PLS-MGA P-Value

[LilianPLS]

Hello!
I am currently doing PLS-MGA. At the SmartPLS Website I found the following quote: "A result is significant at the 5% probability of error level, if the p-value is smaller than 0.05 or larger than 0.95 for a certain difference of group-specific path coefficients".
In a previously published article I noticed that the corresponding authors only tested p-values smaller than 0.05, so I would appreciate if you could help me with an explanation of the theoretical basis.

Thank you very much in advance.

[janschreier]

Hi there,

that is the difference between one- and two-sided tests:
https://en.wikipedia.org/wiki/One-_and_two-tailed_tests

If you have the assumption that group A has a higher value than group B, you make a one sided test and look only for values smaller than 0.05. If you only have the assumption that group A and B are different, you also check for values > 0.95.

Hope that helps,
Jan

[jmbecker]

If you want to test a two-sided test you should, however, use 2.5% and 97.5% as the cut-off values for a 5% alpha level. Otherwise, you are testing on a 10% alpha level.

[borakali]

with reference to LilianPLS who wrote "A result is significant at the 5% probability of error level if the p-value is smaller than 0.05 or larger than 0.95 for a certain difference of group-specific path coefficients". I have similar results in my research.
Would anyone send me a reference publication which supports the above significance level?
Thanks,
Borak