multiple mediation - regression & macro SPSS?

[Anter]

Hello!

 

I have a design with 2 mediators and 3 predictors, 1 outcome variable. 

 

1.The simple mediation analysis, each mediator with each predictor was significant. Now I want to test for the multiple mediation. How can I do this in SPSS? I understand that first I must verify that each IV positively predicts each mediator, controlling for the effect of the other IV's and the other mediator. If one of my IV's doesn't significantly predict a mediator do I use it in further analysis? Also controlling for would mean hierarchical regression, with IV's and mediator1 in step 1 and  the IV predicting mediator2 in step 2?

2. There is a macro on multiple mediation written by Andrew Hayes called Process (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html) but it gives me errors in SPSS 16. I'm not familiar with macros so I don't know if there's something I'm doing wrong or it's because of the older version of SPSS. My question is do I change any part of the code, where do I insert the name of my variables, after  !charend before it etc, do I use spaces, are there any more slashes or paranthesis. An error sounds something like :

 

Error # 6820 in column 27.  Text: AVER_TI_Z
>There is no left parenthesis after one of the DEFINE command parameter
>specifications, e.g.  after !TOKENS.
>This command not executed.

 

Does anyone know what it means?

 

This is the syntax:

 

define process (vars = !charend('/')/model = !charend('/') !default(75)/y = !charend('/')/m = !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(0)/covmy = !charend('/') !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/') !default(random)).

Any thoughts greatly welcomed.

Thanks much,

 

Andra

[David Marso]

2. Did you bother to read the instructions in the PDF file?
http://www.afhayes.com/public/process.pdf
DO NOT MODIFY THE CODE!  

From the instructions you did not read:
"Make sure you run the code exactly as downloaded in a syntax or program file.  Modifying the code will produce errors.  Do not modify the code at all. "
---
--

Anter wrote

Hello!
 
I have a design with 2 mediators and 3 predictors, 1 outcome variable. 
 
1.The simple mediation analysis, each mediator with each predictor was significant. Now I want to test for the multiple mediation. How can I do this in SPSS? I understand that first I must verify that each IV positively predicts each mediator, controlling for the effect of the other IV's and the other mediator. If one of my IV's doesn't significantly predict a mediator do I use it in further analysis? Also controlling for would mean hierarchical regression, with IV's and mediator1 in step 1 and  the IV predicting mediator2 in step 2?

2. There is a macro on multiple mediation written by Andrew Hayes called Process (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html) but it gives me errors in SPSS 16. I'm not familiar with macros so I don't know if there's something I'm doing wrong or it's because of the older version of SPSS. My question is do I change any part of the code, where do I insert the name of my variables, after  !charend before it etc, do I use spaces, are there any more slashes or paranthesis. An error sounds something like :
 
Error # 6820 in column 27.  Text: AVER_TI_Z
>There is no left parenthesis after one of the DEFINE command parameter
>specifications, e.g.  after !TOKENS.
>This command not executed.
 
Does anyone know what it means?
 
This is the syntax:
 
define process (vars = !charend('/')/model = !charend('/') !default(75)/y = !charend('/')/m = !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(0)/covmy = !charend('/') !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/') !default(random)).

Any thoughts greatly welcomed.
Thanks much,
 
Andra

[Anter]

Yes, I did bother to read the PDF and the others from mediation and some articles... However programming isn't something I know. I don't really understand where to put the name of my variables, if not in the syntax shown below? If I write the names here

INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal = n/contrast = t/percent = p/bc = b/bca = d */.

nothing happens, no output appears. This is why I asked if there are errors because of the older version of spss or because I am writing it wrong.

Thank you.

Andra

--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 3:21 PM

2. Did you bother to read the instructions in the PDF file?
http://www.afhayes.com/public/process.pdf
DO NOT MODIFY THE CODE!

From the instructions you did not read:
"Make sure you run the code exactly as downloaded in a syntax or program
file.  Modifying the code will produce errors.  Do not modify the code at
all. "
---
--
Anter wrote

