Pair Mating

[John Norton]

Hi Listers,
 
Can anyone suggest a way to match cases within the same data set into pairs?  I have a file of cases who fall into two groups: those who received a transplant from a related donor, and those who received theirs from an unrelated donor.  I need to match the cases in one group with cases in the other on a 1:1 basis using 3 variables, which are disease type (categorical), disease status (categorical) and age (scale, and is considered a match when age is within 3 years of each other).  Finally, I'll need to create a pair ID.
 
Might anyone have an programming example for how this could be done in SPSS?
 
Many thanks in advance,
 
 

[Swank, Paul R]

You might try propensity scores. Use logistic regression with the three
matching variables as predictors and the group as the DV. Then the
predicted scores are referred to as propensity scores. Then you match on
the propensity scores.


Paul R. Swank, Ph.D.
Professor, Developmental Pediatrics
Director of Research, Children's Learning Institute
Medical School
University of Texas Health Science Center at Houston

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
John Norton
Sent: Friday, September 29, 2006 11:52 AM
To: [hidden email]
Subject: Pair Mating

Hi Listers,

Can anyone suggest a way to match cases within the same data set into
pairs?  I have a file of cases who fall into two groups: those who
received a transplant from a related donor, and those who received
theirs from an unrelated donor.  I need to match the cases in one group
with cases in the other on a 1:1 basis using 3 variables, which are
disease type (categorical), disease status (categorical) and age (scale,
and is considered a match when age is within 3 years of each other).
Finally, I'll need to create a pair ID.

Might anyone have an programming example for how this could be done in
SPSS?

Many thanks in advance,

[Dan Zetu]

I second Paul's advice. Propensity matching is ideal for your situation. In
this case, propensity score is the probability of havong received a
transplant (a logistic regression output). If you need a 1:1 match, you sort
both groups in descending order of propensity score and match accordingly.

Dan

>From: "Swank, Paul R" <[hidden email]>
>Reply-To: "Swank, Paul R" <[hidden email]>
>To: [hidden email]
>Subject: Re: Pair Mating
>Date: Mon, 2 Oct 2006 09:32:26 -0500
>
>You might try propensity scores. Use logistic regression with the three
>matching variables as predictors and the group as the DV. Then the
>predicted scores are referred to as propensity scores. Then you match on
>the propensity scores.
>
>
>Paul R. Swank, Ph.D.
>Professor, Developmental Pediatrics
>Director of Research, Children's Learning Institute
>Medical School
>University of Texas Health Science Center at Houston
>
>-----Original Message-----
>From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
>John Norton
>Sent: Friday, September 29, 2006 11:52 AM
>To: [hidden email]
>Subject: Pair Mating
>
>Hi Listers,
>
>Can anyone suggest a way to match cases within the same data set into
>pairs?  I have a file of cases who fall into two groups: those who
>received a transplant from a related donor, and those who received
>theirs from an unrelated donor.  I need to match the cases in one group
>with cases in the other on a 1:1 basis using 3 variables, which are
>disease type (categorical), disease status (categorical) and age (scale,
>and is considered a match when age is within 3 years of each other).
>Finally, I'll need to create a pair ID.
>
>Might anyone have an programming example for how this could be done in
>SPSS?
>
>Many thanks in advance,

[John Norton]

Thanks Paul and Dan!  I'll give this a try.
 
JN

>>> Dan Zetu <[hidden email]> 10/2/2006 10:02:18 AM >>>

I second Paul's advice. Propensity matching is ideal for your situation. In
this case, propensity score is the probability of havong received a
transplant (a logistic regression output). If you need a 1:1 match, you sort
both groups in descending order of propensity score and match accordingly.

Dan

>From: "Swank, Paul R" <[hidden email]>
>Reply-To: "Swank, Paul R" <[hidden email]>
>To: [hidden email]
>Subject: Re: Pair Mating
>Date: Mon, 2 Oct 2006 09:32:26 -0500
>
>You might try propensity scores. Use logistic regression with the three
>matching variables as predictors and the group as the DV. Then the
>predicted scores are referred to as propensity scores. Then you match on
>the propensity scores.
>
>
>Paul R. Swank, Ph.D.
>Professor, Developmental Pediatrics
>Director of Research, Children's Learning Institute
>Medical School
>University of Texas Health Science Center at Houston
>
>-----Original Message-----
>From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
>John Norton
>Sent: Friday, September 29, 2006 11:52 AM
>To: [hidden email]
>Subject: Pair Mating
>
>Hi Listers,
>
>Can anyone suggest a way to match cases within the same data set into
>pairs?  I have a file of cases who fall into two groups: those who
>received a transplant from a related donor, and those who received
>theirs from an unrelated donor.  I need to match the cases in one group
>with cases in the other on a 1:1 basis using 3 variables, which are
>disease type (categorical), disease status (categorical) and age (scale,
>and is considered a match when age is within 3 years of each other).
>Finally, I'll need to create a pair ID.
>
>Might anyone have an programming example for how this could be done in
>SPSS?
>
>Many thanks in advance,