Search found 1175 matches

by jmbecker
Sat Jan 22, 2022 4:09 pm
Forum: FAQ (Methodology)
Topic: SIGN OF T-VALUES IN BOOTSTRAP
Replies: 6
Views: 3633

Re: SIGN OF T-VALUES IN BOOTSTRAP

The t-value threshold for 5% alpha level (or 95% confidence interval) would be 1.96 for a two sided test and 1.645 for a one-sided test. The p-value you are reporting seems to be calculated based on a two-sided test. For a two-sided test the 1.645 t value represents the 10% error level (or 90% confi...
by jmbecker
Sat Jan 15, 2022 11:39 am
Forum: FAQ (Methodology)
Topic: Discriminant Validity: HTMT of Higher Order Constructs
Replies: 1
Views: 3814

Re: Discriminant Validity: HTMT of Higher Order Constructs

Have you calculated the HTMT for the higher order construct PI correctly? You cannot use the numbers from SmartPLS but have to calculate this by hand if you use the repeated indicator approach: https://www.researchgate.net/publication/333877607_How_to_Specify_Estimate_and_Validate_Higher-Order_Const...
by jmbecker
Sat Jan 15, 2022 11:35 am
Forum: FAQ (Methodology)
Topic: Endogeneity in PLS-SEM
Replies: 1
Views: 4032

Re: Endogeneity in PLS-SEM

Generally, you should include all relevant copulas at the same time. However, adding copula terms for all predictor variables might not be very efficient and the results are not very trustworthy, because you need very large datasets to reliably estimate such models. Thus, I would first think about w...
by jmbecker
Sat Jan 15, 2022 11:27 am
Forum: FAQ (Methodology)
Topic: FIMIX segmentation
Replies: 1
Views: 3811

Re: FIMIX segmentation

1. In the results you will find the "Segment assignment". This first tab is the "FIMIX probabilities of segment membership". This is the native result of the algorithm which create a probabilistic segment assignment (i.e., each observation/case has a specific probability of belon...
by jmbecker
Fri Jan 14, 2022 9:13 am
Forum: FAQ (Methodology)
Topic: Bootstrapping Method
Replies: 1
Views: 3904

Re: Bootstrapping Method

There are many different versions of bootstrapping described in these texts, but what PLS algorithms rely on (not only SmartPLS) is usually a simple case bootstrap. That is resampling entire cases (i.e., observations, rows) of the data matrix with replacement.
by jmbecker
Fri Jan 14, 2022 8:56 am
Forum: FAQ (Methodology)
Topic: Unstandardized path coefficients
Replies: 1
Views: 3826

Re: Unstandardized path coefficients

The PLS algorithm always require standardization of the latent variables. However, in SmartPLS 3 you get unstandardized path coefficients in the IPMA analysis. There latent variable scores are unstandardized using the Groenhold method and then unstandardized regression weights are calculated on thes...
by jmbecker
Fri Jan 14, 2022 8:52 am
Forum: FAQ (Methodology)
Topic: Outer VIF Calculation
Replies: 1
Views: 4650

Re: Outer VIF Calculation

VIF = 1/ (1-Ri²), where Ri² is the r-square of the focal predictor i, when regressing all other predictors on this variable. In the outer model, the focal predictor i would be a specific manifest variable of a latent variable, and all other predictors are the other manifest variables of that predict...
by jmbecker
Fri Jan 14, 2022 8:48 am
Forum: FAQ (Methodology)
Topic: Bootstraping Samples
Replies: 1
Views: 4358

Re: Bootstraping Samples

ATTR->COMP means the inner model relationship between ATTR (Attractiveness) and COMP (Competence) in the corporate reputation model. What is shown depends on the report you are looking at (which you have cut out). But assuming that you are in the "Path Coefficient" report, these are the pa...
by jmbecker
Fri Jan 14, 2022 8:43 am
Forum: FAQ (Methodology)
Topic: Sensitivity Analysis
Replies: 1
Views: 2082

Re: Sensitivity Analysis

Maybe the reviewer wants to see some robustness checks on the findings like explained here: https://www.researchgate.net/publication/330566055_Structural_model_robustness_checks_in_PLS-SEM Or he wants to see if alternative models like CB-SEM or regression on sum-scores come to similar conclusions. W...
by jmbecker
Fri Jan 14, 2022 8:41 am
Forum: FAQ (Methodology)
Topic: PLS and Meta-analysis
Replies: 1
Views: 441

Re: PLS and Meta-analysis

I have not seen any literature that uses PLS to analyze data from a meta analysis. I think it is an interesting idea, but you probably have to build the literature yourself.
by jmbecker
Fri Jan 14, 2022 8:37 am
Forum: FAQ (Methodology)
Topic: Sample size calculation using G*power Analysis
Replies: 21
Views: 61562

Re: Sample size calculation using G*power Analysis

Your sample size is sufficient to detect effects with an effect size of 0.05 with 5% error probability and 80% power in each group. However, that does not mean that you are able to detect differences between the groups. That would be another different power analysis.
by jmbecker
Tue Jan 11, 2022 8:48 pm
Forum: FAQ (SmartPLS 2.0)
Topic: Unstandardized regression coefficients
Replies: 4
Views: 6928

Re: Unstandardized regression coefficients

Smartpls 2 has never been updated and thus still contains the bug. SmartPLS 3 is doing the calculations correctly.
by jmbecker
Wed Apr 28, 2021 3:33 pm
Forum: FAQ (Methodology)
Topic: SmartPLS basics
Replies: 3
Views: 12666

Re: SmartPLS basics

This would be a quite unusual practice, but you may also have good theoretical arguments. If a question does not fit to the other questions, you might see such behavior. If you can now argue that the questions (theoretically) also is the one that fits the least to the construct (based on your constr...
by jmbecker
Wed Apr 28, 2021 12:36 pm
Forum: SmartPLS Small Talk Corner
Topic: PLS-MGA p-value result missing
Replies: 1
Views: 2767

Re: PLS-MGA p-value result missing

My first guess would be that you have the "hide zeros" option on so that all zeros are not shown. This option is quite useful for some results, but not for all.
by jmbecker
Mon Apr 19, 2021 12:20 pm
Forum: Requested Features
Topic: Simple mediator effect
Replies: 1
Views: 2032

Re: Simple mediator effect

No. That would not be correct. There exists a formula to calculate standard errors for mediation using Sobels formulas. However, Sobel is not recommended anymore because it is less accurate and robust in many situations. Thus, bootstrapping the indirect effect is the preferred choice nowadays.