Page 1 of 1
Can i modify the results of my moderating effect
Posted: Mon Jan 22, 2018 11:14 am
by amira-mehdi
- m.PNG (20.24 KiB) Viewed 2588 times
- moderating.PNG (30.46 KiB) Viewed 2588 times
Hello, in my model i have the social media use as a moderator variable that amplifies the positive impact of social CRM on firm performance. My proble is that moderating effect when testing it with the smartpls3 was not significant (p-value= 0.549) + the direct relation between Social CRM and firm performance was also not significant (p-value= 0.664)which is total illegical !
My question is how can i do to make my results better and to prove that the moderator+ the direct impact both have a significant impact ??
Re: Can i modify the results of my moderating effect
Posted: Tue Jan 23, 2018 11:30 am
by cringle
We strongly suggest not to "torture" methods until it confesses what you expect.
In case of your situation, the coefficients are so small that no variations would lead to a different solution. The problem may be with your data. Please check them again (or collect more data). But if the data is good, take the results as they are: There may be no relevant moderating effect!
Sorry, there is (as always) not a straightforward solution when estimations do not meet expectation (what is the usual case in statistical analyses).
Best regards
Christian Ringle