Hi,

I am trying to run a multiple imputation to correct for measurement error. Using "Int J Epidemiol. 2006 Aug;35(4):1074-81.Multiple-imputation for measurement-error correction.
Cole SR1, Chu H, Greenland S paper" for guidance, I first ran a linear regression of my validated exposure measurements on the mismeasured exposure.

regress X (blood levels of a toxin) = a + b(mismeasured sweat levels of a toxin)

I save the outcomes for the beta coefficients and variances as matrices next.
matrix beta = e(b)
matrix var = e(V)

matrix list beta

beta[1,2]
b _cons
y1 .76620142 .57014411


matrix list var

symmetric var[2,2]
b _cons
b .00131926
_cons -.00323015 .00828289

The next step is to generate new beta coefficients by including a component of variation, drawn from a multivariate normal distribution with mean(beta) cov(var)
clear
set obs 100
number of observations (_N) was 0, now 100

drawnorm w x y cons, n(100) cov(var) means(beta)

However when I try running this, it comes up as a conformability error. I am not sure what I am doing wrong. I would be really grateful for any thoughts.

Thanks,
Io