hi all
In terms of the Moderating Hypothesis, i have X as independent construct and Y as dependent construct. Z is the Moderating construct.
before insert z as a moderating construct, i run smart pls, result is:
x--->y path co is 0.49
R square for y is 0.24
and when run bootstrap with 500 re-s t result is:
x--->y t is 8.05
in second step i insert z as a moderating construct and run smartpls, result is:
x--->y path co is 0.45
z--->y path co is 0.086
x*z---->y path co is -0.18
R square for y is 0.28
and
x--->y t is 5.20
z--->y t is 0.0819
x*z---->y t is 0.78
i my hypothesis i assume that the relationship between x,y will be stronger with moderation effect of z!
1- is this hypothesis supported?
2- how much R square must be changed to a hypothesis do support?
3- critical value for supp/reject a hypothesis with moderating effect?
4- if before we insert a moderated variable, t x--->y was less than 1.96 and after a insert moderated variable higher than 1.96, what we must do?
thx
moderating effect
-
christian.nitzl
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Dear Resoul!
ad 1and 3) Your interaction term isn't significant because the t-value of x*y is <1,65. Therefore your moderating effect isn't significant and your hypothesis cannot supported.
as 2) Normally we use the Effect Size for interpretation of changes in R2.
ad 4) I don´t understand this question, sorry.
Best regards
Christian
ad 1and 3) Your interaction term isn't significant because the t-value of x*y is <1,65. Therefore your moderating effect isn't significant and your hypothesis cannot supported.
as 2) Normally we use the Effect Size for interpretation of changes in R2.
ad 4) I don´t understand this question, sorry.
Best regards
Christian
-
christian.nitzl
- PLS Expert User
- Posts: 248
- Joined: Sat Jul 25, 2009 1:34 pm
- Real name and title:
Your welcome!
ad 1) Generally, only the interaction term (x*y) is used for interpretation (e.g. Henseler/Chin (2010), p. 85-86.)
ad 2) Some authors argue, that the interaction coefficient shouldn’t the only basis for assessing the interaction effect, because the interaction coefficient could be biased. As complement they suggest the effect size. (e.g. Henseler/Chin (2010), p. 105)
If you can read German, you can find further information on this topic in my working paper on pages 47-49 (http://www.ibl-unihh.de/ap21.pdf).
But here the cited English literature:
Henseler, J./Chin, W. (2010): A Comparison of Approaches for the Analysis of Interaction Effects Between Latent Variables Using Partial Least Squares Path Modeling, in: Structural Equation Modeling: A Multidisciplinary Journal, Vol. 17, No. 1, pp. 82-109.
An intresting comment in these board from Sasa Saric you can find here:
viewtopic.php?t=1316&highlight=
Best regards
Christian
ad 1) Generally, only the interaction term (x*y) is used for interpretation (e.g. Henseler/Chin (2010), p. 85-86.)
ad 2) Some authors argue, that the interaction coefficient shouldn’t the only basis for assessing the interaction effect, because the interaction coefficient could be biased. As complement they suggest the effect size. (e.g. Henseler/Chin (2010), p. 105)
If you can read German, you can find further information on this topic in my working paper on pages 47-49 (http://www.ibl-unihh.de/ap21.pdf).
But here the cited English literature:
Henseler, J./Chin, W. (2010): A Comparison of Approaches for the Analysis of Interaction Effects Between Latent Variables Using Partial Least Squares Path Modeling, in: Structural Equation Modeling: A Multidisciplinary Journal, Vol. 17, No. 1, pp. 82-109.
An intresting comment in these board from Sasa Saric you can find here:
viewtopic.php?t=1316&highlight=
Best regards
Christian
-
jmbecker
- SmartPLS Developer
- Posts: 1301
- Joined: Tue Mar 28, 2006 11:09 am
- Real name and title: Dr. Jan-Michael Becker
Two articles that do not specifically address latent constructs, but moderation in general and are therefore also very helpful are:
Sharma, S., Durand, R.M., and Gur-Arie, O. "Identification and Analysis of Moderator Variables," Journal of Marketing Research (JMR) (18:3) 1981, pp 291-300.
Carte, T.A., and Russell, C.J. "IN PURSUIT OF MODERATION: NINE COMMON ERRORS AND THEIR SOLUTIONS," MIS Quarterly (27:3) 2003, pp 479-501.
Sharma, S., Durand, R.M., and Gur-Arie, O. "Identification and Analysis of Moderator Variables," Journal of Marketing Research (JMR) (18:3) 1981, pp 291-300.
Carte, T.A., and Russell, C.J. "IN PURSUIT OF MODERATION: NINE COMMON ERRORS AND THEIR SOLUTIONS," MIS Quarterly (27:3) 2003, pp 479-501.