I don't disagree with the case for a percent format. There is a
difference, however.
Many models are difficult for users to implement, even with a major
programming effort. New models are features that often trigger purchase
or upgrade decisions from users. StataCorp is a business, after all,
wanting to maximise sales!
Minor graphics or management features do not fall into either category.
It may be a nuisance that you have to multiply by 100 or spell out what
text you want as axis labels, but you can do either without any
intervention by StataCorp.
Of course, Stata users expect a variety of improvements, big and small,
but resources are finite. In a crunch I imagine that most users would
prefer StataCorp to make the impossible possible, rather than make the
awkward easy.
Nick
[email protected]
David Radwin
At 5:48 PM +0100 6/30/08, Nick Cox wrote:
>
>I did once suggest a percent format to StataCorp -- without eliciting
>any interest.
No interest from StataCorp, perhaps, but this suggestion has appeared
before on Statalist, so there is obviously some interest from Stata
users. I am another interested user.
Outside of certain academic disciplines, I suspect it is much more
typical to portray fractions in the form of percentages, such as "41%
of households do X" rather than the Stata graph label default "0.41
of households do X."
It seems to me that this feature would be simple to implement and
would appeal to a wider range of potential users than highly
specialized estimation methods with names like "fixed effects
instrumental variables nonlinear multinomial probit mixed models for
heteroscedastic censored data with bootstrapped standard errors."
It's not that the latter doesn't have its place, but it would be nice
for the rest of us to put percentages on graph labels without jumping
through hoops.
Maybe we just need to lobby StataCorp more?
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
