Re: st: Graph Mean SD of continuous variable over categorical variable - but with a twist

[Nick Cox <njcoxstata@gmail.com>]

See also

SJ-5-4  gr0023  . . . . Stata tip 27: Classifying data points on scatter plots
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q4/05   SJ 5(4):604--606                                 (no commands)
        tips for using separate, gray-scale gradation, and text
        characters as class symbols to classify data points on
        a scatter plot

for technique relevant to the "for each group" part of the question.
In a sentence, use -separate- to separate the response into different
variables.

Nick

On Tue, Oct 25, 2011 at 8:35 AM, Maarten Buis <maartenlbuis@gmail.com> wrote:
> On Tue, Oct 25, 2011 at 9:02 AM, Lars Folkestad wrote:
>> I am trying to depict the differences in mean+-SD of two continuous
>> Variables over a categorical variable <snip>
>> scatter mean cat || rcap upper lower cat
> <snip>
>> But this does not give me the option to change the colour of the
>> Mean (dot) or the SD (bars) for each group, which visually would
>> Be preferable.
>
> It does give you that possibility, you just need to find the right
> options. That command consists of two parts separated by -||-. The
> first part graphs mean against the categorical variable and the second
> graphs the standard deviation bars against the categorical variable.
> To find the options possible after the first command type in Stata
> -help twoway scatter- and to find the options possible after the
> second command type in Stata -help twoway rcap-. In the example below
> I give a silly (and extremely ugly) example of how you can change the
> way your graph looks:
>
> *--------------------- begin example ----------------------
> sysuse auto, clear
> egen mean = mean(price) , by(rep78)
> egen sd = sd(price), by(rep78)
> gen upper = mean + sd
> gen lower = mean - sd
>
> scatter mean rep78 ,        /// plots the mean against rep78
>    msymbol(S)              /// change the symbol to a solid square
>    msize(huge)             /// make the marker size huge
>    color(purple)           /// change the color of the symbol to purple
>    ||                      /// start a new overlay graph
>    rcap upper lower rep78, /// plots the standard deviation against rep78
>    lpattern(longdash)      /// make the line dashed
>    lwidth(thick)           /// make the line thick
>    lcolor(pink)            //  make the the line pink
> *---------------------- end example -----------------------
> (For more on examples I sent to the Statalist see:
> http://www.maartenbuis.nl/example_faq )
>

*
*   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/