{smcl}
{* 22Jun12}{...}
{cmd:help plotbeta}
{hline}

{title:Title}

{hi:Plots the estimates from a series of lincom commands}

{title: Syntax}

{p 8 27 2}
{cmdab:plotbeta} {it:exp}
[{cmd:,} {it:options}]

{synoptset 20 tabbed}{...}
{synopthdr}
{synoptline}
{syntab:Main}
{synopt:{opt p:osition(numlist)} } specifies the position of the confidence intervals. {p_end}
{synopt:{opt eform} } specifies that the exponentiated form the parameters be plotted. {p_end}
{synopt:{opt ver:tical} } specifies that the confidence intervals are plotted vertically. {p_end}
{synopt:{opt l:evel(#)} } specifies the confidence level. {p_end}
{synopt:{opt labels} } specifies that variable labels are used in the plot. {p_end}
{synopt:{opt savedata(filename)} } specifies that the data in the final plot should be saved. {p_end}
{synopt:{opt addplot(graph)} } specifies that an additional twoway graph is added to the plot. {p_end}
{synopt:{help twoway_options } } {p_end}
{synoptline}
{p2colreset}{...}

{p}
where {it:exp} is defined

         {it:lincom_exp} [|] 

{p}
and where {it:lincom_exp} may be any linear combination of coefficients that is valid syntax for {help test}. Note, however, that {it:exp} may not contain an equal sign.


{title:Description}

{p 0 0}
This command will plot the parameter estimates and confidence intervals from a series of lincom commands.

{title:Updating this command using SSC}

{p 0 0} 
To obtain the latest version click the following to install the new version

{stata ssc install plotbeta,replace}

{title:Options}

{phang}
{cmdab:p:osition}{cmd:(}{it:numlist}{cmd:)} specifies the position of the confidence intervals. The default is that the point estimate and confidence intervals are plotted at points 1 2 3 ... It is possible to group confidence intervals in different parts of the plots e.g.  1 2 3 3.1 3.2, here the last three confidence intervals will be much closer to each other. 

{phang}
{cmdab:eform} specifies that the exponentiated form the parameters be plotted.

{phang}
{cmdab:labels} specifies that if the lincom expression is a variable then that variable's label is used instead of the variable name. 
This option is particularly useful if the linear predictor is made up entirely of main effects.

{phang}
{cmdab:ver:tical} specifies that the confidence intervals are plotted vertically rather than the default of horizontally.

{phang}
{cmdab:l:evel}{cmd:(}{it:#}{cmd:)} specifies the confidence level, in percent, for confidence intervals; see {help level}.

{phang}
{opt savedata(filename)} specifies that the data in the final plot should be saved.

{phang}
{opt addplot(graph)} specifies that an additional twoway graph is added to the plot. This is particularly useful for adding additional datapoints using
{help twoway_scatteri}.

{title:Examples}

{p 0 0}
A linear regression using the auto data, click the highlighted text in order

  {stata sysuse auto, clear}
  {stata reg price mpg turn length}
  {stata plotbeta mpg | turn | length}
  {stata plotbeta mpg | turn | length, labels}
  {stata plotbeta mpg | turn | length, addplot((scatteri 2.5 2.5, ms(S) msize(*2))) labels}

{p 0 0}
To remove the green text on the right hand side of the plot you need to add in the following two options

{p 2 2}
{stata plotbeta mpg | turn | length, addplot((scatteri 2.5 2.5, ms(S) msize(*2))) labels ylab(none, axis(2)) ytitle("", axis(2))}

{p 0 0}
This is the most basic graphic and using twoway options the look can be improved. 
The following command puts the confidence intervals

{p 2 2}
{stata plotbeta mpg | turn | length | mpg-length , xtitle(Parameters) yscale(range(0.7 4.3) axis(1)) yscale(range(0.7 4.3) axis(2)) title(Coefficients and Confidence Intervals) subtitle(from a Simple Linear Regression) xline(0, lp(dash)) }

An example using Survival data, click the highlighted text in order

  {stata sysuse cancer}
  {stata stset studytime, f(died)}
  {stata "xi:stcox i.drug age"}
  {stata plotbeta _Idrug_2 | _Idrug_3}
  {stata plotbeta _Idrug_2 | _Idrug_3, labels eform savedata(temp,replace)} 

Using the twoway options to solve a problem with labelling

  {stata "plotbeta age | _Idrug_2 | _Idrug_3, eform labels xline(1, lp(dash)) "}

{p 0 0}
{bf: Note} that by adding {hi:age} into the estimates the ylabels are off the graph. This
is due to the single quote in the variable label. Either change the label or alter the graphregion size

{p 2 2}
{stata "plotbeta age | _Idrug_2 | _Idrug_3, eform labels xline(1, lp(dash)) graphregion(margin(l+40 r+15)) "}

{title:Author}

{p}
Adrian Mander, MRC Biostatistics Unit, Cambridge, UK.

Email {browse "mailto:adrian.mander@mrc-bsu.cam.ac.uk":adrian.mander@mrc-bsu.cam.ac.uk}

{title:Also see}

Related commands

HELP FILES
{help lincom}