Library: | eiv |
See also: | simex |
Quantlet: | reca | |
Description: | RECA (REgression CAlibration) is a method in which replacing the unobserved x by its expected value E(x|w,z) and then to perform a standard analysis. |
Usage: | {beta,bv} = reca(y,w,z,su2) | |
Input: | ||
y | n x 1 matrix, the design variables | |
w | n x 1 matrix | |
z | n x 1 matrix | |
su2 | the variance of the measurement error | |
Output: | ||
beta | vector, the estimate | |
bv | matrix, the variance of the estimate |
library("xplore") library("eiv") n=100 randomize(n) y=floor(uniform(n)+0.25) w=uniform(n)^2 z=floor(uniform(n)+0.4) su2=var(w)/4 res=reca(y,w,z,su2) res.beta res.bv
Contents of beta [1,] -2.7659 [2,] 1.531 [3,] 0.96504 Contents of bv [1,] 0.49574 -0.68988 -0.24432 [2,] -0.68988 1.7197 0.06315 [2,] -0.24432 0.06315 0.33938