mediating effect problem
Posted: Wed Oct 31, 2012 2:35 am
Dear PLS experts
Hope my message find you all well...really in need for advice
I know there has been some post on this issue but could not gain a good idea on what should be done.
I am interested in testing mediating effect of one variable, and
since Baron and Kenny approach has been attacked I wanted to test it using Preacher and Hayes approach. therefore I run there macro on spss using 5000 bootstrap.
the problem is that:
1) their direct and indirect reporting is different from smartpls results (within 5%) so can I ask you why I am getting different results? and
2) in in the first case is it ok to use both smartpls and macro together, knowing that the macros work using OLS??
3) finally if i decided to use smartpls results only what is the approach to use in deciding if these indirect relationships are partial or complete?
thank you for any help i can get.
Hope my message find you all well...really in need for advice
I know there has been some post on this issue but could not gain a good idea on what should be done.
I am interested in testing mediating effect of one variable, and
since Baron and Kenny approach has been attacked I wanted to test it using Preacher and Hayes approach. therefore I run there macro on spss using 5000 bootstrap.
the problem is that:
1) their direct and indirect reporting is different from smartpls results (within 5%) so can I ask you why I am getting different results? and
2) in in the first case is it ok to use both smartpls and macro together, knowing that the macros work using OLS??
3) finally if i decided to use smartpls results only what is the approach to use in deciding if these indirect relationships are partial or complete?
thank you for any help i can get.