www.cmm.bristol.ac.uk/forum

Dear Forum,

I would have another question regarding the meaning of the "cases in use" bracket in the lower right side of the equations window.
I am currently analysing data from a diary study where several predictors have been measured 3 times per day for five days accumulating to a total sample of 660 measurement points.
In the same time period my DV's have only been measured once at the end of each of the five days accumulating to a sample of 201 measurement points. I declared the other 2 time points each day as "missing values" for the respective outcome variables.

Now when I model data with these daily DV's the lower right side of the equation window reads "201 of 660 cases in use". Does this mean that MLwiN merely takes into consideration the last measurement of each day since this also incorporates the DV?
What should I do to overcome this issue? Should I just paste the same value of the DV for all 3 daily measurement occasions? Or should I average the other variables to the day level (e.g. average three measurement occasions to one) so that then all variables refer to the day level?

Thank you very much for your ideas!

Jake

Unless you are doing imputation MLwiN will ignore any rows where there are missing values for data used in the model when calculating estimates. I suspect that how you should handle the missing values will depend on what the variables are and what information you are trying to get out of the model, for example whether they likely to vary over the day or if time of day is of interest.

Thank you for your reply. It was very helpful indeed!