.-
help for ^lrplot^
.-

Plots coefficients from a logistic regression
----------------------------


    ^lrplot^  [ ^, XLab LEGgend SAVing(string) REPLace ^
                   ^ti_s(integer) ax_s(integer) leg_s(integer) xlab_s(integer)^
                   ^ti1(string) t2(string)^
                   ^GROPT(string) MATrix(string) debug ^]

Description
-----------

Plots the coefficients from a logistic regression
with confidence intervals, on a log scale.

Simplest use: do lrplot immediately after a logistic
regression.


Options
-------

^xlab^    gives coefficient labels on the x-axis
^legend^  gives a legend to the right
^sav^     saves the graph to a Stata graph file
^replace^ replaces the graph if it already exists
^ti1,ti2^ specifies the two title lines
^ti_s^
^ax_s^
^leg_s^
^xlab_s^  specify title, axis, legend and xlab-name size
          as a percentage (0-400)
^gropt^   gives options to pass on to graph
^matrix^  is used to specify a matrix that holds
          the coefficients and odds ratios that are to be
          used 
^debug^   is intended for developers. More or less useful.

Remarks 
-------
The utility @lrmatx@ is required. Not very extensively tested. 
Please inform the author of any bugs. Presently can only 
calculate 95% C.I. Modify lrmatx if you want this changed.

The coefficients are put in a matrix by the utility @lrmatx@.
Run it immediately after a logistic regression command
or use:
^logistic regression command^
^lrmatx^                    and
^matrix name=r(or),r(ci)^   to store results. Then use 
^lrplot , matrix(name)^     to get results

The advantages of putting results in a matrix are 
are that interactions and  different scalings of 
continuous variables can be made a lot more acessible 
graphically. 

For plotting of interactions, see @lrint@.

Examples
--------

    . ^logistic^ response age gender
    . ^lrplot^, legend
    . ^lrplot^, legend ti_s(150) ax_s(120)
    . ^lrplot^, xlab sav("model a")


    . ^xi: logistic age gender i.gender^
      [code fragment to generate easily 
       interpretable values of the 
       interaction terms. See ^lrint^]
    . ^lrplot^, matrix(model) legend

Author
-------

    Jan Brogger, University of Bergen, Norway
    jan.brogger@@med.uib.no


Shameless plug :)
--------------
 I am looking for internship & residency  in internal 
 medicine/pulmonary medicine US/EU. Am doing PhD in respiratory 
 epidemiology now.