Hi,
I am using smartPLS for the data analysis of my master thesis. I performed the data (n=65) in smartPLS and got the output, but I got some questions with regards to 'individual item reliability'. To my knowledge, the observed variables should have a loading of 0.70 or more.
1. Do outer loadings in the output of smartPLS mean factor loadings?
2. Can outer loadings be used to examine 'individual item reliability'?
Besides, I got the results of outer loadings. Some loadings of the indicators are below 0.70 (0.4298, 0.6254,0.6141, 0.5551, 0.6772, 0.6835). but values of Cronbachs Alpha (>0.70), Composite Reliability (>0.70) and AVE (>0.50) exceed the suggesting level. What should I do with those poor indicators?
Thanks!
Best regards,
Lelia
individual item reliability
- Diogenes
- PLS Super-Expert
- Posts: 899
- Joined: Sat Oct 15, 2005 5:13 pm
- Real name and title:
- Location: São Paulo - BRAZIL
- Contact:
Hi Lelia,
1) No, the outer loading = factor loading
Outer weights are used to compute the latent variable scores.
Score = o.w.1 * X1 + o.w.2 * X2 + o.w.3 * X3
2) Yes. Thinking that each indicator measures the same construct, we could think that each indicator is a different way to measure the same latent variable (same idea of reliability).
About the values of the outer loadings:
a) In a confirmatory study, all the indicators should be kept.
When we dropped the worst indicators (low outer loading), we are adjusting the model to the data, for this reason a second sample (validation) should be used.
b) Another reason to keep the indicator with low outer loading is the validity content, remembering that in your case some individual loading are lower than 0.7, but the mean (AVE) is bigger than 0.5 (ok).
Best regards,
Bido
1) No, the outer loading = factor loading
Outer weights are used to compute the latent variable scores.
Score = o.w.1 * X1 + o.w.2 * X2 + o.w.3 * X3
2) Yes. Thinking that each indicator measures the same construct, we could think that each indicator is a different way to measure the same latent variable (same idea of reliability).
About the values of the outer loadings:
a) In a confirmatory study, all the indicators should be kept.
When we dropped the worst indicators (low outer loading), we are adjusting the model to the data, for this reason a second sample (validation) should be used.
b) Another reason to keep the indicator with low outer loading is the validity content, remembering that in your case some individual loading are lower than 0.7, but the mean (AVE) is bigger than 0.5 (ok).
Best regards,
Bido