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

st: reshaping data and count models

From   Maarten buis <>
To   stata list <>
Subject   st: reshaping data and count models
Date   Sun, 14 Dec 2008 16:51:52 +0000 (GMT)

Nikolaos Pandis had trouble posting to the statalist because he sent
the message as formatted text rather than plain text. I am rather busy
so I am just forwarding his message in the hope that one of you will
answer him.

-- Maarten

Somehow I cannot get my postings through to statalist and I was
wondering if you would please give me some advice on the following
I have counted the number of articles that show significant and non
significant results in 5 different journals, over a five year period.
The count data is further divided in 5-6 categories depending on the
subject of the article.

The data should look as follows:

Journal year issue category significant nonSign. ratio
A       2008 1     1        4           1        4
A       2008 2     2        5           2        2.5
A       2008 3     3        6           2        3
B       2008 4     2        4           0        .
B       2008 5     3        8           1        8

1. Would you please help me set the data in long format like above? I
cannot fit the data to show you how I have it exactly. It has the above
variables with the category var 1-6 in a wide format. I hope this is

2. The objective of the study, apart from posting the descriptive stats
which on their own will be interesting, would be to perform a
statistical analysis also. The analysis would determine the effect on
the count ratio (countsignificant/countnonsig) of the journal type,
year and subject (subject would be the most important for us). I was
thinking of some sort of regression for count data.

It is possible that a very small number of ratios would have zero for
denominator (how should I handle this?-Would it be better to use the
reverse ratio-countnonsig/countsignificant- so I get zero values for
the cases were the numerator is zero? The countsignificant  var does
not have any zero counts)  

3. Any references you might have in mind or links would be welcome.

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