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

Re: st: Creating new variables from variable values

From   Maarten buis <>
Subject   Re: st: Creating new variables from variable values
Date   Mon, 29 Dec 2008 21:50:08 +0000 (GMT)

--- Eric Uslaner <> wrote:
> I have a data set (European Social Survey) with several variables
> indicating the respondents' and parents' country of birth.  The
> values for these string variables (which I know that I can convert)
> are mostly letters (some are numbers and I am not sure what they mean
> but I can set them to missing).  

Not knowing what a code in your data means should make you very very
worried. It should certainly not be a reason to set the unknown codes
to missing. It should be a reason to go back to the codebook, and look
up what those codes mean. (you can probably get them from )

When I teach my students about how to deal with data, I always read a
section from the "Hitchhiker's guide to the galaxy" by Douglas Adams
where "the answer to the great question of Life, The Universe, And
Everything" turns out to be 42. I think that this is a nice
illustration of why any data on its own is absolutely worthless unless
you have access to the questions and the way the answers are coded
(i.e. the codebook). It is actually quite a long section of the book
that needs to be read, so it takes up quite some time in the lecture.
But I think it is worth it as this is a very important point (data
isn't magic, it is just lots of answers written down in a table) and it
is fun.

-- 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 N515

+31 20 5986715

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index