Yule: From a two by two table, find the Yule coefficients of association, convert to phi, or polychoric, recreate table the table to create the Yule coefficient.
Description
One of the many measures of association is the Yule coefficient. Given a two x two table of counts
lll{
a b
c d
}
Yule Q is (ad - bc)/(ad+bc).
Conceptually, this is the number of pairs in agreement (ad) - the number in disagreement (bc) over the total number of paired observations. Warren (2008) has shown that Yule's Q is one of the ``coefficients that have zero value under statistical independence,maximum value unity, and minimum value minus unity independent of the marginal distributions" (p 787).
ad/bc is the odds ratio and Q = (OR-1)/(OR+1)
Yule's coefficient of colligation is Y = (sqrt(OR) - 1)/(sqrt(OR)+1)
Yule.inv finds the cell entries for a particular Q and the marginals (a+b,c+d,a+c, b+d). This is useful for converting old tables of correlations into more conventional phi or polychoric correlations.
Yule2phi and Yule2poly convert the Yule Q with set marginals to the correponding phi or polychoric correlation.
Usage
Yule(x,Y=FALSE) #find Yule given a two by two table of frequencies
Yule.inv(Q,m) #find the frequencies that produce a Yule Q given the Q and marginals
Yule2phi(Q,m) #find the phi coefficient that matches the Yule Q given the marginals
Yule2poly(Q,m) #Find the tetrachoric correlation given the Yule Q and the marginals
Arguments
x
A vector of four elements or a two by two matrix
Y
Y=TRUE return Yule's Y coefficient of colligation
Q
The Yule coefficient
m
A two x two matrix of marginals or a four element vector of marginals
Value
QThe Yule Q coefficient
RA two by two matrix of counts
Details
Yule developed two measures of association for two by two tables. Both are functions of the odds ratio
References
Yule, G. Uday (1912) On the methods of measuring association between two attributes. Journal of the Royal Statistical Society, LXXV, 579-652
Warrens, Matthijs (2008), On Association Coefficients for 2x2 Tables and Properties That Do Not Depend on the Marginal Distributions. Psychometrika, 73, 777-789.