 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: generating correlated data
At 12:32 PM 6/14/2009, Jeph Herrin wrote:
Try -drawnorm-
Whether you use -corr2data- or -drawnorm- depends on exactly what it
is you want to do. With -corr2data-, you get exactly the
correlation you specify, e.g. if you want a correlation of .6 the
data you create will have a correlation of .6. With -corr2data-, it
is like you have an entire population in your created data set. No
matter how many times you run the command, no matter what seeds you
use, the data set will always have the exact same means, standard
deviations and correlations.
With -drawnorm- you are sampling from a population with the given
correlation. So, if you specify a correlation of .6, the generated
data might have a correlation of .57, or .63, or whatever, i.e. it
won't be exactly .6 because of sampling variability. If you create
different samples with different seeds you will get results that are
at least slightly different each time.
One use for -corr2data- is when you have the published means,
correlations and standard deviations for a data set, but not the
original data itself. With -corr2data-, you could replicate the
published analyses, and do a limited amount of tweaking, e.g. adding
or deleting variables from the models. For examples, see pages 8-10 of
http://www.nd.edu/~rwilliam/stats1/OLS-Stata9.pdf
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/