>
> Hello!
> Â
> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
> Â
> 1.The simple mediation analysis, each mediator with each predictor was
> significant. Now I want to test for the multiple mediation. How can I do
> this in SPSS? I understand that first I must verify that each IV
> positively predicts each mediator, controlling for the effect of the other
> IV's and the other mediator. If one of my IV's doesn't significantly
> predict a mediator do I use it in further analysis? Also controlling for
> would mean hierarchical regression, with IV's and mediator1 in step 1 andÂ
> the IV predicting mediator2 in step 2?
>
> 2. There is a macro on multiple mediation written by Andrew Hayes called
> Process
> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html)Â but it
> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
> if there's something I'm doing wrong or it's because of the older version
> of SPSS. My question is do I change any part of the code, where do I
> insert the name of my variables, after  !charend before it etc, do I use
> spaces, are there any more slashes or paranthesis. An error sounds
> something like :
> Â
> Error # 6820 in column 27.  Text: AVER_TI_Z
>>There is no left parenthesis after one of the DEFINE command parameter
>>specifications, e.g.  after !TOKENS.
>>This command not executed.
> Â
> Does anyone know what it means?
> Â
> This is the syntax:
> Â
> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
> = !charend('/')/m = !charend('/')/x = !charend ('/')
>    /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
> !charend('/') !default(95)
>    /hc3 = !charend('/') !default(0)/cluster = !charend('/')
> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
> !charend('/') !default(999)/
>    vmodval = !charend('/') !default (999)/qmodval = !charend('/')
> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
> !charend('/') !default (999)
>   /boot = !charend('/') !default(1000)/center = !charend('/')
> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
> !default(0)/normal = !charend('/') !default(0)
>    /varorder = !charend('/') !default(2)/total = !charend('/')
> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
> !default(1)
>    /iterate = !charend('/') !default(10000)/converge = !charend('/')
> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
> !charend('/') !default(0)/
>      coeffci = !charend('/') !default(0)/covmy = !charend('/')
> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
> !default(random)).
>
> Any thoughts greatly welcomed.
> Thanks much,
> Â
> Andra
>

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (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]

DISCLAIMER:  I have never used any of the macros you are talking about here.  

Have you seen the example of how to call the INDIRECT macro on p. 890 of http://www.quantpsy.org/pubs/preacher_hayes_2008b.pdf?

Anter wrote

Yes, I did bother to read the PDF and the others from mediation and some articles... However programming isn't something I know. I don't really understand where to put the name of my variables, if not in the syntax shown below? If I write the names here
INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal = n/contrast = t/percent = p/bc = b/bca = d */.
nothing happens, no output appears. This is why I asked if there are errors because of the older version of spss or because I am writing it wrong.
Thank you.
Andra

--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:


From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 3:21 PM


2. Did you bother to read the instructions in the PDF file?
http://www.afhayes.com/public/process.pdf
DO NOT MODIFY THE CODE!

From the instructions you did not read:
"Make sure you run the code exactly as downloaded in a syntax or program
file.  Modifying the code will produce errors.  Do not modify the code at
all. "
---
--
Anter wrote
>
> Hello!
> Â
> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
> Â
> 1.The simple mediation analysis, each mediator with each predictor was
> significant. Now I want to test for the multiple mediation. How can I do
> this in SPSS? I understand that first I must verify that each IV
> positively predicts each mediator, controlling for the effect of the other
> IV's and the other mediator. If one of my IV's doesn't significantly
> predict a mediator do I use it in further analysis? Also controlling for
> would mean hierarchical regression, with IV's and mediator1 in step 1 andÂ
> the IV predicting mediator2 in step 2?
>
> 2. There is a macro on multiple mediation written by Andrew Hayes called
> Process
> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html) but it
> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
> if there's something I'm doing wrong or it's because of the older version
> of SPSS. My question is do I change any part of the code, where do I
> insert the name of my variables, after  !charend before it etc, do I use
> spaces, are there any more slashes or paranthesis. An error sounds
> something like :
> Â
> Error # 6820 in column 27.  Text: AVER_TI_Z
>>There is no left parenthesis after one of the DEFINE command parameter
>>specifications, e.g.  after !TOKENS.
>>This command not executed.
> Â
> Does anyone know what it means?
> Â
> This is the syntax:
> Â
> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
> = !charend('/')/m = !charend('/')/x = !charend ('/')
>    /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
> !charend('/') !default(95)
>    /hc3 = !charend('/') !default(0)/cluster = !charend('/')
> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
> !charend('/') !default(999)/
>    vmodval = !charend('/') !default (999)/qmodval = !charend('/')
> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
> !charend('/') !default (999)
>   /boot = !charend('/') !default(1000)/center = !charend('/')
> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
> !default(0)/normal = !charend('/') !default(0)
>    /varorder = !charend('/') !default(2)/total = !charend('/')
> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
> !default(1)
>    /iterate = !charend('/') !default(10000)/converge = !charend('/')
> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
> !charend('/') !default(0)/
>      coeffci = !charend('/') !default(0)/covmy = !charend('/')
> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
> !default(random)).
>
> Any thoughts greatly welcomed.
> Thanks much,
> Â
> Andra
>




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
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 Marso]

There are something like 14 macros on that site and I was assuming you are running the first one (PROCESS) but then you switch to the second one (INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a macro in the first place.
Follow the instructions and look at the several annotated examples provided by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some basic level of reading comprehension and a capacity to follow the instructions.  

Anter wrote

Yes, I did bother to read the PDF and the others from mediation and some articles... However programming isn't something I know. I don't really understand where to put the name of my variables, if not in the syntax shown below? If I write the names here
INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal = n/contrast = t/percent = p/bc = b/bca = d */.
nothing happens, no output appears. This is why I asked if there are errors because of the older version of spss or because I am writing it wrong.
Thank you.
Andra

--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:


From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 3:21 PM


2. Did you bother to read the instructions in the PDF file?
http://www.afhayes.com/public/process.pdf
DO NOT MODIFY THE CODE!

