1

I'm quite new to R and am trying to run a number of separate multiple regression analyses on a dataset with some 250 rows. I've successfully done a few of these on columns with a small number of NA's, however it's refusing to work on columns with larger numbers of NA's (48 NA's) and giving me the error message "Error in mkRespMod(fr, REML = REMLpass) : NA/NaN/Inf in 'y'". The command I'm using is 

summary(with(Gratios,  lmer(log(G_ratio) ~ -1 + Type + scaled_date + I(scaled_date^2) + (1|Site) + (1|Crop) +(1|Harvest.status), REML = F) ))

Is there a way to make R run the analysis anyway? I'm sorry if my question is not very specific, I'm very new to R and not quite sure of what the issue even is, please no one bite my head off and thanks for the help

StBebe
  • 11
  • 1
  • 3
  • I suspect your problem is with your choices regarding the statistics methods. I suspect that the log-transformation is not really appropriate here (do you have zero values as ratios perhaps?). You should probably use a GLMM instead. – Roland Feb 25 '16 at 12:26
  • You specify a number of random effects. Could it be that the NAs cluster in one of these groups (Site, Crop, Harvest.status), so that you end up having no observations on your dependent variable for one or more of these clusters ? – Dominix Feb 25 '16 at 13:15

0 Answers0