Displaying 1 result from an estimated 1 matches for "fieldrookery".
2003 Apr 22
0
Correct SE in a poisson model.
...ormally distributed it is correct, but it is
overdipersed count data.
I make a model
> m.poisson <- glm(slugs~field,family=quasipoisson)
And I have these coefficients:
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.2429 0.2490 0.976 0.3323
fieldRookery 0.5790 0.3112 1.861 0.0666 .
The estimate mean = mean
1.275 = exp(0.2429)
2.275 = exp(0.2429+0.5790)
But and the correct standard error of mean? How to obtain this? Exist any
function for calculate this? Exist another better measure than SE for
non-normal errors (poisson, quasi, binom...