From Agresti's 2010 ordinal book, an eye disease data set involving ordinal measurements of retinopathy in left and right eyes of insulin-taking young diabetics.
data("eyed")A data frame with 720 observations on the following 23 variables.
rrla numeric vector
lrla numeric vector
rla numeric vector
rmeRight eye macular oedema (absent = 0, present = 1)
lmeLeft eye macular oedema (absent = 0, present = 1)
rreRight eye refraction index
lreLeft eye refraction index
riopRight eye intraocular pressure
liopLeft eye intraocular pressure
ageAge at diagnosis (values are less than 30 years)
diabDuration of diabetes (years)
ghGlycosylated haemoglobin level
sbpSystolic blood pressure (mmHg)
dbpDiastolic blood pressure (mmHg)
bmiBody mass index, in kg / m2, rounded to an integer
pra numeric vector
sexmale = 0, female = 1
protProteinuria (absent = 0, present = 1)
dosea numeric vector
resida numeric vector
rerlRight eye severity of retinopathy,
an ordered factor with levels
None, Mild, Moderate,
Proliferative
lerlSame as rerl but for
left eye
rl2a numeric vector
The data is used as the example in
Section 9.1.3 of Agresti (2010),
entitled
Eye Disease Risk Factors.
Williamson and Kim (1996) describe a
Wisconsin epidemiological study
in which insulin-taking young diabetics
were examined for diabetic retinopathy
(both prevalence and severity)
in both eyes.
One could fit a proportional odds model
to each eye.
The retinopathy severity is ordinal:
None, Mild, Moderate,
Proliferative.
Variables lerl and rerl can
be considered the primary responses.
Agresti, A. (2010). Analysis of Ordered Categorical Data, Second Edition. Wiley: Hoboken, NJ, USA.
Williamson, J. and Kim, K. (1996). A global odds ratio regression model for bivariate ordered categorical data from opthalmologic studies. Statistics in Medicine 15, 1507--1518.