From the instructions you did not read:
"Make sure you run the code exactly as downloaded in a syntax or program
file.  Modifying the code will produce errors.  Do not modify the code at
all. "
---
--
Anter wrote
>
> Hello!
> Â
> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
> Â
> 1.The simple mediation analysis, each mediator with each predictor was
> significant. Now I want to test for the multiple mediation. How can I do
> this in SPSS? I understand that first I must verify that each IV
> positively predicts each mediator, controlling for the effect of the other
> IV's and the other mediator. If one of my IV's doesn't significantly
> predict a mediator do I use it in further analysis? Also controlling for
> would mean hierarchical regression, with IV's and mediator1 in step 1 andÂ
> the IV predicting mediator2 in step 2?
>
> 2. There is a macro on multiple mediation written by Andrew Hayes called
> Process
> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html) but it
> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
> if there's something I'm doing wrong or it's because of the older version
> of SPSS. My question is do I change any part of the code, where do I
> insert the name of my variables, after  !charend before it etc, do I use
> spaces, are there any more slashes or paranthesis. An error sounds
> something like :
> Â
> Error # 6820 in column 27.  Text: AVER_TI_Z
>>There is no left parenthesis after one of the DEFINE command parameter
>>specifications, e.g.  after !TOKENS.
>>This command not executed.
> Â
> Does anyone know what it means?
> Â
> This is the syntax:
> Â
> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
> = !charend('/')/m = !charend('/')/x = !charend ('/')
>    /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
> !charend('/') !default(95)
>    /hc3 = !charend('/') !default(0)/cluster = !charend('/')
> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
> !charend('/') !default(999)/
>    vmodval = !charend('/') !default (999)/qmodval = !charend('/')
> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
> !charend('/') !default (999)
>   /boot = !charend('/') !default(1000)/center = !charend('/')
> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
> !default(0)/normal = !charend('/') !default(0)
>    /varorder = !charend('/') !default(2)/total = !charend('/')
> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
> !default(1)
>    /iterate = !charend('/') !default(10000)/converge = !charend('/')
> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
> !charend('/') !default(0)/
>      coeffci = !charend('/') !default(0)/covmy = !charend('/')
> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
> !default(random)).
>
> Any thoughts greatly welcomed.
> Thanks much,
> Â
> Andra
>




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

[Anter]

For what I need to do they both, Indirect and Process, are equally usefull. I run the macro, after introducing the data according to the instructions, and nothing happens ... the output doesn't appear ... I think it is from spss, otherwise I can't explain it. I've run the Indirect macro from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm and it worked and know it just gives me <Unexpected error; quiting>. Don't think reinstalling spss would solve it.




--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 6:10 PM

There are something like 14 macros on that site and I was assuming you are
running the first one (PROCESS) but then you switch to the second one
(INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a
macro in the first place.
Follow the instructions and look at the several annotated examples provided
by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some
basic level of reading comprehension and a capacity to follow the
instructions.


Anter wrote

>
> Yes, I did bother to read the PDF and the others from mediation and some
> articles... However programming isn't something I know. I don't really
> understand where to put the name of my variables, if not in the syntax
> shown below? If I write the names here
> INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal
> = n/contrast = t/percent = p/bc = b/bca = d */.
> nothing happens, no output appears. This is why I asked if there are
> errors because of the older version of spss or because I am writing it
> wrong.
> Thank you.
> Andra
>
> --- On Tue, 7/24/12, David Marso &lt;david.marso@&gt; wrote:
>
>
> From: David Marso &lt;david.marso@&gt;
> Subject: Re: multiple mediation - regression & macro SPSS?
> To: SPSSX-L@.UGA
> Date: Tuesday, July 24, 2012, 3:21 PM
>
>
> 2. Did you bother to read the instructions in the PDF file?
> http://www.afhayes.com/public/process.pdf
> DO NOT MODIFY THE CODE!
>
> From the instructions you did not read:
> "Make sure you run the code exactly as downloaded in a syntax or program
> file.  Modifying the code will produce errors.  Do not modify the code at
> all. "
> ---
> --
> Anter wrote
>>
>> Hello!
>> Â
>> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
>> Â
>> 1.The simple mediation analysis, each mediator with each predictor was
>> significant. Now I want to test for the multiple mediation. How can I do
>> this in SPSS? I understand that first I must verify that each IV
>> positively predicts each mediator, controlling for the effect of the
>> other
>> IV's and the other mediator. If one of my IV's doesn't significantly
>> predict a mediator do I use it in further analysis? Also controlling for
>> would mean hierarchical regression, with IV's and mediator1 in step 1
>> andÂ
>> the IV predicting mediator2 in step 2?
>>
>> 2. There is a macro on multiple mediation written by Andrew Hayes called
>> Process
>> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html)Â but it
>> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
>> if there's something I'm doing wrong or it's because of the older version
>> of SPSS. My question is do I change any part of the code, where do I
>> insert the name of my variables, after  !charend before it etc, do I use
>> spaces, are there any more slashes or paranthesis. An error sounds
>> something like :
>> Â
>> Error # 6820 in column 27.  Text: AVER_TI_Z
>>>There is no left parenthesis after one of the DEFINE commandÂ

