www.cmm.bristol.ac.uk/forum

Thanks for your reply Chris. I was concerned that the difference between the models given the small data would hide in the third or fewer decimal points. So I expanded the same data, and ran the models. As you see below, I do still get a difference within Mlwin. Not sure if this is unique to my PC ...

Chris, I ran a test testing the difference between the two id styles. I used the data given in LEMMA section 11.2: A Three-Level Variance Components Model using the the STATA data 11.2.dta.


library(foreign)
tut <- read.dta("11.2.dta")


# run with unique Id
unique.id.performance <- rep(NA, 10 ...

Thanks to both of you. I missed this solution while browsing through the topics. I confirm this works.

On a related note. I wonder how rp- and fpsandwish works for weighting? The manual states:

weighting: a deprecated option for specifying weights in IGLS estimation: see fpsandwich
and ...

Could you try adding the debugmode=TRUE option and then checking within the MLwiN interface whether any errors are reported?

When I enter debugmode I discover that Mlwin crashes. I tried varying the following arguments, but it still crashes:

optimat=,
resi.store = ,
x64= ,
weighting = list ...

ok, thanks Chris. It was not immediately clear from the R2mlwin documentation (https://cran.r-project.org/web/packages ... 2MLwiN.pdf), pp. 22-23. An additional note is may be required.

Hi Chris,

I would have to check, but it's possible that the R display is using unique codes instead of changes to the ID value as in MLwiN to determine different units.

Ok, thanks, that would be great. Again, I could, of course, use unique id, but performance is noticeably slower with large ...

Hi

The mymodel@Hierarchy slot does not get the correct number of intermediate levels (e.g. class id in a data of students nested in class nested in schools) when these intermediate unit id are renumbered.

For example:

School class student
1 1 1
1 1 2
1 2 1
1 2 2
2 1 1
2 1 2
2 2 1
2 3 1


The ...

Hi Chris,

Trying to add weights in a four-level mode and only for the first level. I assume this is a correction estimation option argument:

estoptions = list(EstM = 0, optimat=T, resi.store = TRUE, debugmode=F,
weighting = list(weightvar=c(NA,NA,NA, "wg"), standardised=T))

f2 = as.formula ...

Hi

(A)
How can I define the reference category for non-numeric covariates (binary, ordinal, or multinomial)? According to the manual "[] When placed immediately after an explanatory variable, indicates that the variable is categorical.The string in the [] represents the reference category; if empty ...

Ok, thanks, Chris.

Looking forward to reading Kelvyn's module.