the effect of the 6 continuous variebles on Y?

Thanks a lot

[Carlo Lazzaro]

jasmine (for the umpteenth time, lease note the strong preference for family names, too on this forum. Thanks):
You seem to ignore all the recommendations made in replying to your previous post (please, see http://www.statalist.org/forums/foru...need-your-help). However, they still apply, as your situation has not changed a bit.
Reposting the same query many times and ignoring replies, is the best way to reduce the chance to get replied in the near future. Sorry for this.