Fw: [SPSSX-L] RIM weighting

[Jon K Peck]

The current version, which you can get from the Extension Commands Collection on the SPSS Community site, allows up to 10 variables.  But be careful, the larger the number of dimensions, the longer it takes and the greater the likelihood of extreme weights.

More than 10 would be really extreme, so you might want to combine some dimensions and/or categories.  Bear in mind that ten variables each with four categories implies a table of over 1,000,000 cells.

If you really have more than 10 but you want to proceed, you could do the least important ones first and feed those weights into another application of the command.  That would disturb the first set of weights, but it might give you a result closer to what you want.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        "pm." <[hidden email]>
To:        Jon K Peck/Chicago/IBM@IBMUS
Date:        04/05/2012 02:10 PM
Subject:        Re: [SPSSX-L] RIM weighting

---

Thank you, but this "calculates case weights to match control totals for categories of one to five variables." I have more than five variables. :S I need a loop.

2012/4/5 Jon K Peck <peck@...>
Just use the SPSSINC RAKE extension command for this.  Available from the SPSS Community website (www.ibm.com/developerworks/spssdevcentral).  It is included in the Python Essentials, which can be found under Downloads for SPSS Statistics.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@...
new phone: <a href="tel:[phone]" target=_blank>720-342-5621




From:        MariaPetsch <maria.pecs@...>
To:        [hidden email]
Date:        04/05/2012 01:53 PM
Subject:        [SPSSX-L] RIM weighting
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

/<small>(First: sorry for my english, I'm from Hungary.)/</small>

I got a database and I have to do a rim weighting in it. I need more than 10
iterations, and I have to write a loop, but I can't. I mean I don't know
how. 'Cause numbers are changing in every iteration...
Any solutions?

*Without loop my syntax:*
COMPUTE help=universe/sample.
EXECUTE.

WEIGHT by help.

fre var1.

if (var1 = 1) weight = var1univ / var1samp .
if (var1 = 2) weight = var1univ / var1samp .
EXECUTE.

compute weight = help * weight.
execute.

weight by weight.

fre var2.

if (var2 = 1) weight1 = var2univ / var2samp .
if (var2 = 2) weight1 = var2univ / var2samp.

compute weight = weight * weight1 .

weight by weight.

... and so on for lots of variables.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/RIM-weighting-tp5621247p5621247.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD