Correctness of using Split File in Paired Sample T Test

[Mark Webb-5]

  I have before/after data [5 point scale] in the same record.
The intervention was a seminar.
I want to compare after to before averages on a batch of scale variables
that were measured before & after and determine the significance of the
differences.
For the total sample, one chooses the before & after variables and run
Paired samples.
But now I want to see if there is any difference between pre/post and sex.
Can I split the file on sex and run the Paired sample test as before ?

It works in SPSS - but is this logic correct ?
Regards

--
Mark Webb

Line [phone]
Cell [phone]
Fax to email [phone]
Skype  webbmark
Email  [hidden email]

=====================
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

[Alex Reutter]

It would be better to use MIXED or GLM; I don't think you can adequately answer the question of whether sex makes a difference using SPLIT FILE and T-TEST.  The online case studies work through a situation similar to what you describe below; see Help > Case Studies... then

* Advanced Statistics Option > Linear Mixed Models > Using Linear Mixed Models to Analyze Repeated Measurements, which works through the dietstudy.sav dataset that ships with the product, including restructuring the data to work with the MIXED procedure instead of T-TEST

* Advanced Statistics Option > Multivariate General Linear Modeling > GLM Repeated Measures > A Doubly Multivariate Analysis of Variance, which works through the dietstudy.sav dataset that ships with the product.  

Alex

From:	Mark Webb <[hidden email]>
To:	[hidden email]
Date:	08/21/2010 05:31 AM
Subject:	Correctness of using Split File in Paired Sample T Test
Sent by:	"SPSSX(r) Discussion" <[hidden email]>

---

  I have before/after data [5 point scale] in the same record.
The intervention was a seminar.
I want to compare after to before averages on a batch of scale variables
that were measured before & after and determine the significance of the
differences.
For the total sample, one chooses the before & after variables and run
Paired samples.
But now I want to see if there is any difference between pre/post and sex.
Can I split the file on sex and run the Paired sample test as before ?

It works in SPSS - but is this logic correct ?
Regards

--
Mark Webb

Line [phone]
Cell [phone]
Fax to email [phone]
Skype  webbmark
Email  [hidden email]

=====================
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