Search found 111 matches

by billb
Wed Aug 23, 2017 10:37 am
Forum: MLwiN user forum
Topic: Goodness of fit test for Multilevel logistic models
Replies: 2
Views: 993

Re: Goodness of fit test for Multilevel logistic models

Hi ddswick,
To confirm the Hosmer-Lemeshew test has not been implemented in MLwiN. In the multilevel setting there are residuals available and the diagnostics that go with testing these are available as in normal models.
Hope that answers your question.
Best wishes,
Bill.
by billb
Wed Aug 23, 2017 10:33 am
Forum: R2MLwiN user forum
Topic: Interpretation of coefficient in multilevel cumulative logit model
Replies: 4
Views: 1596

Re: Interpretation of coefficient in multilevel cumulative logit model

Hi Vivian, Thanks for your question. I wonder whether you can clarify a bit? When you talk about reversing the sign of the coefficient is this because of the base category you have chosen in the cumulative logit? I.e. if you are using the lowest category as base then the values for the thresholds wo...
by billb
Wed Aug 09, 2017 8:15 am
Forum: runmlwin user forum
Topic: Advice on cross-classified models with large data sets
Replies: 11
Views: 2257

Re: Advice on cross-classified models with large data sets

Hi RDMCDOWELL, If you fit a cross-classified model then it's best to use MCMC in MLwiN and so you would have higher levels/classifications for doctors and patients. I wouldn't suggest you use the clustering as described. I think the IGLS method would really struggle. The multiple membership approach...
by billb
Wed Aug 09, 2017 8:10 am
Forum: R2MLwiN user forum
Topic: Explained variance logistic/ordered multinomial models estimated with MCMC [VPC?]
Replies: 3
Views: 766

Re: Explained variance logistic/ordered multinomial models estimated with MCMC [VPC?]

Dear Gerine, I will agree with Kelvyn that this is indeed a challenging problem due to the scaling that occurs. I don't know Tony Fielding's paper that well so don't think this has been implemented in MLwiN unless he wrote macros himself. With regard VPC estimation there are 2 references (Goldstein,...
by billb
Wed Aug 09, 2017 8:03 am
Forum: R2MLwiN user forum
Topic: Wald Test in MCMC estimation
Replies: 3
Views: 1173

Re: Wald Test in MCMC estimation

Hi Gerine, A better approach would be to extract the chains for the 2 dummies and subtract their values as this would then give you a chain for the differences. You could then look to see if the quantiles of this chain contain 0. This is better as it will deal with any correlation between the parame...
by billb
Fri Jun 02, 2017 9:23 am
Forum: MLwiN user forum
Topic: low unexplained variance on third level, but third level variable with high explanatory power
Replies: 1
Views: 784

Re: low unexplained variance on third level, but third level variable with high explanatory power

Hi Marjolin,
Don't see any problem here - your variables appears to explain 2 thirds of the level 3 variation and the Z score is only 4.3 which although very significant you do have 6000 data points and so it is possible to have much bigger Z scores.
Best wishes,
Bill.
by billb
Wed May 10, 2017 2:29 pm
Forum: runmlwin user forum
Topic: test for a difference in Bayesian DIC
Replies: 2
Views: 1033

Re: test for a difference in Bayesian DIC

Hi Zuzanna, The DIC is an information criterion like the AIC and BIC in classical statistics and basically such quantities do not have an underlying statistical distribution and so the concept of significance doesn't work for such measures. It is therefore just the case that rules of thumb are used ...
by billb
Wed Apr 19, 2017 9:45 am
Forum: MLwiN user forum
Topic: Assumptions cross-classification models
Replies: 1
Views: 828

Re: Assumptions cross-classification models

Hi Griet,
In theory a nested structure is a special case of the more general crossed-structure with no crossing !! So yes there is no restriction here and if you tried selecting cross-classified for a nested structure you should get the same answers.
Bill.
by billb
Wed Mar 01, 2017 1:23 pm
Forum: MLwiN user forum
Topic: Classification order in cross-classified multilevel models
Replies: 1
Views: 1087

Re: Classification order in cross-classified multilevel models

Hi Griet, It shouldn't matter but (1) you should make sure you tick the cross-classified button so that it is fitting the XC model! (2) you should ensure you run the MCMC estimation for sufficiently long that the chains converge to similar answers - they will not be exactly the same due to the Monte...
by billb
Fri Feb 10, 2017 9:58 am
Forum: MLwiN user forum
Topic: multilevel-analysis, predictor
Replies: 4
Views: 1322

Re: multilevel-analysis, predictor

Hi Ine, What you are proposing would be a false replication i.e. you will be pretending you have more data than you have (particularly if you didn't put in the level 2 random effect). In fact once you put the random effect in unsurprisingly it will show all the variance is at level 2 and might in so...