SPSSX Discussion

reciprocal relationship

Classic

List

Threaded

5 messages Options

pseudoyou

reciprocal relationship

Hi,

Could anyone tell me how to use SPSS to confirm a reciprocal relationship?
I only have two variables and I would like to statistically test if they are
reciprocally affecting each other.

Thank You
Jon

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

David Greenberg

Re: reciprocal relationship

If you only have observations on two variables at a single point in time, withno a priori knowledge of the nature of a relationship between them, there is no statistical
technique that will tell you whether there are reciprocal influences. Find any statistics textbook that deals with simultaneous equation systems and you will find out why. David Greenberg, Sociology Department, New York University

----- Original Message -----
From: pseudoyou <[hidden email]>
Date: Friday, February 8, 2008 8:06 pm
Subject: reciprocal relationship
To: [hidden email]

> Hi,
>
> Could anyone tell me how to use SPSS to confirm a reciprocal relationship?
> I only have two variables and I would like to statistically test if
> they are
> reciprocally affecting each other.
>
> Thank You
> Jon
>
> =====================
> 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

Bob Schacht-3

Re: reciprocal relationship

At 03:42 PM 2/8/2008, David Greenberg wrote:
>If you only have observations on two variables at a single point in time,
>withno a priori knowledge of the nature of a relationship between
>them, there is no statistical technique that will tell you whether there
>are reciprocal influences. Find any statistics textbook that deals with
>simultaneous equation systems and you will find out why. David Greenberg,
>Sociology Department, New York University

Amplifying on Greenberg's point just a little bit: If this is all you have
(i.e., no info on nature of the relationship, no temporal time structure to
tell you which comes first, etc.), the most you can hope for is
"correlation". Just to be clear, correlation does not imply causation, and
does not tell you direction of influence, and does not even assure you that
the relationship is direct (it could be mediated through a third variable).
But given your data, correlation seems to be as much as you are going to
get with respect to "reciprocal relationship". At least correlation can
tell you the likelihood that there is any relationship at all.

Bob Schacht

>----- Original Message -----
>From: pseudoyou <[hidden email]>
>Date: Friday, February 8, 2008 8:06 pm
>Subject: reciprocal relationship
>To: [hidden email]
>
>
> > Hi,
> >
> > Could anyone tell me how to use SPSS to confirm a reciprocal relationship?
> > I only have two variables and I would like to statistically test if
> > they are
> > reciprocally affecting each other.
> >
> > Thank You
> > Jon
> >
> > =====================
> > 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

Robert M. Schacht, Ph.D. <[hidden email]>
Pacific Basin Rehabilitation Research & Training Center
1268 Young Street, Suite #204
Research Center, University of Hawaii
Honolulu, HI 96814

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

pseudoyou

Re: reciprocal relationship

Actually, I have a varible called "DM", which is a sum of three scales,
"mon", "emo"
and "cog". I want to know whether there is a reciprocal relationship
between "DM" and another variable "TIT".
I used 2SLS and specified "DM" as the predictor, while "TIT" as the
dependent variable. "mon" "emo" and "cog" are three instruments. Did I do
it correctly? The output I had is similar to MLS but I don't know whether
the result confirmed the reciprocal relationship.

Also, I think SEM can deal with a non-recursive model like this, but I would
like to know if it can be tested on SPSS.

Thanks
Jon

On Feb 8, 2008 9:31 PM, Bob Schacht <[hidden email]> wrote:

> At 03:42 PM 2/8/2008, David Greenberg wrote:
> >If you only have observations on two variables at a single point in time,
> >withno a priori knowledge of the nature of a relationship between
> >them, there is no statistical technique that will tell you whether there
> >are reciprocal influences. Find any statistics textbook that deals with
> >simultaneous equation systems and you will find out why. David Greenberg,
> >Sociology Department, New York University
>
> Amplifying on Greenberg's point just a little bit: If this is all you have
> (i.e., no info on nature of the relationship, no temporal time structure
> to
> tell you which comes first, etc.), the most you can hope for is
> "correlation". Just to be clear, correlation does not imply causation, and
> does not tell you direction of influence, and does not even assure you
> that
> the relationship is direct (it could be mediated through a third
> variable).
> But given your data, correlation seems to be as much as you are going to
> get with respect to "reciprocal relationship". At least correlation can
> tell you the likelihood that there is any relationship at all.
>
> Bob Schacht
>
>
> >----- Original Message -----
> >From: pseudoyou <[hidden email]>
> >Date: Friday, February 8, 2008 8:06 pm
> >Subject: reciprocal relationship
> >To: [hidden email]
> >
> >
> > > Hi,
> > >
> > > Could anyone tell me how to use SPSS to confirm a reciprocal
> relationship?
> > > I only have two variables and I would like to statistically test if
> > > they are
> > > reciprocally affecting each other.
> > >
> > > Thank You
> > > Jon
> > >
> > > =====================
> > > 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
>
> Robert M. Schacht, Ph.D. <[hidden email]>
> Pacific Basin Rehabilitation Research & Training Center
> 1268 Young Street, Suite #204
> Research Center, University of Hawaii
> Honolulu, HI 96814
>
> =====================
> 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
>

Peck, Jon

Re: reciprocal relationship

If I understand your model, it is
dm = mon + emo + cog
tit = a + b * dm + e1
dm = c + d * tit + e2

As it stands, this model is underidentified, so no estimation technique can answer your question. The accounting identity for dm doesn't give you a proper set of instruments.

HTH,
Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of pseudoyou
Sent: Friday, February 08, 2008 7:55 PM
To: [hidden email]
Subject: Re: [SPSSX-L] reciprocal relationship

Actually, I have a varible called "DM", which is a sum of three scales,
"mon", "emo"
and "cog". I want to know whether there is a reciprocal relationship
between "DM" and another variable "TIT".
I used 2SLS and specified "DM" as the predictor, while "TIT" as the
dependent variable. "mon" "emo" and "cog" are three instruments. Did I do
it correctly? The output I had is similar to MLS but I don't know whether
the result confirmed the reciprocal relationship.

Also, I think SEM can deal with a non-recursive model like this, but I would
like to know if it can be tested on SPSS.

Thanks
Jon

On Feb 8, 2008 9:31 PM, Bob Schacht <[hidden email]> wrote:

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