Can external validity value be larger than 1?
I tried to calculate using Henseler et al. 2009 for a second-order measure.
First, I calculated the CR by hand for the reflective second-order measure, which was .88. Second, I regressed the formative second-order measure on reflective second- order measure and got their correlation, which was .99.
Finally, I put all these numbers in the formula which gave me -.14. I subtracted this amount from 1 and the result was 1.14 :(
Is this value correct? If not where am I making a mistake?
External validity for a formative second-order LV
- Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Hi,
If I understand you correctly,
You are correction the correlation for attenuation (or disattenuation of the correlation).
r-dis = r / CR
This formula is an approximation and could result in values greater than 1, in these cases we assume that the correlation is one, see:
NUNNALLY, J. C.; BERNSTEIN, I. H. Psychometric theory. 3rd ed. New York: McGraw-Hill, 1994.
Curiosity: already in the beginning of 1900, Pearson disagreed of Spearman about this.
Best regards,
Bido
If I understand you correctly,
You are correction the correlation for attenuation (or disattenuation of the correlation).
r-dis = r / CR
This formula is an approximation and could result in values greater than 1, in these cases we assume that the correlation is one, see:
NUNNALLY, J. C.; BERNSTEIN, I. H. Psychometric theory. 3rd ed. New York: McGraw-Hill, 1994.
Curiosity: already in the beginning of 1900, Pearson disagreed of Spearman about this.
Best regards,
Bido