You can combined 3 dummy variables as long as the three dummies are no perfectly collinear

Ibrahim,

It is possible to have many dummy variables. I attach an example where 12 dummy variables were used for each month. See R. Pindyck and D. Rubinfeld, "Econometric Models & Economic Forecast" Third Edition, 1991 McGraw-Hill for this example. One must be careful to avoid the dummy variable trap.

Thank you dear friends Lall B. Ramrattan Peterson K Ozili

Yes you can have 3 dummy variables in a regression.

You can use dummies as much as you wish (just be careful to your degree of freedom). Besides, even dummies could be colinear so you have to ensure their independence with adequate test (not the conventional FIV).

You can use as many variables (this includes dummies) as your degrees of freedom allows. Just remember that, if you include an intercept, for each categorical variables, you should include only "nº of categories -1" dummies.

I hope it helps.

Your answers are highly helpful. Thank you all dear José-Ignacio Antón Georgi Kiranchev Chiraz Karamti Anna Ebers Broughel

You are most welcome

I agree with all the answers. I would just like to comment on the page shown by Lall B. Ramrattan. There is a wrong conclusion in the sentence just after the table of seasonal coefficients (" The seasonal coefficients are all significant at the 5 percent level, suggesting that the seasonal variations in the volume of certificates of deposit are quite important ") due to a specification error in that example taken from R. Pindyck and D. Rubinfeld, "Econometric Models & Economic Forecast" Third Edition, 1991 McGraw-Hill. I had noticed a long time ago and even documented in my book in French "Méthodes de prévision à court-terme". Indeed, the authors omit the intercept to avoid a collinearity problem. But since their 12 dummies are all multiplied by PI (personal income), the sum of their "dummies" is PI, not 1. A more correct specification shows that only one of the seasonal coefficients is significantly different from 0.

Dear Ibrahim.

I agree with all the earlier contributions, you can include 2 to 3 dummy variables but note that addition of more explanatory variables can affect your final results.

Also be sensitive towards your degree of freedom.

Regards

Great thanks dear Kehinde Mary Bello

Dear all,

you can use as much as you whish but be carefull to the multicolinearity

best