st: Metareg: meta-regression and squared term

["Maurizio La Rocca" <m.larocca@unical.it>]

Please, I would like a support in my meta-regression analysis concerning a 
financial management topic.

I'm investigating the moderator role of ownership, the percentage of shares 
that managers own of a firm, on the relation between debt (as explanatory 
variable) and firm's value (my dependent variable). The theoretical 
literature suggests the existence of complementarity between Debt and 
Ownership in affecting firm's value. I want to verify if this is true and if 
ownership amplifies or reduces the effect of Debt on Firm's Value.
I'm using the metareg command. I calculated the effect size of the effect of 
Debt on Value (Zd-w) and this is my dependent variable in the 
meta-regression. Then I use a Dummy Ownership (D_O), that is a dummy equal 
to 1 if the study account for ownership or 0 otherwise, to see if the effect 
of Debt on Value is affected by the inclusion in the model of the variable 
Ownership.

Here my question.

Due to the fact that in some paper it is also included the squared term of 
Ownership, because it is possible a non linear relation between Ownership e 
Value, it is methodologically correct to include in the meta-regression also 
a Dummy Ownership Squared (D_O2) that is a dummy equal to 1 if the variable 
ownership squared is present in the study to account for a quadratic 
functional form or 0 otherwise. Theoretically, the effect of ownership on 
the relation between debt and value can be different for low or high level 
of ownership . thus, if I include in the meta-regression D_O and D_O2  I 
should account for this . is it true?

Thanks a lot in advance for any suggestion.

Maurizio La Rocca
Assistant Professor in Business Economics, Management and Finance
University of Calabria - Italy
m.larocca@unical.it
www.mauriziolarocca.it


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/