[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: When to use Log (1+x) Transformation?

From   Maarten buis <>
Subject   Re: st: When to use Log (1+x) Transformation?
Date   Sun, 6 Apr 2008 21:00:08 +0100 (BST)

--- Song <> wrote:
> I plan to estimate the following model by 3SLS  (reg3).
> ln(Revenue in each country) = a0 + a1*time window + controls.
> ln(time window) = b0 + b1*(expected) Revenue in each country +
> controls.
> Time window is difference in launching time (in days) of a product
> between the US and international launch. Both revenue and time window
> are skewed, so I wanted to use log transformation. The problem is
> that time window includes a lot of zeros, so I am not sure what is
> the best method. 

One transformation you can use is the cube root:
gen cr_time_widow = time_window^(1/3)

An issue you might want to look into is that a zero time may be
qualitatively different from the other values, but I don't know how to
implement such a thing for a endogenous variable, except that I
recommend that you take a very good look at the residuals.

-- Maarten

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Yahoo! For Good helps you make a difference
*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index