RE: st: How to interpret the incidence-rate ratio (IRR) of the interaction terms between binary and continuous variable

[Kieran McCaul <kieran.mccaul@uwa.edu.au>]

...

You will probably get a quicker and more accurate reply to this question if you (1) submitted it as a new post rather than a reply to an already existing thread and (2) included the explicit Stata regression command(s) that you used to estimate these mortality declines.



-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Ubydul Haque
Sent: Tuesday, 17 September 2013 3:41 AM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: How to interpret the incidence-rate ratio (IRR) of the interaction terms between binary and continuous variable

Dear all,
My findings is 'Total malaria, severe malaria and malaria-associated mortality declined by 65%, 78% and 93%, respectively in 2012 from the base year 2008'. Please let me know how I can calculate 95% confidence intervals to convey the precision. Is there any way in STATA (I need code too) that I can calculate 95% CI from 65%, 78% and 93% reduction?
Thank you.

Regards
Ubydul

On Mon, Aug 19, 2013 at 12:53 PM, Somrat Punurai <somratp@hotmail.com> wrote:
> Dear All,
>
> Recently, I read the Stata tip#87 written by Dr. Maarten Buis and his 
> posted reply in http://www.stata.com/statalist/archive/2011-10/msg00938.html.
>
> I have some questions regarding the interpretation of incidence-rate ratio (IRR) from the negative binomial regression.  In particular, it is the incidence-rate ratio of interaction terms between the binary variable (dummy) and the continuous variables.
>
> In my study, the model specification of negative binomial regression 
> is
>
> Y = constant + D1 + b1X1 +  b2X2 + b3D1X1 + b4D1X2
>
> Numerical example is
> Y = -1.013 + -1.129 + -0.068X1 + -0.610X2 + 0.054D1X1 + 1.590D1X2
>
> Y = the number of mergers and acquisitioin deals
> D1 = the dummy variable taking on value 1 if developing countries, and 
> 0 advanced coungries
>
> X1 = the difference of corporate tax rate between a country pair  
> (Continuous variable)
> X2 = the presence of double tax treaties (binary variable, value 1 if 
> two countries has double tax treaties, and 0 otherwise)
>
> D1X1 = interaction term between the binary and the continuous variable
> D1X2 = interaction term between the binary and the binary variable
>
> My question is that
>
> 1.  Is it a right way to think that "when the nation is from developing nation", the IRR of the factor X1 is  exp(b1 + b3)  and "when the nation is from advanced nation", its IRR of factor X1 is exp(b1)?
>
> 2.  How can I interpret the IRR of b3 when the nation is from developing nation or when the nation is from advanced nation.
>
> 3.  Is it correct to think that when the nation is from developing nation, the IRR of factor X2 is exp(b2+b4) and "when the nation is from advanced nation", its IRR of factor X2 is exp(b2)?
>
>
> Best regards,
>
> Somrat Punurai
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/



--
Ubydul Haque
Post Doctoral Fellow
Johns Hopkins University
615, North Wolfe Street
Baltimore, Maryland
21218, USA

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

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