Re: Guidelines or advice for data management? [correction]

[John F Hall]

The link in my earlier message link was inadvertently author only access to
the Weebly site.  The correct link is:
http://surveyresearch.weebly.com/guide-to-spss-tutorials.html

----- Original Message -----
From: John F Hall
To: Robert Lundqvist ; [hidden email]
Sent: Tuesday, November 30, 2010 10:06 AM
Subject: Re: Guidelines or advice for data management?


Robert

Have a look at the entry-level, syntax-based SPSS tutorials on my website,
especially 1.3.3.9  Some general advice on file building in SPSS.

The tutorials are specific to the management and analysis of data from
questionnaire surveys and are intended for newbies, students, researchers,
tutors and support staff.  Most of them use SPSS 18, but some are still in
15.  All syntax examples will work with releases back to 11, but the GUI
menu examples are all from 18.  There are around 400 pages altogether and
everything is available for free download.

There nothing specific about which files to save and how, but you'll pick up
a lot of good practice by reading through the tutorials right from the very
beginning and noting the steps taken to create and save syntax and saved
files (and the occasional hint on professional practice).  There's nothing
about how to organise folders, but this is implicit in the steps to create
folders and sub-folders as the tutorials progress step-by-step (with
full-colour screenshots at each step) through the entire research process
from data capture to statistical analysis.

John Hall
[hidden email]
http://surveyresearch.weebly.com


.  ----- Original Message -----
From: Robert Lundqvist
To: [hidden email]
Sent: Tuesday, November 30, 2010 9:09 AM
Subject: Guidelines or advice for data management?


As a relatively new SPSS user, I have still not found a stable way to
organize my files and away to deal with all datasets. What files should be
saved and when, and what is the best way to protect the original data and
not end up with loads of files? Some people have told me that they never
save any variable or value labels, but prefer to keep the basic file
unchanged. For each analysis they repeat a syntax sequence to create those
lables. Would this be considered as a good approach? Furthermore, if it was
my decision, I could probably find a way that suits my personal needs.
However, I am in the position where less experienced users come to me for
advice in statistical matters, and what should I tell them? What would be a
good "data handling praxis" for someone who don't want to use syntax?

There are obviously no correct answer. We are all different, and the needs
also differs a lot depending on the actual kind of data. I would still
appreciate either suggestions or pointers to links or literature.

Robert
*****************
Robert Lundqvist
Norrbotten regional council
Sweden

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

[Bianco, Joseph]

Hello Listers,

Apologies for the remedial nature of this question, but here goes:

I'm conducting a multinomial logistic regression (3 levels of the DV) and, based on significant chi-squares/ANOVAs, would like to enter predictors hierarchically (i.e., in blocks, as in binary logistic regression and regular linear regression). On SPSS, that's not an option for multinomial; instead, you can "force entry" certain predictors and then add others in a stepwise fashion.  What I'm not clear about is how to conduct this such that I force entry of specific predictors in three iterations. If this were a binary logistic regression, I would do this:

Block 1:  demographic predictors (age, education)
Block 2:  treatment factors (treatment condition [2 levels], therapy attendance)
Block 3:  psychotherapy process variables (alliance, group cohesion, group similarities)

In other words, I'd like to see if/how the psychotherapy process variables predict group membership when predictors from Blocks 1-2 are controlled for. I hope this makes sense. Any advice will be appreciated.

Best,
Joe



*************************************
Joseph A. Bianco, Ph.D.
Department of Geriatric Medicine/Gerontology
Ohio University College of Osteopathic Medicine
353 Grosvenor
Athens, OH 45701
[phone]
Email: [hidden email]
Web: http://ohio.academia.edu/JosephBianco

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

[Bruce Weaver]

Bianco, Joseph wrote

Hello Listers,

Apologies for the remedial nature of this question, but here goes:

I'm conducting a multinomial logistic regression (3 levels of the DV) and, based on significant chi-squares/ANOVAs, would like to enter predictors hierarchically (i.e., in blocks, as in binary logistic regression and regular linear regression). On SPSS, that's not an option for multinomial; instead, you can "force entry" certain predictors and then add others in a stepwise fashion.  What I'm not clear about is how to conduct this such that I force entry of specific predictors in three iterations. If this were a binary logistic regression, I would do this:

Block 1:  demographic predictors (age, education)
Block 2:  treatment factors (treatment condition [2 levels], therapy attendance)
Block 3:  psychotherapy process variables (alliance, group cohesion, group similarities)

In other words, I'd like to see if/how the psychotherapy process variables predict group membership when predictors from Blocks 1-2 are controlled for. I hope this makes sense. Any advice will be appreciated.

Best,
Joe

Hi Joe.  It's been a while since I used NOMREG, but I just took a quick look at it, and it would appear that you have to run a separate NOMREG command for each step in your hierarchical model building.  (This is the case for MIXED too, by the way).  

One issue you have to be aware of is that when there are more variables in the later models, you could lose cases that appeared in the earlier models, due to missing data on later added variables.  That's a problem, because the comparison of nested models via the change in -2LL is only valid when all models are computed using exactly the same cases.  A very useful  trick you can use to get around that problem is to list ALL of the variables that will appear in the final model each time, but limit which ones are used in a particular model on the /MODEL sub-command.  E.g.,

* Step 1 .
NOMREG Y (BASE=FIRST ORDER=ASCENDING)
            BY { all categorical variables in the final model }
            WITH { all continuous variables in the final model }
  /MODEL = { all variables forced in on step 1 }
  /INTERCEPT=INCLUDE
  /PRINT=PARAMETER SUMMARY LRT CPS STEP MFI IC.

* Step 2 .
NOMREG Y (BASE=FIRST ORDER=ASCENDING)
            BY { all categorical variables in the final model }
            WITH { all continuous variables in the final model }
  /MODEL = { all step 1 variables plus all step 2 variables }
  /INTERCEPT=INCLUDE
  /PRINT=PARAMETER SUMMARY LRT CPS STEP MFI IC.

* Step 3 .
NOMREG Y (BASE=FIRST ORDER=ASCENDING)
            BY { all categorical variables in the final model }
            WITH { all continuous variables in the final model }
  /MODEL = { all step 1 & 2 variables plus all step 3 variables }
  /INTERCEPT=INCLUDE
  /PRINT=PARAMETER SUMMARY LRT CPS STEP MFI IC.

You'll have to compute your own tests on the change in -2LL from one step to the next.  Remember that it is distributed (approximately) as chi-squared with df = the change in the number of model parameters from one step to the next.

HTH.