Page 1 of 1

Filtering rows to fit in existing equations

Posted: Sun Apr 07, 2019 12:49 pm
by rimberdrimberd
Hello! :)

I have a dataset (n=500), where I need to fit and store around 50+ different equations. (I have 10 depentend variables, which I need to check with different combinations of explanatory variables.) After the set up of these equations and assessment of the estimators as well as the check of assumptions, I am planning to filter rows where assumptions are set in my data to make my results more reliable

Is there a way to leave my prior 50 equations as is and to change my filtered rows where I then automatically see the change in my equations result?

What would be the best approach to make it least time-consuming? At the moment I only saw a way of building up the equations again with a totally new dataset, which would mean that I have to to is all again.

Thank you for your help. :P I am new here I hope my question does not come across to naive.. :oops:

Re: Filtering rows to fit in existing equations

Posted: Mon Apr 08, 2019 11:12 am
by ChrisCharlton
You can filter the rows included within the model via the options button in Model->Hierarchy Viewer, or via the EXCL command (see the MLwiN help file). The model results will not automatically change after doing this, so you will need to re-run the model.

If you are fitting a series of models you probably want to do this via macros in MLwiN, or via R2MLwiN or runmlwin ( if you are more comfortable with those environments. Within a worksheet you can only have one live equation at a time, however you can store the models results for later comparison via the "stored model results" functionality. For examples of running and comparing models this way within MLwiN see chapter 4 of the manual supplement ( ... nuals.html).