COBOT: Conditional Ordinal-By-Ordinal Test
URL of this page:
http://biostat.mc.vanderbilt.edu/OrdinalRegression
We developed a method for testing for association between two ordinal variables,
Y and
X, while adjusting for covariates
Z. The data analysis program (
COBOT-analysis.r) is written in R (
www.r-project.org). The main analysis functions are
COBOT.emp()
and
COBOT.scores()
for calculating empirical and asymptotic p-values when proportional odds models are used to model P(
Y|
Z) and P(
X|
Z). Please see the beginning lines of the file for detailed descriptions.
Version history
12/06/2010: Old version did not work for ordinal variables with 3 levels. The new version should work.
12/04/2011: The R Design package has been replaced by R rms package. The code is updated to reflect the change.
07/23/2012: a bug fix in the
ifelse()
calls inside function
COBOT.scores()
.
Simulation code and other simulation results
Our simulation code is in
COBOT-simulation.r.
Other simulation results not included in the article:
ordinal-otherresults.pdf
Software package
The code given above has now been incorporated into an R package, PResiduals. This package is freely available at
http://cran.r-project.org/web/packages/PResiduals/index.html.
Contact and paper
For questions, please contact
chun.li@vanderbilt.edu or
bryan.shepherd@vanderbilt.edu .
The paper that describes the method is:
Li C, Shepherd BE (2010) Test of association between two ordinal variables while adjusting for covariates.
Journal of the American Statistical Association 105:612-620