forum.smartpls.com

Hello,

I'just switched from AMOS to (Smart)PLS and have some questions about the selection of estimation methods/algorithms.

I suppose the PLS alorithm is used to compute the model and interpret the beta's and loadings. Subsequently, to verify the significance of the paths Bootstrapping can be used (in order to compute t-values) (?). My questions are:

1) Is it o.k. to interpret the results from both estimation methods (PLS algorithm and bootstrapping) to arrive at overall conclusions for a structural model?

2) What are the differences between the PLS algorithm and Bootstrapping. Is there any literature on this subject available?

3) When should blindfolding be used?

With kind regards
dr. Tibert Verhagen
VU Amsterdam

PLS algorithm is used only to compute the construct correlations, ave values and alpha values. It will not test the significance of any path as it is non parametric in nature. To test path significances, you need resampling methods like bootstrap and jack knife provided by the software.

You need to use both the procedures to completely judge the model effectiveness.

I hope this helps

Vivek
India

Dear Vivek,

Thank you very much!

kind regards
Tibert

Hi,

In regard to what Mr Vivek said.

I understand the what each is used for (PLS Algorithm and Bootstrapping)

But what exactly is PLS Algorithm?

I realized I know we use it for mostly the reflective measurement model, then we get our rsquare for our structural model from the PLS algorithm.

And Bootstrapping is for significance testing.

But if I were to have the explain the difference. I can't.

I essentially don't know what PLS Algorithm is. I'd be grateful for some layman explanation on PLS Algorithm.

Many Thanks,
Priyanka

Maybe this video helps: https://www.youtube.com/watch?v=wGjIFC1HlHk
(shameless self promotion ;)