Small discrepancy in ER cv when covariates are in the detection function

[LHMarshall]

Values from the software matched the hand calculations up to 4 decimal places but differed slightly after that.

Further comparison of er_method=1 and er_method=2 when covariate is in detection function. Mock example with ducknests

ducknests.pdf

Originally posted by @erex in #182

[LHMarshall]

From Eric: I suspect those differences might be magnified if the variability in encounter rate between transects was larger. For the ducknests, encounter rate is fairly homogeneous between the 20 transects