forum.smartpls.com

Hi, I am not sure, but it looks like that multilevel analysis (dependent variable in the lower level (like individuals) and predictors in the same level or upper) was more suitable. Bickel, R. (2007). Multilevel Analysis for Applied Research : It’s just regression! (p. 355). New York: The Guilford P...

Hi, The bootstrap is based on random resampling with replacement the original sample, for this reason, every time that you run the bootstrap the resamples will be different from the previous, and paths (mean) and t-values will be different too. Some hints: 1) Do not use the option “No sign changes”,...

Hi,

Everything is ok, it is just a preview of your dataset.

If you want see all data used, run the PLS algorithm + Report + Manifest variables scores (original)

And a final hint, in the next doubts, look for the answer in the “Search” function (right corner)
Best regards,
Bido

Hi,
It is interpreted in the same way of standardized regression coefficients (betas).

Y <-- - 0.30 * X

In this example, when X increase 1 standard deviation, Y will decrease 0.30 s.d..

Best regards,
Bido

Hi, In the 1st step you don’t include the 2nd order LV, just the 15 first order LVs. You could analyze the measurement model: AVE, crossloadings, correlations between LVs, but don’t use the path coefficients for nothing. It is a good way to show that the first order LV are correlated, before put the...

Hi Omid, In your case (all LV are reflective) you have two options: 1) Just repeat the indicator of the first order LV in the second order LV (see Wetzels, p.181). Wetzels, M., Odekerken-Schröder, G., & Van Oppen, C. (2009). Using PLS path modeling for assessing hierarchical construct models: gu...

Hi,
in the same way, use the "outer weights".

SmartPLS already computes the LV scores, see::
Report / PLS / Calculation results / Latent variable scores /

Best regards,

Bido

Hi,

AVE and CC for the 2nd order LV, ou will have to compute by hand.

The "factor loadings" are in the report, but you should copy them from the path coefficients report.

Best regards,
Bido

Hi, It is simplier: - the relations between 2nd order LV and their 1st order LV are "measurement", for this reason, these relations should be interpreted as "outer loadings" or "factor loadings". - but the report of SmartPLS (and others softwares) will include these rel...

Hi,

- run SmartPLS (PLS algorithm)
- Report
- Copy the "standardized LV scores"
- Paste them in a spreadsheet
- Run the correlation analysis (SPSS or R is better than Excel to compute p-values)

Best regards,

Bido

Hi,

(*.sav) --> save as --> (*.csv)

Open SmartPLS
File / New / Create new project / Name / Next / Choose the (*.csv) file / ...

For more details, see the video (SmartPLS User Manual ) at downloads.php

Best regards,

Bido

Hi Kamal,
The "search" that I was talking about is in the upper right corner in this forum.

The results of bootstrap are t-values
If t > 1,96, the path or loading is significant at 5%.

Best regards,
Bido

Hi,
run the bootstrap procedure.

Use the search function to find details about it.

Best regards,
Bido

Hi,
maybe you could import the *.splsm file from SmartPLS to R,
and to run the things that you want.
See:http://127.0.0.1:18671/library/semPLS/h ... splsm.html

Best regards,
Bido

Hi,
it is a matter of multicollinearity.

See viewtopic.php?t=2477

Best regards,

Bido