Hi,
please see viewtopic.php?p=832#832
SmartPLS is showing "Commulality" for each LV, but the correct will be "Average Comunnality".
Thanks.
Bido
Communality X AVE = Average Communality
- cringle
- SmartPLS Developer
- Posts: 818
- Joined: Tue Sep 20, 2005 9:13 am
- Real name and title: Prof. Dr. Christian M. Ringle
- Location: Hamburg (Germany)
- Contact:
Hi,
this is not really a bug!
All computations for communality are fine.
I understand that the term "Average Communality" is the precise naming (but that should be pretty much clear if computed for a latent variable and not on the manifest variable level). However, we keep this on our to-do-list.
Best regards
Christian
this is not really a bug!
All computations for communality are fine.
I understand that the term "Average Communality" is the precise naming (but that should be pretty much clear if computed for a latent variable and not on the manifest variable level). However, we keep this on our to-do-list.
Best regards
Christian
Prof. Dr. Christian M. Ringle, Hamburg University of Technology (TUHH), SmartPLS
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
- 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 we have the LV_2nd with 3 reflective LV_1st, with these loadings:
LV1 ==> 0.6
LV2 ==> 0.7
LV3 ==> 0.8
The communalities will be:
LV1 ==> 0.6^2 = 0.36 = 36% (variance of the LV1 explained by the LV_2nd)
LV2 ==> 0.7^2 = 0.49 = 49%
LV3 ==> 0.8^2 = 0.64 = 64%
The AVE will be: (0.36+0.49+0.64)/3 = 0.49666 = 40.6%
The composite reliability of the second order also must be computed by hand.
Best regards,
Bido
if we have the LV_2nd with 3 reflective LV_1st, with these loadings:
LV1 ==> 0.6
LV2 ==> 0.7
LV3 ==> 0.8
The communalities will be:
LV1 ==> 0.6^2 = 0.36 = 36% (variance of the LV1 explained by the LV_2nd)
LV2 ==> 0.7^2 = 0.49 = 49%
LV3 ==> 0.8^2 = 0.64 = 64%
The AVE will be: (0.36+0.49+0.64)/3 = 0.49666 = 40.6%
The composite reliability of the second order also must be computed by hand.
Best regards,
Bido
-
- PLS User
- Posts: 11
- Joined: Sat Dec 03, 2011 3:31 pm
- Real name and title:
-
- PLS User
- Posts: 11
- Joined: Sat Dec 03, 2011 3:31 pm
- Real name and title:
Hi all,
I found the formula for calculating Composite reliability:
Henseler et al. (2009) p.300
I checked using various examples in published papers; i.e. Wetzels et al. 2009, p: 188; and it worked for calculating Cr for second, third and fourth order constructs.
However, AVE for second, third, fourth order constructs works as suggested by Prof. Bido.
Hope this helps others :)
Best
Marwa
I found the formula for calculating Composite reliability:
Henseler et al. (2009) p.300
I checked using various examples in published papers; i.e. Wetzels et al. 2009, p: 188; and it worked for calculating Cr for second, third and fourth order constructs.
However, AVE for second, third, fourth order constructs works as suggested by Prof. Bido.
Hope this helps others :)
Best
Marwa
- 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 we have the LV_2nd with 3 reflective LV_1st, with these loadings:
LV1 ==> 0.6
LV2 ==> 0.7
LV3 ==> 0.8
The variance of error terms will be:
LV1 ==> 1 – 0.6^2 = 0.36 = 0.64
LV2 ==> 1 – 0.7^2 = 0.49 = 0.51
LV3 ==> 1 – 0.8^2 = 0.64 = 0.36
The CR will be: (0.6 + 0.7 + 0.8)^2/[(0.6 + 0.7 + 0.8)^2 + (0.64 + 0.51 + 0.36)] = 0.745
The composite reliability of the second order is computed in the same way that we compute the CR for the first order LV.
The problem with SmartPLS when computing the AVE and CR for the higher order LV is that it uses the loadings of the repeated indicators, not the loading from the second order LV to its first order LV (these loadings are reported as path coefficients).
Some references about AVE and CR:
FORNELL, C.; LARCKER, D. F. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, v. 18, p. 39-50, 1981.
HAIR JR., J. F.; BLACK, W. C.; BABIN, B. J.; ANDERSON, R. E. Multivariate Data Analysis. 7th ed. Upper Side River, NJ: Prentice Hall, 2010.
TENENHAUS, M.; ESPOSITO VINZI, V.; CHATELIN, Y.-M.; LAURO, C. PLS path modeling. Computational Statistics & Data Analysis, v. 48, n. 1, p. 159-205, 2005.
Best regards,
Bido
if we have the LV_2nd with 3 reflective LV_1st, with these loadings:
LV1 ==> 0.6
LV2 ==> 0.7
LV3 ==> 0.8
The variance of error terms will be:
LV1 ==> 1 – 0.6^2 = 0.36 = 0.64
LV2 ==> 1 – 0.7^2 = 0.49 = 0.51
LV3 ==> 1 – 0.8^2 = 0.64 = 0.36
The CR will be: (0.6 + 0.7 + 0.8)^2/[(0.6 + 0.7 + 0.8)^2 + (0.64 + 0.51 + 0.36)] = 0.745
The composite reliability of the second order is computed in the same way that we compute the CR for the first order LV.
The problem with SmartPLS when computing the AVE and CR for the higher order LV is that it uses the loadings of the repeated indicators, not the loading from the second order LV to its first order LV (these loadings are reported as path coefficients).
Some references about AVE and CR:
FORNELL, C.; LARCKER, D. F. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, v. 18, p. 39-50, 1981.
HAIR JR., J. F.; BLACK, W. C.; BABIN, B. J.; ANDERSON, R. E. Multivariate Data Analysis. 7th ed. Upper Side River, NJ: Prentice Hall, 2010.
TENENHAUS, M.; ESPOSITO VINZI, V.; CHATELIN, Y.-M.; LAURO, C. PLS path modeling. Computational Statistics & Data Analysis, v. 48, n. 1, p. 159-205, 2005.
Best regards,
Bido