Typically use MLwiN for recoding, or other stat pkg?

[mlm4tw]

Hi! (This is my first posting...)

I'm teaching myself MLwiN. Used to be a SPSS user, then Stata, but my new R.A. work entails MLM-ing, so...

I tend to prefer command-line useage rather than the GUI, so I've started by reading the Command Manual first. The structure of the command language seems to be more ''stripped down'' than Stata's or SPSS's -- I'm still trying to get my head around it.

I see two possible routes for myself: One is to do all my data cleaning and recoding in Stata, and save MLwiN for just the actual multilevel models. This means a shorter learning curve (just familiarize myself with the GUI and call it good). But it means that if I think of additional recodes I need to do, I'll have to do a **lot** of clicking (with the GUI), or else export it back to Stata, do the manipulations, and re-import it into MLwiN.

The other option is to tackle the learning curve, ''go native'' -- and once the data is into MLwiN, just do all my manipulations there.

Any thoughts/insights/experiences you're willing to share? (I'd also be interested in what proportion of users on this forum use the command language, versus just doing the GUI.)


Thanks!

[mlm4tw]

Hmm -- 19 views and no replies.

May I infer that I'm the only one foolhearty enough to try using MLwiN primarily thorough the command language? :)


--Travis

[mlm4tw]

Sorry -- my initial posting actually asked two distinct questions. No responses yet, so let me revise... :)

1) How many of you use MLwiN for your recoding -- and how many do all your data cleaning and manipulating in another stats package, and then just import it to MLwinN just for the specific MLM analyses and graphing? Thanks!

(And: Did you start one way, but then drift over to another -- e.g. started by just importing to MLwiN, but then gradually did more and more of your data manipulations natively?)


Just a bit of a straw poll. :)


(The other question, I'll put in a new thread with a more descriptive subject line.)

[bosawas]

Like some of the folks who are directly associated with CMM, I tend to use MLwiN only for the statistical analysis, not for routine data management.