{smcl}
{* April 15, 2008 @ 21:14:32}{...}
{cmd:help metandiplot} {right: (Roger Harbord)}
{right:also see:  {help metandi}}
{hline}

{title:Title}

    {hi:metandiplot} {hline 2} SROC plot of results from metandi


{title:Syntax}

{p 8 17 2}
{cmdab:metandiplot}
[{it: tp fp fn tn}]
[{it:{help if}}]
[{it:{help in}}]
[{it:{help weight}}]
[{cmd:,}
{cmdab:notr:uncate}
{cmdab:l:evel(}{it:#}{cmd:)}
{cmdab:predl:evel(}{it:{help numlist}}{cmd:)}
{cmdab:n:points(}{it:#}{cmd:)}
{it:{help metandiplot##subplot_options:subplot_options}}
{cmd:addplot(}{it:{help addplot_option:plot}}{cmd:)}
{it:{help twoway_options}}
]

{p 4 8 2}
{cmd:aweight}s, {cmd:fweight}s, and {cmd:pweight}s are allowed;
see {help weight}.


{title:Description}

{p 4 4 2}
{cmd:metandiplot} graphs the results from {helpb metandi}
on a summary receiver operating characteristic (SROC) plot.
By default the display includes:{p_end}

{p 8 12 2}- a {hi:summary point}
showing the summary sensitivity and specificity{p_end}
{p 8 12 2}- a {hi:confidence contour}
outlining the confidence region for the summary point{p_end}
{p 8 12 2}- one or more {hi:prediction contours}
outlining the prediction region
for the true sensitivity and specificity in a future study{p_end}
{p 8 12 2}- the {hi:HSROC curve}
from the hierarchical Summary ROC (HSROC) model

{p 4 4 2}
If the optional variables
{it: tp fp fn tn} are included on the command line,
the plot also includes {hi: study estimates}
indicating the sensitivity and specificity
estimated using the data from each study separately.

{p 4 4 2}
Any of these features may be customised or turned off using the
{help metandiplot##subplot_options:subplot_options}.


{title:Options}

{p 4 8 2}
{cmd:notruncate} specificies that the HSROC curve will {it:not} be truncated
outside the region of the data.
By default, the HSROC curve is not shown when
the sensitivity or specificity is less than
its smallest study estimate.


{p 4 8 2}
{cmd:level(}{it:#}{cmd:)} specifies the confidence level, in percent,
for the confidence contour; see help {help level}.

{p 4 8 2}
{cmd:predlevel(}{it:{help numlist}}{cmd:)} specifies the levels,
in percent, for the prediction contour(s).
The default is a single contour
at the same probability level as the confidence region.
Up to five prediction contours are allowed.

{p 4 8 2}
{cmd:npoints(}{it:#}{cmd:)} specifies the number of points to use in drawing
the outlines of the confidence and prediction regions.
The default is 500.

{p 4 8 2}{marker subplot_options}
{it:subplot_options}:
{cmdab:su:mmopts()}, {cmdab:co:nfopts()},
{cmdab:pr:edopts()}, {cmdab:cu:rveopts()} and {cmdab:st:udyopts()}
specify options that control the display of the
summary point, confidence contour, prediction contour(s), HSROC curve
and study symbols respectively.
The options within each set of parantheses
are simply passed through to the appropriate {helpb twoway} plot.
In addition, any of the plots can be turned off
by specifying, for example, {cmd:summopts(off)}.

{p 4 8 2}
{cmd:addplot(}{it:plot}{cmd:)} allows adding additional
{cmd:graph} {cmd:twoway} plots to the graph;
see {it:{help addplot_option}}.
For example, empirical Bayes predictions could be generated
by using {helpb predict} after {helpb metandi} and added to the graph.
See {help metandipostestimation}.

{p 4 8 2}
{it:twoway_options} are most of the options documented in 
{it:{help twoway_options}}, including options for titles, axes, labels,
schemes and saving the graph to disk.
The {cmd:by()} option is not allowed, however.


{title:Remarks}

{p 4 4 2}
The default is to weight the study estimates
by the total number in each study,
giving symbols (open circles by default)
scaled according to the size of the weights;
see {help scatter##remarks14:weighted markers}.
To make the symbols all the same size,
specify constant weights, e.g. {cmd:[aw=1]}.


{title:Examples}

{p 4 8 2}{cmd:. metandiplot}

{p 4 8 2}{cmd:. metandiplot tp fp fn tn}

{p 4 8 2}{cmd:. metandiplot tp fp fn tn [aw=1], conf(off) curve(off) predlevel(50 80 95 99)}


{title:Also see}

{p 4 13 2}
Online:  {helpb metandi}, {help metandi postestimation}.