TOPICS
Search

Correlation Ratio


Let there be N_i observations of the ith phenomenon, where i=1, ..., p and

N=sumN_i
(1)
y^__i=1/(N_i)sum_(alpha)y_(ialpha)
(2)
y^_=1/Nsum_(i)sum_(alpha)y_(ialpha).
(3)

Then the sample correlation ratio is defined by

 E_(yx)^2=(sum_(i)N_i(y^__i-y^_)^2)/(sum_(i)sum_(alpha)(y_(ialpha)-y^_)^2).
(4)

Let eta_(yx) be the population correlation ratio. If N_i=N_j for i!=j, then

 f(E^2)=(e^(-lambda)(E^2)^(a-1)(1-E^2)^(b-1)_1F_1(a,b;lambdaE^2))/(B(a,b)),
(5)

where

lambda=(Neta^2)/(2(1-eta^2))
(6)
a=(n_1)/2
(7)
b=(n_2)/2,
(8)

and _1F_1(a,b;z) is the confluent hypergeometric limit function. If lambda=0, then

 f(E^2)=beta(a,b)
(9)

(Kenney and Keeping 1951, pp. 323-324).


See also

Correlation Coefficient, Regression Coefficient

Explore with Wolfram|Alpha

References

Kenney, J. F. and Keeping, E. S. Mathematics of Statistics, Pt. 2, 2nd ed. Princeton, NJ: Van Nostrand, 1951.

Referenced on Wolfram|Alpha

Correlation Ratio

Cite this as:

Weisstein, Eric W. "Correlation Ratio." From MathWorld--A Wolfram Web Resource. https://mathworld.wolfram.com/CorrelationRatio.html

Subject classifications