Search found 54 matches

by stefanbehrens
Mon Jul 09, 2007 2:27 pm
Forum: Method and application
Topic: Multi-group comparisons with PLS
Replies: 59
Views: 66430

The version from Chin's website (no. 1 in my earlier post) is definitely the correct one.

Keil et al's formula must be a mistake. (By the way, their own calculated values in the paper cannot be reproduced if you fill in the numbers in their formula... so much for A-rated journals ;)

Cheers,
Stefan
by stefanbehrens
Wed Nov 22, 2006 12:43 am
Forum: Method and application
Topic: comparison of the fit of two models
Replies: 4
Views: 5488

Dear Sandra, if I remember correctly, the goodness-of-fit index proposed by Tenenhaus goes something like this: GOF = sqrt(average(AVE)*average(Rsq)) where average(AVE) is the mean of all AVE values of all LVs and average(Rsq) is the mean of all Rsquare values of all endogeneous LVs. Thus, the GOF s...
by stefanbehrens
Fri Jul 21, 2006 1:23 pm
Forum: Method and application
Topic: Significances / Covariation / Weights /Quality Indicators
Replies: 13
Views: 12843

Hi all, 1) regarding the one- or two-sided significance tests I would apply the same logic as in any test of the significance of correlation coefficients. If you already have an expectation regarding the sign of your coefficient (based on theoretical considerations), then use a one-sided test. If yo...
by stefanbehrens
Fri Jul 21, 2006 1:05 pm
Forum: Method and application
Topic: prediction using PLS
Replies: 1
Views: 2388

Hi, I think that should work. The outer weight of your MV will tell you how much the independent LV will change if your MV moves by one std-dev. The regression coefficient will then allow you to calculate by how much your dependent LV should go up. However, if your independent LV is operationalized ...
by stefanbehrens
Mon Jul 03, 2006 4:54 pm
Forum: Method and application
Topic: Relective or Formative
Replies: 5
Views: 5149

Here's an illustrative example for both types of measurement: Assume you would want to measure how drunk a particular person is. To do it reflectively , you could: - measure his/her blood alcohol level - observe how well he/she can still walk in a straight line - test his/her reaction time - etc. Al...
by stefanbehrens
Mon Jul 03, 2006 4:46 pm
Forum: Method and application
Topic: Changing the significance level
Replies: 2
Views: 3847

Hi Aden et al. Sorry, but I'm not sure I fully understand your question(s). The bootstrapping procedure in SmartPLS produces t-values for many of the parameters/ estimates of your model (e.g., loadings, weights or path coefficients). These t-values are your raw input for running significance tests. ...
by stefanbehrens
Wed Jun 21, 2006 6:52 pm
Forum: Method and application
Topic: Interpretation of latent variable scores
Replies: 7
Views: 6503

Hi Oliver, well, first of all, PLS doesn't really differentiate the way it calculates LV-scores for reflective or formative LVs. Both are simply a weighted sum of their MVs. Hence, no concern there. With regard to the sum of weights: The sum of all weights doesn't add up to 1 because PLS produces st...
by stefanbehrens
Tue Jun 20, 2006 11:11 pm
Forum: Method and application
Topic: Interpretation of latent variable scores
Replies: 7
Views: 6503

Well... a source I don't have for you. It's just following the logic of the PLS algorithm. In a regular PLS run, MVs would first be standardized, before LV-scores are calculated based on the weights. This procedure results in standardized LV-scores, i.e. mean=0, stddev=1. Thus, in principle, I don't...
by stefanbehrens
Tue Jun 20, 2006 12:59 pm
Forum: Method and application
Topic: Missing values
Replies: 3
Views: 3925

I don't think smartPLS' algorithm would allow pair-wise deletion. The only way this could work is to use pair-wise deletion in SPSS/ SAS to calculate a correlation/ covariance matrix. Then feed this matrix into Lohmoeller's LVPLS, which is also freely available on the web, to calculate your model. A...
by stefanbehrens
Tue Jun 20, 2006 9:27 am
Forum: Method and application
Topic: Missing values
Replies: 3
Views: 3925

It doesn't.

I.e. if you had your missings coded with say -999, SmartPLS would take these values for real and your results would obviously be very far off.

Good luck,
Stefan
by stefanbehrens
Sat Jun 17, 2006 6:59 pm
Forum: Method and application
Topic: Comparability of PLS resuts
Replies: 3
Views: 4898

Torben, with your post we are entering the interesting world of multi-group comparisons in PLS. As you might already have suspected, simply comparing the two path coefficients doesn't tell you whether the observed differences are really significant (or just mere coincidence). You have several option...
by stefanbehrens
Sat Jun 17, 2006 6:27 pm
Forum: Method and application
Topic: Interpretation of latent variable scores
Replies: 7
Views: 6503

Hi Oliver, you have to keep in mind that PLS works with standardized scores for the MVs and also produces standardized scores for the LVs. That is, each LV has a mean of 0 and a std-dev of 1 across all cases. That's why the observed ranges are not what you were expecting to find. To produce LV-estim...
by stefanbehrens
Thu Jun 01, 2006 1:49 pm
Forum: Method and application
Topic: Bootstraping and alpha significance
Replies: 2
Views: 3531

We've already had the same discussion in this forum.

Please have a look at:
viewtopic.php?t=151
by stefanbehrens
Wed May 24, 2006 12:25 pm
Forum: Method and application
Topic: Formula for calculating bootstrap t-value
Replies: 1
Views: 2605

Dear Dr. Kidd, the formula for the t-values is as follows: t = beta / stdev_bootstrap(beta) You have to keep in mind that the standard deviation of the bootstrap estimates is equivalent to the standard error of the original parameter estimate. Thus, there is no need to divide by SQRT(1000). Hope thi...
by stefanbehrens
Mon Apr 24, 2006 12:47 pm
Forum: Method and application
Topic: Critical values for t
Replies: 4
Views: 5403

Alistair, your degrees of freedom (df) for the student distribution are not derived from your sample size, but rather from the number of bootstrap runs. E.g., if you ran the bootstrap with 500 runs, your df would be 499. To determine the proper critical t-value, obviously you also need to consider w...