Nick,
I'm afraid it's not the case of several softwares : R/Splus, Statistica,
SPSS, GAUSS, Maple and maybe others
I don't think they have missed up your clear example. We are just talking
about graphing (making a zoom on a subset with respect with all the data
without making a substantial additionnal work)
I do completely agree with your latent & important remark, the researcher
had to explain carefully & precisely what is beeing plotted
by the way,
>> in other words the image you get is not a subset of the full image.
I think we are always getting a subset of the full image, but for sure I
must be wrong..
Best regards
Naji
That is your distinction, not Stata's.
As far as Stata is concerned, either
a data point is on a graph or it isn't,
and -if- is the way to control that.
-line- plots are awkward. If you
omit data, what gets left behind
will get joined up. That's all.
What might have been included but
is not is not taken account of.
You won't get slicing whereby a line is cut
in the way you might want. In
other words the image you get
is not a subset of the full image.
One way to see this:
sysuse auto, clear
gen order = _n
line price order
line price order if price < 10000
Compare the two graphs carefully.
In the case of density functions,
there are good arguments for both
square root and log scales.
Nick
[email protected]
Nassar
>
> I don't want to exclude data, but to focus the graph window
> on some interval
> of the data.
> The yvariable is not directly observed (a kdensity plot) so I
> can't (don't
> know how to) make any selection upon the density..
>
> yscale(range()) expand range of axis ([G]p56) and range()
> never causes the
> data to be omitted from the plot([G] p59)
> one would appreciate either to expand or to REDUCE the axis
> scales using the
> same axis_scale_option
*
* 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/
