Vpo: Variance for estimated Mann-Whitney parameter under...

In asht: Applied Statistical Hypothesis Tests

Variance for estimated Mann-Whitney parameter under proportional odds.

Description

A function to calculate the variance of the estimated Mann-Whitney parameter under the proportional odds model.

Usage

Vpo(PHI, tf, ny, nx)

Arguments

PHI	Mann-Whitney parameter
tf	tie factor
ny	sample size for the y variable
nx	sample size for the x variable

Details

Uses uniroot to find the variance numerically. The approximation using the LAPH method (Lehmann alternative model combined with the proportional hazards model) is much faster and is the default.

Value

the variance of the estimated Mann-Whitney parameter

Examples

Vpo(.7,.8,5,7)
## Compare to LAPH method
VLAPH<- function(PHI,tf=tiefactor,ny=n.y,nx=n.x){
        tf*(PHI*(1-PHI)/(ny*nx))*
          (1+((ny+nx-2)/2)*
             ((1-PHI)/(2-PHI)+PHI/(1+PHI)))
      }   
VLAPH(.7,.8,5,7)      

asht documentation built on Aug. 24, 2023, 5:08 p.m.