Sample size (total sampling)

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
SAW
PLS Junior User
Posts: 1
Joined: Mon Jan 18, 2021 1:33 am
Real name and title: Shondi Ardi Wijaya

Sample size (total sampling)

Post by SAW »

Dear all,
I am conducting a research using smartPLS v3.3.2 student version. I have 6 latent variables , the greatest number of indicators is 6. Data population is 18, total sampling. Data can be run, all tests inner and outer has passed, the bootstrapping process has no problem resulting 2 of 5 hypothesis rejected. But I still have doubts because the sample size is very small. Some opinions said that the sample should 30-100 and less than 30 can't using PLS, some said there's no minimum sample rules for PLS. I'm a student and new in statistic. The questions are:
1. with conditions like the one above, is my research acceptable?
2. how many sample should i really use?
Please help me...
hdobson
PLS Junior User
Posts: 3
Joined: Mon Feb 21, 2022 9:32 pm
Real name and title: Mr Howard Dobson

Re: Sample size (total sampling)

Post by hdobson »

The Hair et al 2017 book has a table of recommended sample sizes, based on the number of arrows pointing at a variable and of course the R2 you have.

I suppose all data is acceptable up to a point when you mention the sample size but ideally you'd need a bigger sample according to Hair.

At 5% significance level you'd need 175, 75, 48 or 39 responses for R2s of 0.10, 0.25, 0.50, 0.75 respectively.
Post Reply