Control Variable

[Kwong Ying Hui]

What is the difference between independent variable and control variable? Why most article related to corporate governance has control variable?
If my dependent variable is audit fee, independent variable is audit committee effectiveness and control variable is board effectiveness, how to use STATA to run the regression?

[Maarten Buis]

A control variable is just an indepedent variable, so just add it to your model as you would with any other indepedent variable.

[Carlo Lazzaro]

Kwong:
as far as your first question is concerned, you may find the followiing thread useful http://stats.stackexchange.com/quest...ther-variables.
Broadly speaking, control variables are indeed independent variables.
However, it is worth nothing that some prominent statisticians prefer the term "adjusting for" vs "controlling for" other predictors; this topic appeared some time ago on Statalist, too (please, see: http://hsphsun3.harvard.edu/cgi-bin/...ticle-267.html).
As far as your Stata question is concerned, assuming that audit fee is a continuous variable, I would point you to -help regress- and related entry in Stata 13.1 .pdf manual.
I would also recommend you to provide more details about which kind of independent variables are you dealing with (eg.: is committee effectiveness a continuous or categorical variable? If the latter is the right guess, is it an ordinal variable, such as poor, acceptable, outstanding committee effectiveness?)

[Kwong Ying Hui]

Maarten and Carlo,

Thank you for your response.
Dependent variable (Audit Fee) is a continuous data.
Independent variables (Number of Audit Committee, Number of Audit Committee Meeting, Number of Audit Committee Member with Financial Expertise and Number of Directors are Independent Non-Executive Director) - all continuous data
Control Variables (Number of Director, Number of Board Meeting, Duality, Number of Independent Non-Executive Director) - all continuous data except for duality is binary data

I am using Pooled OLS, Fixed Effect Model and Random Effect Model to run the regression. The result showed that Fixed Effect Model is fixed to my equation. The result showed Number of Audit Committee a

[Kwong Ying Hui]

I am using Pooled OLS, Fixed Effect Model and Random Effect Model to run the regression. The result showed that Fixed Effect Model is fixed to my equation. The result showed Number of Audit Committee are independent non-executive director, board independent and duality are significantly related to audit fees. However, when I include time effect and industry effect, all variables are not significantly related to audit fees. I do not understand why is it so.

[Carlo Lazzaro]

Kwong:
as per FAQ, please post what you typed and what Stata gave you back.
Please consider that advising on quantities without seeing (at least an excerpt of) them is very difficult (for me, at least).
At a very first glance, I would say that you're hunting for the best model to fit your data, but unfortunately, that approach is rarely successful (and even harder to justify).

[Nico Cuca]

Hi, i have a similar problem to Kwong.
I research the relationship between performance and variables of corporate governance. Most of the articles related to corporate governance consider different control variables (i.e. log total asset of firm).

In my case, I have ROE as dependent variable and others independent variable (number of member of board, % of independent administrator, number of committee, etc)
So, the question is: in a panel data regression is correct consider control variables, even if I run regression with fixed effect? Then, what is the command on stata for a panel data?

[Carlo Lazzaro]

Nico:
welcome to the list.
For the future please, start another thread. Thanks.
That said, in the light of the limited details you give, you may be interested in taking a look at -xtreg-.

[Nico Cuca]

Thanks Carlo, i wrote here because the topis is very similar.
I'm using this syntax: xtreg $ylist $xlist, fe where y is dependent variable, x ind variables but i don't know how consider the control / moderator variables

I don't know if my question is trivial but i hope that you can help me.

[Carlo Lazzaro]

Nico:
as replied above, control variables are simply independent variables, and should be treated in the very same way.
For more positive replies, please report what you typed and what Stata gave you back via Code delimiters (please, see FAQ about these topics). Thanks.