I am running a regression in Stata.
As the dependent variable, I have the market share of smartphones (quarterly) for Apple and Samsung, and independent variables are Functional improvements and Design innovation (scored also quarterly).
My supervisor suggested that I have time fixed in order to account for the Christmas boost and I do not really understand how to do it.
And the second question is, am I capturing the interaction effect correctly?
So far I did...
xtset idcompany qdate
reg marketshare design function
and for interaction effect I did
gen designfunction=design*function
reg marketshare design function designfunction
and I got really good P values and R^2, but my coefficient for design*function is ( -.09) I am very curious how should I interpret it.
Does this all make sense? I am really new to Stata. I would really appreciate any help.
*To account for time fixed effect in Stata, you add simply :
i.qdate in your command.
reg marketshare design function i.qdate
Note: For you good P values. Unfortunately they may suffer from serial correlation. If errors are serially correlated, an increase in the error term in one time period affects the error term in the next and I suppose it's the case for the market share of smartphones. Try to check for this issue.
If you don't have a serial correlation problem, try also to check for heteroskedasticity. Huber Sandwich Estimator can solve this issue by adding the 'robust' option.
reg marketshare design function i.qdate, robust
Fo the interaction effect:
http://www.stata.com/meeting/germany12/abstracts/desug12_royston.pdf
A.A
Your data are quarterly. That means you should introduce a time variable t with t=1, t=2, t=3, t=4, t=5, ... .If you are thinking of a Christmas effect, you examine the series at multiples of 4 as t-value assuming t=1 is a first quarter. Seasonal index may be computed on quarterly basis. Significant variation amongst the indices is an indication of the time interaction with the series.
Hello Alexa Drk. Generally speaking, Statalist (https://www.statalist.org/forums/forum/general-stata-discussion/general) is a better place to post questions about how to do X using Stata.
Second, I would think that if you are using -xtset-, you would want to use -xtreg- rather than -regress-.
- https://www.stata.com/manuals/xtxtreg.pdf
Third, if you use the # or ## operators to include interactions (as in the file Aymen Ammari linked to), you'll be able to use -margins- and -marginsplot- to explore the nature of the interactions. See section 11.4.3 here:
- https://www.stata.com/manuals/u11.pdf
HTH.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software