Error "shifted too far"

Welcome to the forum for MLwiN users. Feel free to post your question about MLwiN software here. The Centre for Multilevel Modelling take no responsibility for the accuracy of these posts, we are unable to monitor them closely. Do go ahead and post your question and thank you in advance if you find the time to post any answers!

Remember to check out our extensive software FAQs which may answer your question: http://www.bristol.ac.uk/cmm/software/s ... port-faqs/
Post Reply
gilleecl
Posts: 27
Joined: Fri Dec 17, 2010 11:24 am

Error "shifted too far"

Post by gilleecl »

Hi,

I'm having problems with version 2.30. When I click 'notation' and untick 'general', I get an error message "shifted too far". When I click ok on this, I get an error "shifted too far(set)". Then when I click again, MlwiN stops working and the program closes. Would you be able to advise please?

thanks
Lorraine
ChrisCharlton
Posts: 1348
Joined: Mon Oct 19, 2009 10:34 am

Re: Error "shifted too far"

Post by ChrisCharlton »

There's no obvious reason why this would be happening here. Could your provide some information about your model hierarchy (see the Model->Hierarchy viewer window), i.e. number of units at each level, range of unit size, etc. If it was possible for you to send me a worksheet with a similar structure that demonstrates the problem, so that I could investigate it further, then that would also be useful.
gilleecl
Posts: 27
Joined: Fri Dec 17, 2010 11:24 am

Re: Error "shifted too far"

Post by gilleecl »

hi Chris, there are 3125 level 1 units and 190 level 2 units. The following steps lead to the error:
- Open MlwiN
- File > New worksheet
- File > Import > Import > Browse to file
- Note: 'Some input variables held data in more precision than MlwiN supports, these have been rounded > OK
- Click on y: Select 'reading'. N levels: 2-ij. level 2=IDCLASS; level 1=pupid. Done.
- (0 of 0 cases in use)
- Click 'Notation', unselect 'general'. Done.
Then errors occur, e.g. "shifted too far"; "shifted too far(set)" or on one occasion, no error message but only 0 of 1 cases in use and then error message that MlwiN needs to close.

Many thanks for your help.
Lorraine
Attachments
Reading.xls
(177.8 KiB) Downloaded 784 times
ChrisCharlton
Posts: 1348
Joined: Mon Oct 19, 2009 10:34 am

Re: Error "shifted too far"

Post by ChrisCharlton »

Thank you for the additional information - I have now been able to replicate the error that you were seeing and this should be fixed in the next release of MLwiN. In the mean time you can work around the error by either setting notation to simple prior to setting up the response and level structure or by creating a variable named cons of the same length of the rest of your data prior to changing the notation.
gilleecl
Posts: 27
Joined: Fri Dec 17, 2010 11:24 am

Re: Error "shifted too far"

Post by gilleecl »

Many thanks for your help.

Lorraine
chinmaychien
Posts: 1
Joined: Tue Jan 26, 2016 5:44 am

Re: Error "shifted too far"

Post by chinmaychien »

Hi,

I am facing the same error too as I am fitting a 3-level model for binary response using the PQL2 estimation procedure. What could be the possible reason?

thanks!

May Chien
ChrisCharlton
Posts: 1348
Joined: Mon Oct 19, 2009 10:34 am

Re: Error "shifted too far"

Post by ChrisCharlton »

Can you confirm that you are using the most recent release of MLwiN (2.35)? If so then it would be useful if you can provide an example with a structure that replicates the behaviour so that I can look into it further.
kimjuuul
Posts: 1
Joined: Mon Sep 28, 2020 11:30 pm

Re: Error "shifted too far"

Post by kimjuuul »

Hi,

I am using version 3.00 and having the same issue. I am trying cross-classification MCMC model with large data and 4-levels (5-levels including cross-classifying levels). I have 234,216 observations in my data in total.

I was able to set the models in the equation window. However, when I run IGLS with my model, "shifted too far" suddenly appears and the MLwiN quit.

Could it be related to the size of the data or the worksheet? I tried to 1) manipulate the size of the worksheet as well but it did not help, and 2) take a random sample of the data (decreased down to 234 observations). I would really appreciate if you could let me know how to address this issue.
Post Reply