Hello,
I have a question regarding generating tables of baseline characteristics for publication (whether I should breakdown results by outcomes, or by predictors).
BACKGROUND:
Suppose I have the outcomes Foreign Car, Domestic Car, and No Car.
I also have the predictor variables Price $40,000.
I would then want to generate the following table:
Foreign Domestic No Car
Price $40,000
Now, I have 2 options:
1. I can use -by(outcomes), sort: ci price40-, which will give me the breakdown of price within each outcome. In this case, the percentages within each group add to 100%.
2. I can use -by(price), sort: ci foreign domestic nocar-, which will give me the breakdown of price across each outcome, thus allowing me to compare differences between each outcome. In this case, the percentages across groups add to 100%.
QUESTION 1:
Is there a better way to do option 2? If I do it the way stated above, I would have to run that command for each of my predictor variables, which is cumbersome.
QUESTION 2:
In what scenarios would each of the 2 methods for breaking-down-the-population be used best? Do they each have their place, or is one preferred over the other?
Thank you!
_________________________________________________________________
Create a cool, new character for your Windows Live™ Messenger.
http://go.microsoft.com/?linkid=9656621
*
* 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/
