Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: conception confusion - "fixed effects" and time effect on data with time factor

From   Maarten Buis <>
Subject   Re: st: conception confusion - "fixed effects" and time effect on data with time factor
Date   Mon, 24 Oct 2011 10:25:38 +0200

On Fri, Oct 21, 2011 at 10:56 PM, House Wang <> wrote:
> I think I have learned something from your response. "models are
> supposed to be useful, not true." This is the most important of the
> discussion. Thank you very much.

Just to be sure: I am not claiming authorship. It is a loosely based
on George Box's "all models are wrong, but some are useful"

> For the model, I want to measure the maturity of an institution in
> manipulating rules, which are supposed to increase across years.
> Proximately, it is like a kid grow up. This is why I use year as a
> trend variable.

Two comments:
1) Calender time is wrong here. To answer such a question you need the
age of the institution. You said earlier that age is not in the
dataset, but if you have the date when the institution was founded and
the time of the interview, you can trivially add age to your dataset
by -gen age = interview_date - founding_date-. If you have only one
instition than this won't matter much. However, if you are following
multiple institutions with different founding dates, than this will be
very important.

2) If you think of a process like growing up or maturing of an
institution than that process will definately not be linear. So I
would suggest you look into things like splines (-help mkspline-) or
fractional polynomials (-help fracpoly-) to model age.

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index