[R-sig-ME] Mixed cumulative link modelling - error message

[Georgina Southon]

Hi Dan,
Thanks - that worked a treat!
Georgina




On 21 Aug 2015, at 18:51, Dan McCloy <drmccloy at uw.edu> wrote:

If str() says that bob$rating is of type int, then it is of type int.  Observe:

foo <- data.frame(bar=seq(1,3,1), baz=factor(seq(4,6,1)), qux=factor(seq(7,9,1), ordered=TRUE))
str(foo)
'data.frame':    3 obs. of  3 variables:
 $ bar: num  1 2 3
 $ baz: Factor w/ 3 levels "4","5","6": 1 2 3
 $ qux: Ord.factor w/ 3 levels "7"<"8"<"9": 1 2 3

you need to do bob$rating <- factor(bob$rating, ordered=TRUE) before running your model.
-- dan

Daniel McCloy
http://dan.mccloy.info/
Postdoctoral Research Fellow
Institute for Learning and Brain Sciences
University of Washington



On Fri, Aug 21, 2015 at 9:36 AM, Georgina Southon <g.southon at sheffield.ac.uk> wrote:
Hi,

I hope someone can help. I have problems fitting a cumulative link mixed model to my data.
Looking at the str output, rating shows as an integer, despite the fact that it shows as an ordered factor when asked is.factor. Is this the problem or is there something else I have overlooked?

str(bob)
'data.frame':   283 obs. of  4 variables:
 $ site     : Factor w/ 10 levels "ADD","BH","BR",..: 4 4 4 4 4 4 4 4 4 4 ...
 $ person   : Factor w/ 283 levels "ADD10","ADD11",..: 80 91 102 104 105 106 107 108 109 81 ...
 $ treatment: Factor w/ 2 levels "CONTROL","MEADOW": 2 2 2 2 2 2 2 2 2 2 ...
 $ rating   : int  4 3 4 2 3 4 5 5 5 3 ...
>
> m1<-clmm2(rating~treatment,random=person,data=bob)

Error in clm2(location = rating ~ treatment, data = bob, subset = c("1",  :
  response needs to be a factor

Thanks for any feedback / advice!



        [[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models



	[[alternative HTML version deleted]]