Seeking suggestion regarding formative constructs in SmartPLS

[Ahamed AFM Jalal]

Dear Sir,

Couple of months before I got some valuable suggestion from you. Now, I am struggling with some other problem, I would be obliged if you would kindly give your prudence suggestion on this. I am working with a formative construct with five items used to measure it. I am not sure how to report the reliability and validity of the formative construct in SmartPLS. Should I delete any of the five items, specially the negative one (item -4)?

 

The factor loading, AVE and communalities are as follows (I run the model in SmartPLS 2.0):

 

Number of items	Loading	AVE	Communality	Redundancy
Item 1	0.13	0.0	0.28	0.0020
Item 2	0.57
Item 3	0.15
Item 4	-0.62
Item 5	0.80

 

Thanks and best regards

Sincerely

Jalal Ahamed

 

 

[Tsipi Heart]

Dear Ahmed,

Please see in the following link: http://aisel.aisnet.org/icis2009/193/

Good luck,

Tsipi.

--
You received this message because you are subscribed to the Google
Groups "PLS-SEM" group.
To post to this group, send email to pls...@googlegroups.com
To unsubscribe from this group, send email to
pls-sem+u...@googlegroups.com

--
Dr. Tsipi Heart
Department of Industrial Engineering & Management
Ben-Gurion University of the Negev
Office: +972-8-6472255

[Ned Kock]

While I cannot speak for SmartPLS, I am providing here an answer based on WarpPLS to help clarify the different nature of formative versus reflective measurement and the assessment of formative latent variables. This answer also reflects some of my opinions based on various WarpPLS analyses by users and their own struggles. For example, sometimes latent variables don’t fit neatly one specific type – formative or reflective.

 

I strongly recommend reading the ICIS paper linked by Tsipi, which provides an overview of various issues related to PLS-based SEM.

 

The validity and reliability of formative latent variables is not tested in the same way as for reflective latent variables. They are tested in a way that could be characterized as referring to “validity and reliability”, but not in a classical sense, based on the outputs in the indicator weights window. Indicator weights are provided much in the same way as indicator loadings are, in a table format.

 

In the indicator weights table, all cross-weights are zero, because of the way they are calculated through PLS regression. Each latent variable score is calculated as an exactly linear combination of its indicators, where the weights are multiple regression coefficients linking the indicators to the latent variable. This brings the error term to zero in the regression equation relating the indicators to the latent variable.

 

P values are provided for weights associated with formative latent variables. These values can also be seen, together with those for loadings associated with reflective and moderating latent variables, as the result of a confirmatory factor analysis. In research reports, users may want to report these P values as an indication that formative latent variable measurement items were properly constructed.

 

As in multiple regression analysis, it is recommended that weights with P values lower than 0.05 be considered valid items in a formative latent variable measurement item subset. Formative latent variable indicators whose weights do not satisfy this criterion may be considered for removal.

 

In addition to P values, variance inflation factors are also provided for the indicators of formative latent variables. These can be used for indicator redundancy assessment. In reflective latent variables indicators are expected to be redundant. This is not the case with formative latent variables. In formative latent variables indicators are expected to measure different facets of the same construct, which means that they should NOT be redundant.

 

The VIF threshold of 3.3 has been recommended in the context of PLS-based SEM in discussions of formative latent variable measurement. A rule of thumb rooted in the use of this software for many SEM analyses in the past suggests an even more conservative approach: that capping VIFs to 2.5 for indicators used in formative measurement leads to improved stability of estimates.

 

The multivariate analysis literature, however, tends to gravitate toward higher thresholds. Also, capping variance inflation factors at 2.5 may in some cases severely limit the number of possible indicators available. Given this, it is recommended that variance inflation factors be capped at 2.5 if this does not lead to a major reduction in the number of indicators available to measure formative latent variables. One example would be the removal of only 2 indicators out of 16 by the use of this rule of thumb. Otherwise, the criteria below should be employed.

 

Two criteria, one more conservative and one more relaxed, are recommended by the multivariate analysis literature in connection with variance inflation factors in this type of context. More conservatively, it is recommended that variance inflation factors be lower than 5; a more relaxed criterion is that they be lower than 10. High variance inflation factors usually occur for pairs of indicators in formative latent variables, and suggest that the indicators measure the same facet of a formative construct. This calls for the removal of one of the indicators from the set of indicators used for the formative latent variable measurement.

 

These criteria are consistent with formative latent variable theory. Among other things, formative latent variables are expected, often by design, to have many indicators. Yet, given the nature of multiple regression, indicator weights will normally go down as the number of indicators go up, as long as those indicators are somewhat correlated, and thus P values will normally go up as well. Moreover, as more indicators are used to measure a formative latent variable, the likelihood that one or more will be redundant increases. This will be reflected in high variance inflation factors.

 

Formative latent variables, as well as other related issues and references, are also discussed in the User Manual for WarpPLS, and in the following articles, which are available from the “Publications” area on the www.warppls.com website.

 

Kock, N. (2011), Using WarpPLS in e-collaboration studies: Mediating effects, control and second order variables, and algorithm choices. International Journal of e-Collaboration, 7(3), 1-13.

 

Kock, N. (2010). Using WarpPLS in e-collaboration studies: An overview of five main analysis steps. International Journal of e-Collaboration, 6(4), 1-11.

[Yuyu Zhang]

Very useful comments and link. Thanks!

 

Yuyu

--

[Geoffrey Hubona]

I have not been keeping up with this thread as well as I should and it certainly is not because of a lack of interest.....but please see attached flyer....for my friends Christian and Joe Hair....I already have an advance pdf and I can promise that it will be a landmark piece (book by Springer) as far as explaining PLS.

 

And the best part? I don't know how these guys arm-wrestled the publisher down but the book's price will only be $35 for all 400 pages.....I've NEVER seen that in the academic publishing world, especially for a book that only a handful of people on the planet are capable of authoring (one of those is Ned Kock, in case you did not know that by now).

 

geoff hubona

[Ahamed AFM Jalal]

Dear All,

 

Thank you very much for your kind and useful suggestions. My special thanks to Professor Ned Kock, Professor José L. Roldán and Dr. Tsipi Heart. I am obliged

 

Thanks and best regards

Jalal Ahamed