I have one construct in SmartPLS model, which has more than 30 items, so please guide me in this regard that can I have that much items with one construct to analyse it in SmartPLS?
Please give me suggestions in this regard.
See this link
Article Convergent validity assessment of formatively measured const...
Hello Javed,
For CB-SEM or PLS-SEM, the more manifest indicators you have associated with a latent variable:
1. The more opportunity there is for "mis-fit," since you're asking the model to account for estimation of k * (k - 1) / 2 relationships (given k indicators).
2. Functionally, it may be harder to come up with more indicators that are demonstrably associated with (only one) latent variable, without introducing trivially similar indicators (e.g., answers to items such as "I love good health care," "I like good health care," "I prefer good health care").
3. The greater the number of indicators, the greater the number of model parameters there are to be estimated, and therefore the larger your sample size should be.
But no, there's no theoretical limit on the number of indicators.
Good luck with your work.
Hello David Morse....Thank you so much for your suggestions......is there any refernce available which says that there is no theoritcal limit on the number of indicators?
Smart-pls is flexible to run with minimum items for a construct, even one item for each variable.
For example, whenever we used Gender, there is only one item for gender and we can apply the structural model.
For more details, you can read the manual of Smart-PLS.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude