Unexplicably high loading and R²

[antonin]

Hi all,

I tried to model mediating effect on smartpls. I found an unexpectedly high R² (>0,9) and loadings for a latent variable. I checked the database but the items are not equal at all. Where can this output come from? Is it normal to have a latent variable with 4 loading equal to 1?




By advance thank you,

[Hengkov]

Hi,

You have high collinearity problem.
For solve this problem, choice one or more options below:
1. Add sample size
2. Used PLS Regression
3. Check data outlier
4. Used latent score
5. Check your specificy model using Ramsay test
6. Delete your correlated variable

Best Wishes,