parameter

>>>specifications, e.g.  after !TOKENS.
>>>This command not executed.
>> Â
>> Does anyone know what it means?
>> Â
>> This is the syntax:
>> Â
>> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
>> = !charend('/')/m = !charend('/')/x = !charend ('/')
>>    /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
>> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
>> !charend('/') !default(95)
>>    /hc3 = !charend('/') !default(0)/cluster = !charend('/')
>> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
>> !charend('/') !default(999)/
>>    vmodval = !charend('/') !default (999)/qmodval = !charend('/')
>> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
>> !charend('/') !default (999)
>>   /boot = !charend('/') !default(1000)/center = !charend('/')
>> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
>> !default(0)/normal = !charend('/') !default(0)
>>    /varorder = !charend('/') !default(2)/total = !charend('/')
>> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
>> !default(1)
>>    /iterate = !charend('/') !default(10000)/converge = !charend('/')
>> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
>> !charend('/') !default(0)/
>>      coeffci = !charend('/') !default(0)/covmy = !charend('/')
>> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
>> !default(random)).
>>
>> Any thoughts greatly welcomed.
>> Thanks much,
>> Â
>> Andra
>>
>
>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@.UGA (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
>

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714416.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (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

[Jon K Peck]

I'm not familiar with these macros, but when you run them, watch the status bar to see what is running.  You should see the expanded macro commands flash by.  If not, then you haven't actually gotten the macro to run.

It is very unlikely that the SPSS version would matter.

If you can run an actual example from the doc using its example data, that should tell you whether you have it basically set up correctly.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Andra Th <[hidden email]>
To:        [hidden email]
Date:        07/24/2012 12:47 PM
Subject:        Re: [SPSSX-L] multiple mediation - regression & macro SPSS?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

For what I need to do they both, Indirect and Process, are equally usefull. I run the macro, after introducing the data according to the instructions, and nothing happens ... the output doesn't appear ... I think it is from spss, otherwise I can't explain it. I've run the Indirect macro from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm and it worked and know it just gives me <Unexpected error; quiting>. Don't think reinstalling spss would solve it.




--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 6:10 PM

There are something like 14 macros on that site and I was assuming you are
running the first one (PROCESS) but then you switch to the second one
(INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a
macro in the first place.
Follow the instructions and look at the several annotated examples provided
by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some
basic level of reading comprehension and a capacity to follow the
instructions.


Anter wrote
>
> Yes, I did bother to read the PDF and the others from mediation and some
> articles... However programming isn't something I know. I don't really
> understand where to put the name of my variables, if not in the syntax
> shown below? If I write the names here
> INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal
> = n/contrast = t/percent = p/bc = b/bca = d */.
> nothing happens, no output appears. This is why I asked if there are
> errors because of the older version of spss or because I am writing it
> wrong.
> Thank you.
> Andra
>
> --- On Tue, 7/24/12, David Marso &lt;david.marso@&gt; wrote:
>
>
> From: David Marso &lt;david.marso@&gt;
> Subject: Re: multiple mediation - regression & macro SPSS?
> To: SPSSX-L@.UGA
> Date: Tuesday, July 24, 2012, 3:21 PM
>
>
> 2. Did you bother to read the instructions in the PDF file?
> http://www.afhayes.com/public/process.pdf
> DO NOT MODIFY THE CODE!
>
> From the instructions you did not read:
> "Make sure you run the code exactly as downloaded in a syntax or program
> file.  Modifying the code will produce errors.  Do not modify the code at
> all. "
> ---
> --
> Anter wrote
>>
>> Hello!
>> Â
>> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
>> Â
>> 1.The simple mediation analysis, each mediator with each predictor was
>> significant. Now I want to test for the multiple mediation. How can I do
>> this in SPSS? I understand that first I must verify that each IV
>> positively predicts each mediator, controlling for the effect of the
>> other
>> IV's and the other mediator. If one of my IV's doesn't significantly
>> predict a mediator do I use it in further analysis? Also controlling for
>> would mean hierarchical regression, with IV's and mediator1 in step 1
>> andÂ
>> the IV predicting mediator2 in step 2?
>>
>> 2. There is a macro on multiple mediation written by Andrew Hayes called
>> Process
>> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html)Â but it
>> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
>> if there's something I'm doing wrong or it's because of the older version
>> of SPSS. My question is do I change any part of the code, where do I
>> insert the name of my variables, after  !charend before it etc, do I use
>> spaces, are there any more slashes or paranthesis. An error sounds
>> something like :
>> Â
>> Error # 6820 in column 27.  Text: AVER_TI_Z
>>>There is no left parenthesis after one of the DEFINE commandÂ
parameter
>>>specifications, e.g.  after !TOKENS.
>>>This command not executed.
>> Â
>> Does anyone know what it means?
>> Â
>> This is the syntax:
>> Â
>> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
>> = !charend('/')/m = !charend('/')/x = !charend ('/')
>> Â   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
>> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
>> !charend('/') !default(95)
>> Â   /hc3 = !charend('/') !default(0)/cluster = !charend('/')
>> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
>> !charend('/') !default(999)/
>> Â   vmodval = !charend('/') !default (999)/qmodval = !charend('/')
>> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
>> !charend('/') !default (999)
>>   /boot = !charend('/') !default(1000)/center = !charend('/')
>> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
>> !default(0)/normal = !charend('/') !default(0)
>> Â   /varorder = !charend('/') !default(2)/total = !charend('/')
>> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
>> !default(1)
>> Â   /iterate = !charend('/') !default(10000)/converge = !charend('/')
>> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
>> !charend('/') !default(0)/
>> Â Â Â   coeffci = !charend('/') !default(0)/covmy = !charend('/')
>> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
>> !default(random)).
>>
>> Any thoughts greatly welcomed.
>> Thanks much,
>> Â
>> Andra
>>
>
>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@.UGA (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
>




