[R-sig-ME] Error: (maxstephalfit) PIRLS step-halvings failed to reduce deviance in pwrssUpdate

[Kirsty McLaughlin]

To whom it may concern,


I am a novice in the use of R, however I have been trying my best to assess the effect of noise treatment over a three day period on the time spent carrying out certain behaviours. I am using a GLMM as I don't have sample independence and my response variables are non-normally distributed and skewed. I have been using Family Gamma for my analysis.


When running a model on 'TTAgressionCon' (time spent being aggressive) including the predictor variable Day (3 levels) I get the above error message. When the model runs with other predictor variables e.g. Gender, I do not get this error.


I have the most up to date version of lme4.


Below is the model that generated the error message:

model1 <- glmer(TTAggressionCon ~ Treatment + Day + Treatment*Day + (1|Round/Tank/Pair/Fish), data = data1, family = Gamma)


I have attached my data if this helps with diagnosis or suggestions.


Any help or guidance on this matter would be greatly appreciated.



Kind regards
Kirsty McLaughlin

PhD student
Queen's University Belfast
School of Biological Sciences
Medical Biology Centre
97 Lisburn Road
Belfast, BT9 7GT

Tel: 028 90 972103
Email: kmclaughlin23 at qub.ac.uk
Web: www.animal-behaviour.eu<http://www.animal-behaviour.eu>