variables fro factor scores

[danl]

Dear  friends,

can explain me how can create one variable from factor scores. I know that Spss have posibilities but I don`t understand.
I read this expresion: variables x was create like factor scores from variables a, variables b, variables c ...

thanks,

Dan

[Hector Maletta]

You can simply use the SAVE subcommand in the FACTOR command, and this will
create one new variable for the factor scores of each extracted factor from
which you want a variable to be created. SPSS will assign them a standard
rootname by default unless you specify another standard rootname. If your
rootname is FACSCORE, the various factor score variables will be called
FACSCORE1, FACSCORE2, FACSCORE3, etc.
On the other hand, in the FACTOR command output there is a factor score
coefficient matrix showing the weight of each observed variable (in
standardized z-score form) in the calculation of each factor score. This way
you can reconstruct each factor score as a linear combination of observed
variables. Each factor score is itself a z-score variable, i.e. a variable
with zero mean and unit standard deviation.
Hector

-----Mensaje original-----
De: SPSSX(r) Discussion [mailto:[hidden email]] En nombre de danl
Enviado el: Wednesday, August 30, 2006 12:06 PM
Para: [hidden email]
Asunto: variables fro factor scores

Dear  friends,

can explain me how can create one variable from factor scores. I know that
Spss have posibilities but I don`t understand.
I read this expresion: variables x was create like factor scores from
variables a, variables b, variables c ...

thanks,

Dan
--
View this message in context:
http://www.nabble.com/variables-fro-factor-scores-tf2190526.html#a6060890
Sent from the SPSSX Discussion forum at Nabble.com.

[statisticsdoc]

Stephen Brand
www.statisticsdoc.com

Dan,

Following the procedure that Hector outlined in his reply earlier today,
SPSS will allow you to compute and save factor scores using the SAVE
subcommand under FACTOR.  You have some choices about the exact manner in
which the factor score estimates are computed (bear in mind that the scores
are estimates, the estimated scores contain error, and there are different
ways to estimate).  Your options are to estimate scores using Regression
(the default), Bartlett, and Anderson-Rubin.

A lot can be said about the different methods, but in a nutshell, regression
estimates assume that the factor scores are normally distributed, while
Bartlett and Anderson Rubin estimates do not. The core idea in regression
estimates is to minimize the discrepancy between the predicted and true
factor scores.  The regression method is the most common and is the default
method in SPSS.  If you use this estimate method, check to see that the
regression weights resemble the factor pattern, so your estimated factors
have similar conceptual meaning.  Anderson-Rubin is a modification of the
Bartlett procedure that seeks to ensure that the estimated factors are
orthogonal.

A case can also be made for simply computing the sum of all the items that
load on a given factor, giving each item unit weight.  While this may seem
like a very crude way to estimate factor scores, the case can be made that
factor weights can be quite sample specific, and so a simple linear
combination of items may be more robust than a weighted combination of
items.  Check on the internal consistency (Cronbach's alpha) of the
combination.

HTH,

Stephen Brand


For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of
danl
Sent: Wednesday, August 30, 2006 11:06 AM
To: [hidden email]
Subject: variables fro factor scores


Dear  friends,

can explain me how can create one variable from factor scores. I know that
Spss have posibilities but I don`t understand.
I read this expresion: variables x was create like factor scores from
variables a, variables b, variables c ...

thanks,

Dan
--
View this message in context:
http://www.nabble.com/variables-fro-factor-scores-tf2190526.html#a6060890
Sent from the SPSSX Discussion forum at Nabble.com.

[Todd McDonald]

Does SPSS have the ability to create matched groups using propensity scores?

  Todd





---------------------------------
How low will we go? Check out Yahoo! Messenger’s low  PC-to-Phone call rates.

[Dale Glaser]

Hi Todd..I'm accessing my memory bank here as it has been awhile since I created propensity scores, but there isn't an automatic procedure per se in SPSS to create propensity scores, but I recall using the logistic regression procedure (outcome: tx vs control) and based on the covariates of interest (e.g., age, cortisol, etc.) created/saved predicted probability scores and used that as the propensity score....however, I will defer to those who use those more frequently than I.............

  Dale

Todd McDonald <[hidden email]> wrote:
  Does SPSS have the ability to create matched groups using propensity scores?

Todd





---------------------------------
How low will we go? Check out Yahoo! Messenger’s low PC-to-Phone call rates.



Dale Glaser, Ph.D.
Principal--Glaser Consulting
Lecturer--SDSU/USD/CSUSM/AIU
3115 4th Avenue
San Diego, CA 92103
phone: [phone]
fax: [phone]
email: [hidden email]
website: www.glaserconsult.com

[Adam Troy-2]

Hi Todd,

I do this all the time and use syntax.  See reference below.  --Adam

Painter, J. (2004). Propensity matching via SPSS. Syntax for the SPSS software package.  Available online at: http://www.unc.edu/~painter/SPSSsyntax/propen.txt

_____________________________
Adam B. Troy, Ph.D.
Director of Research
GoalQuest, Inc.


----- Original Message -----
From: Todd McDonald <[hidden email]>
Date: Thursday, August 31, 2006 1:12 pm
Subject: propensity scoring

> Does SPSS have the ability to create matched groups using
> propensity scores?
>
>  Todd
>
>
>
>
>
> ---------------------------------
> How low will we go? Check out Yahoo! Messenger’s low  PC-to-Phone
> call rates.
>

[Dennis Deck]

Sometimes stratification, covariate, or weighting make more sense than
matching.

* Here we save the propensity score (CDp), modeling assignment to the
College Dreams program (CD=1 for tx group and 0 for comparison).

LOGISTIC
   Vars= CD with Male LowInc Minority NonTrad Transfer Absences HighRisk

                 GPA5 AvgAch
  /Save=PRED(CDp) /Classplot /External .


* Here we calculate quintiles (CDq) on the propensity score for use in
  stratification and evaluating overlap of the group distributions .

RANK CDp  /NTiles(5) INTO CDq  .


* Here we compute the inverse probability of treatment weights (IPTW)
based
  on the estimated propensity score .

DO IF (CD=1) .
+  COMPUTE iCDp = 1 / CDp .
ELSE .
+  COMPUTE iCDp = 1 / (1-CDp) .
END IF .

* Here we apply the weights in an outcome analysis predicting the odds
of HS
  graduation for treated vs untreated students  .

WEIGHT BY iCDp .
LOGISTIC
   Vars= Graduation with CD Male GPA5 HighRisk
     /Save= Pred(GradP) /Classplot /External .
WEIGHT OFF .


There is more to good propensity score analysis but the syntax is
straightforward - at least when using the PS to define strata,
covariate, or weights.   Matching is more complicated.

Dennis Deck, PhD
RMC Research Corporation
[hidden email]

-----Original Message-----
From: Todd McDonald [mailto:[hidden email]]
Sent: Thursday, August 31, 2006 10:13 AM
Subject: propensity scoring

Does SPSS have the ability to create matched groups using propensity
scores?

  Todd





---------------------------------
How low will we go? Check out Yahoo! Messenger's low  PC-to-Phone call
rates.