[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Before and After Pairwise Plots

From   n j cox <>
Subject   RE: st: Before and After Pairwise Plots
Date   Fri, 19 Oct 2007 17:14:22 +0100

I haven't got a clear picture of what you want
to do, but my guess is that it has been
programmed already. Install these programs:

. ssc inst parplot
. ssc inst pairplot
. ssc inst linkplot

and see if one of those doesn't suit.

A closely overlapping question was asked on Monday. I'll
repeat some references from that:

Cox, N.J. 2006. Assessing agreement of measurements and
predictions in geomorphology. Geomorphology 76: 332-346

is a paper discussing a variety of graphical and numerical
ways of approaching this. Don't let the apparent restriction
to geomorphology trouble you: almost all of the ideas transcend
disciplinary boundaries.

All the analyses in this paper used Stata.

A linked paper in which the Stata side is explicit is

Cox, N.J. 2004. Graphing agreement and disagreement.
Stata Journal 4: 329-349.

In addition, Rino Bellocco gave me this reference at
the Stockholm Users' Meeting. It has material on
graphics too:

Campbell, D. T. and Kenny, D. A. 1999.
A Primer on Regression Artefacts. New York: Guilford Press


C Rose

I'd like to know how to produce a plot that shows how a dependent
variable, measured for a number of individuals, changes after an
intervention. I have managed to find an FAQ on the topic for older
versions of Stata, which explains the situation quite clearly (please
see, as does
the referenced paper. I'm using Stata/IC v10.

It seems the "twoway pcspike" command is appropriate, but I have two
questions about this:

1. This requires one to have data in the form y1 x1 y2 x2. I have two
variables: the dependent variable, Y, and a categorical variable that
indicates whether the corresponding entry in Y was measured pre- or
post-intervention. How can I (easily!) convert my data into the y1 x1
y2 x2 form? (It would be nice if this were to generalise to the case
where I have measurements on Y at more than two time points.)

2. How do I get "twoway pcspike" to place the correct categorical
label on the x-axis, rather than the values I used in x1 and x2? I
can edit the text manually in the graph window, but found that my
text ran off the edge of the display.

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index