forum.smartpls.com

Hello, a while ago, I used a variation of the mediation analyses proposed by Shrout and Bolger [1] with PLS. The basic idea was to use the bootstrapping results for the "a" and "b" paths from each bootstrapping iteration (which are available from the SmartPLS output) and calculat...

Hello, at the risk of pointing out the obvious: Why not a) estimate a path model with the whole dataset, b) create new datasets for each group from the the estimated case values, c) run separate analyses for each group with these case values (and hence identical measurement models), and d) compare p...

Hi, yes, you may use the distribution from the bootstrap samples to obtain CIs. These CIs, however, tend to be too narrow, because they are not bias-corrected [1]. Best, Joachim [1] Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and reco...

Hi, could it be a consequence of small segments, i.e., a) few people and b) many path coefficients in the model? This easily leads to path coefficients > 1 due to multicollinearity. In that case one should probably be careful to have a look at indicators of multicollinearity (tolerance/VIF, intercor...

Hi Lars, in general, usually the same quality criteria are used in PLS as in multiple regression regarding effect sizes, etc. You might, however, consider using f² as a more specific effect size if you're interested in the effects of single exogenous variables. Best wishes, Joachim (BTW, I always wo...

Hello, there's no way to directly do this in SmartPLS (apart from looking at the diverse quality indicators like AVE etc., of course). Instead, you might want to have a look at the "vanishing tetrads" method suggested by Bollen and Ting, which does not involve PLS at all though: Bollen, K....

Dear all, this is an interesting issue. I believe there are some thoughts on this in Wold (1982) [1], but I don't know of any other references. Some time ago, there was a small discussion of this on SEMNET (archives available at http://www.bama.ua.edu/cgi-bin/wa?A0=SEMNET&X=-, July 3). I'll just...

Hello, yes, your scenario sounds very much so. As you probably know, you can model MIMIC models by a) "splitting up" your latent and then b) let your formative latent predict the reflective one. This is mathematically equivalent to using a MIMIC model, a good reference on this is Wold (198...

Hello, thanks for your replies, and sorry for being imprecise—of course I meant the bootstrapping. And also thanks for your moral and methodological support (the Hesterberg et al. chapter really is a good resource). So I guess I will just go along with my indicators... Although R² in the latent is s...

Dear all, because PLS is non-parametrical, it should also handle rather skewed distrubutions adequately (e.g., rare events following a Poisson distribution), right? I was wondering if someone here had a good reference on this. I am also most interested in analyzing statistical power in this case, an...

Hello, I was wondering if anyone used the formula 2 suggested by Chin -- or rather, the suggested adjustment of the degrees of freedom in formula 3. Unfortunately, formula 3 on the web site [1] does not seem to work for me, because the adjusted df values seem to be much too low . I was under the imp...

Hello,

so far the only way I know for inspecting the cross-loadings is to use SPSS. IMHO, it might be more appropriate to use ordinal correlations (Kendall, Spearman) instead of Pearson if you want to stick to non-parametric tests.

Best wishes,

Joachim

Hi all, because this topic seems to come up again every now and then, I thought I might point out that data management has become much easier using SPSS 15. 1) Using a SPSS Syntax like SAVE TRANSLATE OUTFILE='FILENAME.XLS' /TYPE=XLS /MISSING=RECODE [...] . and only keeping numeric variables, of cour...

Hello,

thank you for pointing me to the new 3.0 release (yeah, finally! :-) ) . It really is a major improvement over the previous 2.0 version and much easier to use.

Kind regards,

Joachim

Dear all, I am estimating a model that contains two 2nd order latents. To be on the safe side, I used a 2-step procedure to handle to 2nd-order constructs: 1) I ran separate models to do the basic 2nd-order estimation (cf. Wold, 1982, pp. 40). (E.g., how do several aspects of satisfaction relate to ...