what to do with unstable path coefficients and T-statistics?
what to do with unstable path coefficients and T-statistics?
I am running PLS for one of my paper with smartpls 2.0 but note that the T statistics and path coefficients are not the same every time I run the test. This troubles me especially when the T-statistics for some paths are near the critic values. The changes of T-statistics can have my findings turn from significant to insignificant. What happened with PLS? and what can I do with the results? Do other colleagues have the same issues?
-
Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Hi,
Because SmartPLS used the Bootstrapping approach (random selected), T-statistics and Path Coefficient will change every you run same times. For the solution, you can try used "individual change" and put 2000 resampling.
Another option, WarpPLS 4.0 have feature "Stable Resampling", so the result will same every times.
Best Regards,
Because SmartPLS used the Bootstrapping approach (random selected), T-statistics and Path Coefficient will change every you run same times. For the solution, you can try used "individual change" and put 2000 resampling.
Another option, WarpPLS 4.0 have feature "Stable Resampling", so the result will same every times.
Best Regards,