Moderator and Mediated Variable.

Research topics can be discussed in this area.
Post Reply
iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Moderator and Mediated Variable.

Post by iputu.ariyasa » Tue May 23, 2017 1:52 pm

I am new to PLS modeling and i'am in the middle of trying to analyze a model based on that proposed by Kibler (2013). in Kibler's model, one variable (regional environment) acts as a moderator variable and also as an exogenous mediated variable. Now suppose that the model will be analyzed by SmartPLS and the regional environment variable is a formative variable, thus two stage approach on testing the moderator effect is needed, should we create two separated PLS Path Model, i.e. one testing for mediated effect, and one testing for moderating effect, or not? Thank you.

Kibler, E. (2013). Formation of entrepreneurial intentions in a regional context. Entrepreneurship & Regional Development, 25(3-4), 293-323. doi:10.1080/08985626.2012.721008


-Best Regard

I Putu Ariyasa
Attachments
Screen Shot 2017-05-23 at 8.43.24 PM.png
Screen Shot 2017-05-23 at 8.43.24 PM.png (108.89 KiB) Viewed 4148 times

User avatar
cringle
SmartPLS Developer
Posts: 796
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: Moderator and Mediated Variable.

Post by cringle » Tue May 23, 2017 4:00 pm

You may test the model without the moderator first. In a second analysis step, you may analyze the results of the model including the moderator variable.

Kind regards
CR

iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Re: Moderator and Mediated Variable.

Post by iputu.ariyasa » Wed May 24, 2017 3:30 am

I see,

Thank you very much Sir

Best regards,

I Putu Ariyasa

iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Re: Moderator and Mediated Variable.

Post by iputu.ariyasa » Thu May 25, 2017 7:47 am

cringle wrote:You may test the model without the moderator first. In a second analysis step, you may analyze the results of the model including the moderator variable.

Kind regards
CR
Good day sir, i am sorry for bringing this up again. After some re-reading, i begin to wonder, in a second step analysis of moderating effect, do we still need to include the paths from regional environment to attitude, SN, and PBC? i am wondering about this because i am thinking of a scenario where the mediation effect by Attitude, SN and PBC is not found. thanks.

here i attach Kibler's proposition regarding the model

Proposition 1: The impact of regional conditions on entrepreneurial intentions is mediated by attitude, subjective norm and perceived behavioural control.
Proposition 2: Regional conditions moderate the impact of attitude, subjective norm and perceived behavioural control on entrepreneurial intentions.


Best regards,

I Putu Ariyasa

User avatar
cringle
SmartPLS Developer
Posts: 796
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: Moderator and Mediated Variable.

Post by cringle » Thu May 25, 2017 9:35 am

It looks like need a path from regional environment to entrepreneurial intention as well to test the moderation effect of regional environment on the paths from attitude, SN and PBC on entrepreneurial intention.

What is not clear: Why do you include the paths from regional environment to attitude, SN and PBC? That looks strange. You need to decide if regional environment acts as the source of explanation of other constructs or as a variable to explain group differences (as a moderator). Assigning both roles to that construct seems kind of difficult to justify.

If you like to use regional environment for group effects (or moderation) I would delete the path from regional environment to attitude, SN and PBC and would add a path from regional environment to entrepreneurial intention. Then, you can use the SmartPLS moderator option to create the interaction terms. Alternatively, you can use regional environment as a grouping variable for a PLS multigroup anaylsis (in that case you do not use regional environment as construct in the model).

Best
Christian

iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Re: Moderator and Mediated Variable.

Post by iputu.ariyasa » Thu May 25, 2017 1:58 pm

cringle wrote: You need to decide if regional environment acts as the source of explanation of other constructs or as a variable to explain group differences (as a moderator). Assigning both roles to that construct seems kind of difficult to justify.
Thank you sir for your reply.

Actually, it was based on two research. Ajzen (1991) suggest that external factors, such as demographic or environmental characteristics, do not influence intentions directly, but they are mediated by the three antecedents of intentions, while Cooke and Sheeran (2004) supports the utility of including external moderator variables to enhance the explanatory power of the intentions model. So the research was to test for both possible mediating and moderating effects of the regional environment on entrepreneurial intent.

If this looks weird, should i create two distinct model, one for mediated effect and one for moderation? or should i create two distinct model only in the pls path model analysis, while the model in hypothesis stays the same?

Best regards,

I Putu Ariyasa

User avatar
cringle
SmartPLS Developer
Posts: 796
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: Moderator and Mediated Variable.

Post by cringle » Thu May 25, 2017 2:15 pm

Theory and concept always comes first. If the model make sense to you, then keep it as it is. But to correctly analyze the moderating effect, you need to include a path from regional environment to entrepreneurial intention.

Best
CR

iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Re: Moderator and Mediated Variable.

Post by iputu.ariyasa » Thu May 25, 2017 2:47 pm

cringle wrote:Theory and concept always comes first. If the model make sense to you, then keep it as it is. But to correctly analyze the moderating effect, you need to include a path from regional environment to entrepreneurial intention.

Best
CR
alright sir, i will include the direct path from regional environment to entrepreneurial intention in the moderation analysis. But what not clear is should i exclude the paths from regional environment to attitude, SN, and PBC in the moderation analysis stage?


Best regards,

I Putu Ariyasa

User avatar
cringle
SmartPLS Developer
Posts: 796
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: Moderator and Mediated Variable.

Post by cringle » Thu May 25, 2017 3:10 pm

I would exclude those paths.

Best
CR

iputu.ariyasa
PLS Senior User
Posts: 22
Joined: Mon May 15, 2017 2:10 am
Real name and title: I Putu Ariyasa, ST.

Re: Moderator and Mediated Variable.

Post by iputu.ariyasa » Fri May 26, 2017 2:01 am

cringle wrote:I would exclude those paths.

Best
CR
Thank you very much sir for you reply

Best regards,

I Putu Ariyasa

Post Reply