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714416.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (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

[Dale Glaser]

Anter, I did download the PROCESS macro last week, and it runs fine.....so as has been recommended, you will have PROCESS saved as a syntax file and run the entire (unedited file), with your datafile also open,  and then in a separate file (this example is for multiple mediators in parallel (as opposed to serial) mode) run the following (with your variable names, of course) linear model:


process vars=Soc_Infl Self_Conf Self_Eff Hlth_Mnt /y = Soc_Infl/x = Self_Conf/m= Self_Eff Hlth_Mnt
/total=1/model=4/boot=5000/effsize = 1/.

..and it should run fine....


dale


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

--- On Tue, 7/24/12, Jon K Peck <[hidden email]> wrote:

From: Jon K Peck <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 11:59 AM

I'm not familiar with these macros, but when you run them, watch the status bar to see what is running.  You should see the expanded macro commands flash by.  If not, then you haven't actually gotten the macro to run.

It is very unlikely that the SPSS version would matter.

If you can run an actual example from the doc using its example data, that should tell you whether you have it basically set up correctly.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Andra Th <[hidden email]>
To:        [hidden email]
Date:        07/24/2012 12:47 PM
Subject:        Re: [SPSSX-L] multiple mediation - regression & macro SPSS?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

For what I need to do they both, Indirect and Process, are equally usefull. I run the macro, after introducing the data according to the instructions, and nothing happens ... the output doesn't appear ... I think it is from spss, otherwise I can't explain it. I've run the Indirect macro from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm and it worked and know it just gives me <Unexpected error; quiting>. Don't think reinstalling spss would solve it.




--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 6:10 PM

There are something like 14 macros on that site and I was assuming you are
running the first one (PROCESS) but then you switch to the second one
(INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a
macro in the first place.
Follow the instructions and look at the several annotated examples provided
by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some
basic level of reading comprehension and a capacity to follow the
instructions.


Anter wrote
>
> Yes, I did bother to read the PDF and the others from mediation and some
> articles... However programming isn't something I know. I don't really
> understand where to put the name of my variables, if not in the syntax
> shown below? If I write the names here
> INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal
> = n/contrast = t/percent = p/bc = b/bca = d */.
> nothing happens, no output appears. This is why I asked if there are
> errors because of the older version of spss or because I am writing it
> wrong.
> Thank you.
> Andra
>
> --- On Tue, 7/24/12, David Marso &lt;david.marso@&gt; wrote:
>
>
> From: David Marso &lt;david.marso@&gt;
> Subject: Re: multiple mediation - regression & macro SPSS?
> To: SPSSX-L@.UGA
> Date: Tuesday, July 24, 2012, 3:21 PM
>
>
> 2. Did you bother to read the instructions in the PDF file?
> http://www.afhayes.com/public/process.pdf
> DO NOT MODIFY THE CODE!
>
> From the instructions you did not read:
> "Make sure you run the code exactly as downloaded in a syntax or program
> file.  Modifying the code will produce errors.  Do not modify the code at
> all. "
> ---
> --
> Anter wrote
>>
>> Hello!
>> Â
>> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
>> Â
>> 1.The simple mediation analysis, each mediator with each predictor was
>> significant. Now I want to test for the multiple mediation. How can I do
>> this in SPSS? I understand that first I must verify that each IV
>> positively predicts each mediator, controlling for the effect of the
>> other
>> IV's and the other mediator. If one of my IV's doesn't significantly
>> predict a mediator do I use it in further analysis? Also controlling for
>> would mean hierarchical regression, with IV's and mediator1 in step 1
>> andÂ
>> the IV predicting mediator2 in step 2?
>>
>> 2. There is a macro on multiple mediation written by Andrew Hayes called
>> Process
>> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html)Â but it
>> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
>> if there's something I'm doing wrong or it's because of the older version
>> of SPSS. My question is do I change any part of the code, where do I
>> insert the name of my variables, after  !charend before it etc, do I use
>> spaces, are there any more slashes or paranthesis. An error sounds
>> something like :
>> Â
>> Error # 6820 in column 27.  Text: AVER_TI_Z
>>>There is no left parenthesis after one of the DEFINE commandÂ
parameter
>>>specifications, e.g.  after !TOKENS.
>>>This command not executed.
>> Â
>> Does anyone know what it means?
>> Â
>> This is the syntax:
>> Â
>> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
>> = !charend('/')/m = !charend('/')/x = !charend ('/')
>> Â   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
>> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
>> !charend('/') !default(95)
>> Â   /hc3 = !charend('/') !default(0)/cluster = !charend('/')
>> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
>> !charend('/') !default(999)/
>> Â   vmodval = !charend('/') !default (999)/qmodval = !charend('/')
>> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
>> !charend('/') !default (999)
>>   /boot = !charend('/') !default(1000)/center = !charend('/')
>> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
>> !default(0)/normal = !charend('/') !default(0)
>> Â   /varorder = !charend('/') !default(2)/total = !charend('/')
>> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
>> !default(1)
>> Â   /iterate = !charend('/') !default(10000)/converge = !charend('/')
>> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
>> !charend('/') !default(0)/
>> Â Â Â   coeffci = !charend('/') !default(0)/covmy = !charend('/')
>> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
>> !default(random)).
>>
>> Any thoughts greatly welcomed.
>> Thanks much,
>> Â
>> Andra
>>
>
>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@.UGA (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
>




