What is the difference between corr() and cov() with panel data?

[Brennan RBratton]

When is it more appropriate to use cov() vs corr() for the residual structures for panel data?

Example Code:

xtpoisson Outcome CovariateA##CovariateB CovariateC CovariateD, exposure(area) corr(exchangeable)

xtnbreg Outcome CovariateA##CovariateB CovariateC CovariateD, exposure(area) corr(independent)

mepoisson Outcome CovariateA##CovariateB CovariateC CovariateD, exposure(area) || GroupID:, covariance(unstructured)

menbreg Outcome CovariateA##CovariateB CovariateC CovariateD, exposure(area) || GroupID:, covariance(unstructured)

glm Outcome CovariateA##CovariateB CovariateC CovariateD, exposure(area) vce(robust) family(gamma) link(identity)