forum.smartpls.com

everybady,
I have a question about FIMIX path coefficients,

I think it should be <1,why my output the path coeffucients >1?

I have read some paper,they do not use 2nd LV.

they use all first LV link to the endogenous LV.

is these correct?

I also think ,maybe it is right .why we need 2nd LV?

IF the correlationship of first LV is low.we can use it directly.

or it depens on my theory?

anyone give some comment?

Dear all, I'm new to PLS, in this case what shall I do? I'm trying to prove that 2nd order formative measurement model of perceived value is better than reflective measurement model. That's: reflective measurement model: indicator <- 1st order <- 2nd order formative measurement model: indicator <- ...

Bido thanks a lot ！ Hi Mengchuanjin, ccc moderates the relation between aaa and ddd, and has no effect in the relation between bbb and ddd. Then you will have as exogenous variables: aaa, bbb, ccc, aaa*ccc (interaction term or multiplicative term) and ddd as endogenous variable. Best regards. Bido

my model

aaa--->ddd
bbb--->ddd

if i have a moderate ccc,

when i create a new construct representing the interaction effects of aaa and ccc,

whether i should delete the path bbb--->ddd or not?

if the moderate ccc have no effect on bbb.

thanks

[/url]

both correct

i am new ,i don't understand how to choice?

IN this option " weighting scheme"

what's the difference among "factor" "centroid" "path"


and in this option "metric"

what's the difference between "original" "mean o var 1"


thanks

I am new for PLS,if anyone can give a detail sample about how to read the result ?