[R] dredge solely offset models in MuMIn

[Katarzyna Kulma]

hello everyone,

I'm modelling in lmer an average chick weight defined as
"Total.brood.mass ~ offset(chick.number), with three fixed and two
random effect. Next, I want to use function dredge from MuMIn package
for model averaging. Not sure why, but in consequence the offset
variable is treated as a predictor, so I get a table that mixes models
with and without that offset term (the first crosses to the right):

Model selection table
    (Int) all.rLD bng INF all.rLD:bng all.rLD:INF bng:INF
all.rLD:bng:INF off(log.fld) df logLik   AICc   delta
256 75.70 -0.6761 +   +   +           +           +       +
   +           11 -801.906 1627.4  0.00
168 75.44 -1.1670 +   +                             +
      +             8 -805.477 1627.8  0.43
184 75.68 -0.7785 +   +                +           +
     +             9 -804.390 1627.8  0.46
192 75.65 -0.8870 +   +   +           +           +
    +           10 -803.651 1628.6  1.22
176 75.40 -1.2690 +   +   +                        +
     +             9 -804.832 1628.7  1.35
152 76.25 -0.7686 +   +                +
      +             8 -807.367 1631.6  4.21
136 76.13 -1.1860 +   +
       +             7 -808.533 1631.7  4.35
128 77.34 -0.6853 +   +   +           +           +       +
                11 -804.083 1631.7  4.35
 56  77.33 -0.7899 +   +                +           +
                     9 -806.572 1632.2  4.83
 40  77.08 -1.1880 +   +                             +
                      8 -807.687 1632.2  4.84
160 76.16 -0.8849 +   +   +           +
     +             9 -806.607 1632.3  4.90
144 76.06 -1.2960 +   +   +
      +             8 -807.871 1632.6  5.21

When I try to use a subset function, I get an error:


> offavg<-dredge(avgch, subset=offset(log.fledge))

Error in offset(log.fledge) : object 'log.fledge' not found
In addition: Warning messages:
1: In dredge(avgch, subset = offset(log.fledge)) :
  comparing models fitted by REML
2: In dredge(avgch, subset = offset(log.fledge)) :
  not all terms in 'subset' exist in 'global.model'

So, how do I get a table with offset models only? Not sure if I define
something incorrectly or maybe there's a different way to get around
that? Will appreciate any suggestions

cheers,
Kasia


Katarzyna Kulma

PhD Student
Department of Ecology and Genetics
Institute of Ecology and Evolution/Animal Ecology
Uppsala University
Norbyvägen 18D
SE-752 36 Uppsala, Sweden

email: katarzyna.kulma at ebc.uu.se
Tel.+46 (0)18 471 2672
Fax.+46 18 471 6484