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714416.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (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 Marso]

"watch the status bar to see what is running. "
It will possibly only indicate Running MATRIX... Waiting for END MATRIX since this is one HUGE matrix macro.
"It is very unlikely that the SPSS version would matter."
At what release did SPSS support macro arguments with > 7 length?  Previous releases will truncate with who knows what side effects.
IIRC : OP is on 16.

Jon K Peck wrote

I'm not familiar with these macros, but when you run them, watch the
status bar to see what is running.  You should see the expanded macro
commands flash by.  If not, then you haven't actually gotten the macro to
run.

It is very unlikely that the SPSS version would matter.

If you can run an actual example from the doc using its example data, that
should tell you whether you have it basically set up correctly.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:   Andra Th <[hidden email]>
To:     [hidden email]
Date:   07/24/2012 12:47 PM
Subject:        Re: [SPSSX-L] multiple mediation - regression & macro
SPSS?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>




For what I need to do they both, Indirect and Process, are equally
usefull. I run the macro, after introducing the data according to the
instructions, and nothing happens ... the output doesn't appear ... I
think it is from spss, otherwise I can't explain it. I've run the Indirect
macro from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm and
it worked and know it just gives me <Unexpected error; quiting>. Don't
think reinstalling spss would solve it.




--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 6:10 PM

There are something like 14 macros on that site and I was assuming you are
running the first one (PROCESS) but then you switch to the second one
(INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a
macro in the first place.
Follow the instructions and look at the several annotated examples
provided
by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some
basic level of reading comprehension and a capacity to follow the
instructions.


