Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: RE: three mean and sd plots on the same graph?


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: three mean and sd plots on the same graph?
Date   Thu, 8 Nov 2007 17:04:47 -0000

-over()- indeed does not apply here; I didn't say otherwise. -stripplot-
with -by()- will produce graphs of the kind you sketch: I don't see what
you think is the difficulty. You would need to specify -vertical-. 

Nick 
n.j.cox@durham.ac.uk 

Joseph Wagner

Unfortunately I can't use over with -stripplot- since I can't combine 
()over with more than one variable and I have three that I wanted to 
compare side by side for each value of the X- axis categorical variable.

by() won't work either for that matter since what I want is basically X 
over many Y as follows with a crude representation of my boxplot:


            |
            |   -             -        -             -
            |  +-+     -     +-+      +-+     -     +-+
            |  |-|    +-+    |-|      |-|    +-+    |-|
            |  +-+    |-|    +-+      +-+    |_|    +-+
            |   -     +-+     -        -     +-+     -
            |          -                      -
            +-------------------------------------------
              var_1  var_2  var_3    var_1  var_2  var_3
                     age=8                 age=9

Nick Cox wrote:

> Now my best advice is to use -stripplot- from SSC. Its default is to
> show 
> strips of data points, but you can jitter or stack and add boxes or
> bars. 
>
> . sysuse auto
> (1978 Automobile Data)
> . stripplot trunk turn mpg, bar
>
> . stripplot trunk turn mpg, bar  
>
> . stripplot trunk turn mpg, bar vertical
>
> You can use a -by()- option as well. I'd recommend showing the data, 
> but you can blank it out with -ms(none)-. 
>
> By default the bars are offset from the data which may answer your 
> last question. 
>
> Joseph Wagner
>
> I created box plot graph with of 3 X variables over the same Y
variable 
> but was then asked to produce a similar graph this time using mean
(not 
> median) and +/- 95% CI.   I don't think Stata can do this but thanks
to 
> a post by Nick three years ago, I was able to (sort of) create these:
>
> egen mean = mean(cont), by(cat) 
> egen sd = sd(cont), by(cat) 
> gen upper = mean + sd
> gen lower = mean - sd 
>
> scatter mean cat || rcap upper lower cat 
>
> . . .  is the example Nick gave but how would (or could I?) do this
for
> three different X variables (in Nick's example I suppose it would be 3
> different 'cont' variables) on the same graph?  I can do 
> this for two but not 3 which brings me to my next problem: 
>
> The box plots I created earlier were each side by side for the same
> value of X (rather than on top of one another) but the two -rcap-
graphs
> are on top of each other making it impossible to differentiate 
> the two lines (three would be even worse).  Is there a way to separate
> these lines or do I need to graph these data in another way
altogether?

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index