-------------------------------------------------------------------------------
Title
paragr -- Parallel graphing of a coefficient across different equations
Syntax
paragr variable [, options]
Description
paragr provides a fast and easy way to compare a coefficient across different equations within an estimation.
It can used to visualize the parallel assumption of ordered logit or the equality of coefficients for quantile regression, etc.
By default paragr plots a scatter graph. A quadratic fitted line (qfit) can be superimposed.
Twoway graph options can be specified for the entire graph, or separately for the scatter graph or for the qfit graph.
Options
+------+ ----+ Main +-------------------------------------------------------------
scatter( ) standard options for scatter graph. Use this to set color, range, etc
qfit a quadratic fitted line is superimposed.
qfit( ) standard options for qfit graph. Use this to set color, range, etc
gen( ) a prefix for saving the plot points as variables.
click here for a list of twoway options, including:
nodraw suppress the display of graph
saving( ) save as a Stata graph
name(name[, replace]) specifies the name of the graph
Examples
* reg3 sysuse auto, clear reg3 (price weight rep78) (price weight rep78 mpg) (price weight rep78 mpg headroom) (price weight rep78 mpg headroom length) paragr rep78, xtitle(Equation) gen(reg3_) saving(mygraph)
* sqreg sysuse auto, clear sqreg price mpg rep78 headroom length, q(.10 .20 30 .40 .50 .60 .70 .80 .90) reps(25) paragr rep78, qfit
* Unconstrained gologit sysuse auto, clear egen money=cut(price), group(5) gologit money headroom mpg weight length turn displacemen paragr displacement
* constrained gologit2 from ssc (need Stata 8.2 or better) sysuse auto, clear egen money=cut(price), group(5) gologit2 money headroom mpg weight length turn displacement, auto paragr weight, qfit ytitle(odds ratios) xtitle(cut)
Remarks
qfit requires at least two equations.
Coefficient values goes on the y-axis.
Acknowledgement
Nick Cox provided detailed and helpful suggestions.
Author
Roy Wada roywada@hotmail.com