0

I am trying to run the Dredge function of the MuMIn package and keep getting an error, saying "result is empty". I don't know why and cannot find information on the meaning of this error message.

As far as my study of this function and package, the below code "should" be correct. Basically, I have a General Linear Mixed Model being run and I want to use the dredge function to run a model selection procedure based on AICc. 

options(na.action = "na.fail") # Require for dredge to run

glmm1 <- lmer(cpue_diff ~ year +  p.afraid + s.frequency.monitoring + (1 | f1.name ), data = dat, REML=FALSE)

summary(glmm1)

model_dredge <- dredge(glmm1, beta = FALSE, evaluate = TRUE, rank = "AICc")

options(na.action = "na.omit") # set back to default

The error message is:

"Fixed term is "(Intercept)" Error in dredge(glmm1, beta = FALSE, evaluate = TRUE, rank = "AICc") : result is empty"

Any ideas anyone what this message means and how to correct it?

Much appreciated!

aosmith
  • 34,856
  • 9
  • 84
  • 118
alfajor
  • 1
  • 3
  • Please share a portion of your data `dat`. Without being able to reproduce your error, it is extremely difficult for anyone to help you. – astrofunkswag May 21 '19 at 21:01
  • I'm wondering if the problem is that the minimum model for a mixed model should be the intercept plus the random effects, since **lme4** won't fit a model without a random effect in it. – aosmith May 21 '19 at 21:37

1 Answers1

1

Check the the dimensions of your data frame. I had the same error message and when I checked out my data frame, there was more than a thousand "extra rows" full of NAs. After subsetting the data frame to include only the true rows, everything worked well with the dredge function.

Paulo Paiva
  • 11
  • 1