Anter wrote
>
> Yes, I did bother to read the PDF and the others from mediation and
some
> articles... However programming isn't something I know. I don't really
> understand where to put the name of my variables, if not in the syntax
> shown below? If I write the names here
> INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf =
ci/normal
> = n/contrast = t/percent = p/bc = b/bca = d */.
> nothing happens, no output appears. This is why I asked if there are
> errors because of the older version of spss or because I am writing it
> wrong.
> Thank you.
> Andra
>
> --- On Tue, 7/24/12, David Marso <david.marso@> wrote:
>
>
> From: David Marso <david.marso@>
> Subject: Re: multiple mediation - regression & macro SPSS?
> To: SPSSX-L@.UGA
> Date: Tuesday, July 24, 2012, 3:21 PM
>
>
> 2. Did you bother to read the instructions in the PDF file?
> http://www.afhayes.com/public/process.pdf
> DO NOT MODIFY THE CODE!
>
> From the instructions you did not read:
> "Make sure you run the code exactly as downloaded in a syntax or program
> file.  Modifying the code will produce errors.  Do not modify the code
at
> all. "
> ---
> --
> Anter wrote
>>
>> Hello!
>> Â
>> I have a design with 2 mediators and 3 predictors, 1 outcome
variable.Â
>> Â
>> 1.The simple mediation analysis, each mediator with each predictor was
>> significant. Now I want to test for the multiple mediation. How can I
do
>> this in SPSS? I understand that first I must verify that each IV
>> positively predicts each mediator, controlling for the effect of the
>> other
>> IV's and the other mediator. If one of my IV's doesn't significantly
>> predict a mediator do I use it in further analysis? Also controlling
for
>> would mean hierarchical regression, with IV's and mediator1 in step 1
>> andÂ
>> the IV predicting mediator2 in step 2?
>>
>> 2. There is a macro on multiple mediation written by Andrew Hayes
called
>> Process
>> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html) but
it
>> gives me errors in SPSS 16. I'm not familiar with macros so I don't
know
>> if there's something I'm doing wrong or it's because of the older
version
>> of SPSS. My question is do I change any part of the code, where do I
>> insert the name of my variables, after  !charend before it etc, do I
use
>> spaces, are there any more slashes or paranthesis. An error sounds
>> something like :
>> Â
>> Error # 6820 in column 27.  Text: AVER_TI_Z
>>>There is no left parenthesis after one of theÂ
DEFINEÂ commandÂ
parameter
>>>specifications, e.g.  after !TOKENS.
>>>This command not executed.
>> Â
>> Does anyone know what it means?
>> Â
>> This is the syntax:
>> Â
>> define process (vars = !charend('/')/model = !charend('/')
!default(75)/y
>> = !charend('/')/m = !charend('/')/x = !charend ('/')
>> Â   /w = !charend('/') !default(xxx)/z = !charend('/')
!default(xxx)/v =
>> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
>> !charend('/') !default(95)
>> Â   /hc3 = !charend('/') !default(0)/cluster = !charend('/')
>> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
>> !charend('/') !default(999)/
>> Â   vmodval = !charend('/') !default (999)/qmodval = !charend('/')
>> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
>> !charend('/') !default (999)
>>   /boot = !charend('/') !default(1000)/center = !charend('/')
>> !default(0)/quantile = !charend('/') !default(0)/effsize =
!charend('/')
>> !default(0)/normal = !charend('/') !default(0)
>> Â   /varorder = !charend('/') !default(2)/total = !charend('/')
>> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
>> !default(1)
>> Â   /iterate = !charend('/') !default(10000)/converge =
!charend('/')
>> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
>> !charend('/') !default(0)/
>> Â Â Â   coeffci = !charend('/') !default(0)/covmy = !charend('/')
>> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
>> !default(random)).
>>
>> Any thoughts greatly welcomed.
>> Thanks much,
>> Â
>> Andra
>>
>
>
>
>
> --
> View this message in context:
>
http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html

> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@.UGA (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
>




--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714416.html

Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

[Anter]

Thank you, indeed now I managed to run the syntax and have the results.

The Indirect script that I have from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm seems to work in spss 17, but I didn't manage to get it working in spss 16. But, now it's fine, I can use the macros.

 

Don't know if this is the place to ask, but what do you think about the causal steps that must be met for mediation according to Baron & Kenny? Some authors, e.g. http://leeds.colorado.edu/asset/publication/zhaolynchchen2010jcr.pdf consider that not all steps need to be significant so we can speak of some form of mediation.

 

Which would be more "secure" in a report, analysing mediation results even if some steps (c not significant) aren't met or only if all the steps are significant? Also, can I test the mediation only for the predictor I'm interested in and report those results or is it necessary to control for all the other predictors in the model as well?

 

--- On Tue, 7/24/12, Dale Glaser <[hidden email]> wrote:

From: Dale Glaser <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 11:30 PM

Anter, I did download the PROCESS macro last week, and it runs fine.....so as has been recommended, you will have PROCESS saved as a syntax file and run the entire (unedited file), with your datafile also open,  and then in a separate file (this example is for multiple mediators in parallel (as opposed to serial) mode) run the following (with your variable names, of course) linear model:


process vars=Soc_Infl Self_Conf Self_Eff Hlth_Mnt /y = Soc_Infl/x = Self_Conf/m= Self_Eff Hlth_Mnt
/total=1/model=4/boot=5000/effsize = 1/.

..and it should run fine....


dale


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

--- On Tue, 7/24/12, Jon K Peck <[hidden email]> wrote:

From: Jon K Peck <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 11:59 AM

I'm not familiar with these macros, but when you run them, watch the status bar to see what is running.  You should see the expanded macro commands flash by.  If not, then you haven't actually gotten the macro to run.

It is very unlikely that the SPSS version would matter.

If you can run an actual example from the doc using its example data, that should tell you whether you have it basically set up correctly.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Andra Th <[hidden email]>
To:        [hidden email]
Date:        07/24/2012 12:47 PM
Subject:        Re: [SPSSX-L] multiple mediation - regression & macro SPSS?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

For what I need to do they both, Indirect and Process, are equally usefull. I run the macro, after introducing the data according to the instructions, and nothing happens ... the output doesn't appear ... I think it is from spss, otherwise I can't explain it. I've run the Indirect macro from http://www.ats.ucla.edu/stat/spss/faq/mediation_complex.htm and it worked and know it just gives me <Unexpected error; quiting>. Don't think reinstalling spss would solve it.




