My results from PLS-SEM analysis (in WarpPLS) shows weak effect size (f-squared). The path coefficient is 0.15 and p-value is significant. I am analyzing indirect relationship (mediation). How do I interpret such results? Also the VAF value shows full mediation. I am confused.
Significant but weak means exactly what it sounds like: it's a small effect but it's statistically significant. This is a totally normal and common situation.
Whether mediation is partial vs. full is also a separate issue from those two. Full mediation means that the relationship between X and Y is completely explained by the mediator, whereas partial mediation means there's still some effect on X and Y even after the mediator has been added.
So it sounds like your results mean exactly what you said: M has a small but significant effect on Y, and it totally accounts for whatever apparent effect X looked like it had on Y.
As Stephen Politzer-Ahles said, yes it is a small effect but it's statistically significant.
I don't really understand though, how can be a relationship significant, but weak in the same time. It sounds like a contradiction to me.
Statistical significance is not the same thing as effect size.
Think about the standard t-test formula: it's effect size divided by variance. You can get a large t value by having a big effect, or by having small variance. In other words, an effect that's small but stable might be statistically significant, whereas an effect that's big but highly variable might not be significant. If you are thinking there is something contradictory about an effect being significant but small, then it sounds like you are misunderstanding what statistical significance actually means; I highly recommend you refresh your memory on statistical significance (for instance, the recent ASA statement on p-values is a good resource about this).
See also the figure I provided at https://www.researchgate.net/post/How_to_interpret_a_large_effect_sizeomega_squared_on_an_non-significant_factor
- Badges
- Science topic