Weighted vs unweighted banners for T & Z significance tests.

[Mark Webb-3]

What is best market research practice ?

Run Z & T significance tests on weighted or unweighted banners ?

Please give reasoning with your replies.



Thanks in advance.

Mark

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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

[zstatman]

This has come up several times and the bottom line is to run sig tests on
unweighted data and is explained very well in Hector Maletta's paper on
weighting at Ray Levesque's site, SPSSTools

WMB
Statistical Services

============
mailto: [hidden email]
http:\\home.earthlink.net\~info.statman
============


On 10/27/2008 10:40:10 AM, Mark Webb ([hidden email]) wrote:

> What is best market research practice ?
>
> Run Z & T significance tests on weighted or unweighted banners ?
>
> Please give reasoning with your replies.
>
>
>
> Thanks in advance.
>
> Mark
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (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

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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

[Hector Maletta]

Not exactly on unweighted data, but on data weighted with non-inflationary
weights, i.e. weights that correct proportional imbalances but the sum of
which adds to the total sample size. This is, however, a rough-and-tumble
solution, especially when the sample design involves clustering. For better
solutions use the Complex Samples module in SPSS, or analogous facilities in
SUDAAN and other softwares.
It is all explained in the paper mentioned by Statman.

Hector

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Statmanz
Sent: 27 October 2008 12:57
To: [hidden email]
Subject: Re: Weighted vs unweighted banners for T & Z significance tests.

This has come up several times and the bottom line is to run sig tests on
unweighted data and is explained very well in Hector Maletta's paper on
weighting at Ray Levesque's site, SPSSTools

WMB
Statistical Services

============
mailto: [hidden email]
http:\\home.earthlink.net\~info.statman
============


On 10/27/2008 10:40:10 AM, Mark Webb ([hidden email]) wrote:

> What is best market research practice ?
>
> Run Z & T significance tests on weighted or unweighted banners ?
>
> Please give reasoning with your replies.
>
>
>
> Thanks in advance.
>
> Mark
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (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

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (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