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

st: correlation between time-series of panel data

From   "Christian Winzer" <[email protected]>
To   <[email protected]>
Subject   st: correlation between time-series of panel data
Date   Wed, 23 Sep 2009 10:20:57 +0100

I have spent the past day searching (the archive, faqs, the web) but haven't
found a way to calculate the correlation coefficients between time-series
belonging to different categories in a panel data set.

The data has the format:

Category 	timestep 	value
A		1		5	
A		2		6
A		3		7
B		1		8	
B		2		8	
B		3		10

In this example I would thus calculate the correlation between the
timeseries of 'value' for category A and category B.
So one option would be to change the data format to the following: 

timestep 	valueA	valueB
1		5		8
2		6		8
3		7		10

Then the standard correlate valueA valueB works. 
Unfortunately I don't know how to do that transformation. 

I have thus tried the following direct solutions, but without success:
1) correlate value, by(category) -> 'option by() not allowed'
2) by category: correlate value -> only produces correlation of 'value'
within each category (=1), instead of across categories A and B
3) statsby, by(category): -> 'no; data in memory would be lost'

Thank you very much for your help,


*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index