1. in a research study, I'm doing using SEM, there are six paths which are theoretically well-established in the literature, and three paths (from the exogenous to three mediating constructs) whose literature isn't strong. should I still use PLS-SEM or not?
2. using two different techniques of determining the sample size for PLS-SEM, I ended up with 146 and 147 cases, which are both over the number of population cases (140). this seems a little odd to me.
two methods are Hair et al. (2013) table on sample size recommendation and Gamma exponential method by Kock and Hadaya (2016).
in case of PLS-SEM being the right approach, should I continue with studying all the population (140)?
I recommend hair et al. 2013 , there are will be useful in your case, note that your population less than 200, you should use smart pls ,
Thank you dear Mahmoud, but the point is that using either methods the sample size would be higher than the population size which is 140; using Hair et al (2013) method, the number would be 147. I really doubt whether I should study all the population or just increase the significance level to 10% to study 120 cases out of 140.
It's actually a table in which there are three criteria, and based on them you could determine the sample size.... I haven't heard about Gpower test.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling