www.cmm.bristol.ac.uk/forum

Apologies for flooding the forum, but I've met an error when requesting residuals for level 1 and level 2 in a normally distributed strictly hierarchical model. It only says "MLwiN has stopped working" or with runmlwin's batch: "MLNScript has stopped working". Also tried experimenting with changing the max worksheet cell size, tempmat, etc. The model estimation works fine. Since I haven't had an issue with this before I suspect it has to do with my large dataset? #clusters: L1: 301,000; L2: 14,000; L3: 9,000; L4: 58. Using MLwiN 2.30.

Have you tried asking for just the level-2 residuals as these should be much shorter, so would give an indication of whether it is a memory issue? It might also be worth removing the nopause option and then requesting the residuals via the menus to see if you get the same error there. If you were able to send me a worksheet with a similar structure that also has this problem then I could attempt to investigate the exact cause.

Forgot to write that; I've tried only level 2, with and without batch, nopause, regular GUI (menus) and so on. I'll send you a dropbox link by email Chris, thanks