No data for AVE, Composite Reliability or Cronbach's Alpha

[JohnP]

Dear qualified users,

1) We have an output with 0,0000 values for all LV's concerning AVE, Composite Reliability and Cronbach's Alpha. We understand, that "Communality" in the PLS report output is actual Average Communality and therefore equal to AVE. We have values of "Communality in the range 0.41-0.86, and therefore we know that our AVE's for the LV's is not 0,0000.

We have imported the ECSI file from Smartpls.de, and here values for AVE, Composite Reliability and Cronbach's Alpha are shown - so the problem should not lie in the software, we guess.

Is it really possible, that the fit of our model is that bad? Weights and outer loadings does not suggest this.

In our model we only have 2 levels of LV's and not multiple like in the ECSI file/model. Is this causing the problem?

2) How is it possible to display values of ex. AVE, Composite Reliability and Cronbach's Alpha inside the LV' circles? We can't make it work by adding them in "display settings" under "preferences" under "window".

3) Is it possible to show loadings in the output model instead of weights?

We sincerely thank you in advance for your help.

[VerenaGormanns]

I have the same problem. I don't get no values for Chronbach's alpha, AVE, and Composite Reliability. Please help :)

[JohnP]

I could not make it work with the formative method, but when we reversed the arrows (reflective method), then we got the values for AVE, Composite Reliability and Cronbach's Alpha.

[VerenaGormanns]

Thank you very much. I'll hope this works out.
Best,
Verena

[stephan.kramer]

Please be aware that classical test theory (with AVE, Cronbach's Alpha etc.) does not apply to formative constructs. Thus, the statistics are not included for good reason!!! Before running PLS, you might take deeper consideration about the theoretical properties of your construct and the statistical assessment!

Best regards, Stephan

[JohnP]

Thank you very much Stephan.

Do you know were it is possible to read more about the use of the formative method and the test theory of this method?

/John

[stephan.kramer]

Hi John,

A good overview about the differences between reflective and formative constructs and their different evaluation can be found in:

Henseler, J., Ringle, C.M., & Sinkovics, R.R. (2009). The Use of Partial Least Squares Path Modeling in International Marketing. Advances in International Marketing 20, p. 298 f.

Best, Stephan

[JohnP]

Thank you very much!

[jmbecker]

These two papers are also very helpfull for this topic:

Petter, S., Straub, D., and Rai, A. "Specifying Formative Constructs in Information Systems Research," MIS Quarterly (31:4) 2007, pp 623-656.

Cenfetelli, R.T., and Bassellier, G.v. "Interpretation of Formative Measurement in Information Systems Research," MIS Quarterly (33:4) 2009, pp 689-707.