--- On Tue, 7/24/12, David Marso <[hidden email]> wrote:

From: David Marso <[hidden email]>
Subject: Re: multiple mediation - regression & macro SPSS?
To: [hidden email]
Date: Tuesday, July 24, 2012, 6:10 PM

There are something like 14 macros on that site and I was assuming you are
running the first one (PROCESS) but then you switch to the second one
(INDIRECT)!
DO NOT MODIFY THE MACRO!  That completely defeats the purpose of having a
macro in the first place.
Follow the instructions and look at the several annotated examples provided
by Prof Hayes.
This does not require knowledge of programming.  It merely assumes some
basic level of reading comprehension and a capacity to follow the
instructions.


Anter wrote

>
> Yes, I did bother to read the PDF and the others from mediation and some
> articles... However programming isn't something I know. I don't really
> understand where to put the name of my variables, if not in the syntax
> shown below? If I write the names here
> INDIRECT y = dv/x = iv/m = mlist covlist/c = cov/boot = z/conf = ci/normal
> = n/contrast = t/percent = p/bc = b/bca = d */.
> nothing happens, no output appears. This is why I asked if there are
> errors because of the older version of spss or because I am writing it
> wrong.
> Thank you.
> Andra
>
> --- On Tue, 7/24/12, David Marso &lt;david.marso@&gt; wrote:
>
>
> From: David Marso &lt;david.marso@&gt;
> Subject: Re: multiple mediation - regression & macro SPSS?
> To: SPSSX-L@.UGA
> Date: Tuesday, July 24, 2012, 3:21 PM
>
>
> 2. Did you bother to read the instructions in the PDF file?
> http://www.afhayes.com/public/process.pdf
> DO NOT MODIFY THE CODE!
>
> From the instructions you did not read:
> "Make sure you run the code exactly as downloaded in a syntax or program
> file.  Modifying the code will produce errors.  Do not modify the code at
> all. "
> ---
> --
> Anter wrote
>>
>> Hello!
>> Â
>> I have a design with 2 mediators and 3 predictors, 1 outcome variable.Â
>> Â
>> 1.The simple mediation analysis, each mediator with each predictor was
>> significant. Now I want to test for the multiple mediation. How can I do
>> this in SPSS? I understand that first I must verify that each IV
>> positively predicts each mediator, controlling for the effect of the
>> other
>> IV's and the other mediator. If one of my IV's doesn't significantly
>> predict a mediator do I use it in further analysis? Also controlling for
>> would mean hierarchical regression, with IV's and mediator1 in step 1
>> andÂ
>> the IV predicting mediator2 in step 2?
>>
>> 2. There is a macro on multiple mediation written by Andrew Hayes called
>> Process
>> (http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html)Â but it
>> gives me errors in SPSS 16. I'm not familiar with macros so I don't know
>> if there's something I'm doing wrong or it's because of the older version
>> of SPSS. My question is do I change any part of the code, where do I
>> insert the name of my variables, after  !charend before it etc, do I use
>> spaces, are there any more slashes or paranthesis. An error sounds
>> something like :
>> Â
>> Error # 6820 in column 27.  Text: AVER_TI_Z
>>>There is no left parenthesis after one of the DEFINE commandÂ

parameter

>>>specifications, e.g.  after !TOKENS.
>>>This command not executed.
>> Â
>> Does anyone know what it means?
>> Â
>> This is the syntax:
>> Â
>> define process (vars = !charend('/')/model = !charend('/') !default(75)/y
>> = !charend('/')/m = !charend('/')/x = !charend ('/')
>> Â   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v =
>> !charend('/') !default(xxx)/q = !charend ('/') !default(xxx)/conf =
>> !charend('/') !default(95)
>> Â   /hc3 = !charend('/') !default(0)/cluster = !charend('/')
>> !default(xxx)/wmodval = !charend('/') !default (999)/zmodval =
>> !charend('/') !default(999)/
>> Â   vmodval = !charend('/') !default (999)/qmodval = !charend('/')
>> !default(999)/mmodval = !charend('/') !default (999)/xmodval =
>> !charend('/') !default (999)
>>   /boot = !charend('/') !default(1000)/center = !charend('/')
>> !default(0)/quantile = !charend('/') !default(0)/effsize = !charend('/')
>> !default(0)/normal = !charend('/') !default(0)
>> Â   /varorder = !charend('/') !default(2)/total = !charend('/')
>> !default(0)/plot = !charend('/') !default(0)/detail = !charend('/')
>> !default(1)
>> Â   /iterate = !charend('/') !default(10000)/converge = !charend('/')
>> !default(0.00000001)/percent = !charend('/') !default(0)/jn =
>> !charend('/') !default(0)/
>> Â Â Â   coeffci = !charend('/') !default(0)/covmy = !charend('/')
>> !default(0)/contrast = !charend('/') !default(0)/seed = !charend('/')
>> !default(random)).
>>
>> Any thoughts greatly welcomed.
>> Thanks much,
>> Â
>> Andra
>>
>
>
>
>
> --
> View this message in context:
> http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714408.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> LISTSERV@.UGA (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
>

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/multiple-mediation-regression-macro-SPSS-tp5714407p5714416.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@... (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