Message-ID: <1308768224.4836.YahooMailRC@web27007.mail.ukl.yahoo.com>
Date: 2011-06-22T18:43:44Z
From: Iker Vaquero Alba
Subject: Count data
Dear list:
This may be quite a naive question: I am fitting a model in which the
response variable is the number of chicks hatched in nests of different pairs of
swallows. I assumed that this is count data, so I used family=poisson for the
model. But when fitting the residuals vs. fitted values, I cannot see the
structure that I've been told count data normally has. Also, amount of variance
explained by random effects when assuming count data is this:
Random effects:
Groups Name Variance Std.Dev.
pair:site (Intercept) 2.0757e-02 1.4407e-01
site (Intercept) 6.8285e-17 8.2635e-09
year (Intercept) 2.5513e-03 5.0511e-02
Whereas if I assume gaussian errors, I get this:
Random effects:
Groups Name Variance Std.Dev.
pair:site (Intercept) 4.89955 2.21349
site (Intercept) 0.45493 0.67448
year (Intercept) 0.22061 0.46969
Residual 0.80047 0.89469
I don't know if this is indicative of anything, but I'm concerned I may be
making the wrong assumptions. Graph of residuals vs fitted is attached.
Thank you very much
________________________________________
Iker Vaquero-Alba
Centre for Ecology and Conservation
Daphne du Maurier Building
University of Exeter, Cornwall Campus
Treliever Road
TR10 9EZ Penryn
U.K.
http://biosciences.exeter.ac.uk/cec/staff/postgradresearch/ikervaquero-alba/