## sample size estimation for multi group analysis

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
kini
PLS Junior User
Posts: 2
Joined: Fri Sep 08, 2023 12:44 pm
Real name and title: Kini.S PhD student

### sample size estimation for multi group analysis

Dear All,
I want to run a multi-group analysis in smart pls. For this, I would like to do an apriori estimation of the minimum sample size requirement for each group using G*power3.1. Which test should I use in G*power? Any guidance on this would be appreciated.
gertrudecolclough
PLS Junior User
Posts: 1
Joined: Fri Nov 03, 2023 7:32 pm
Real name and title: smartpls

### Re: sample size estimation for multi group analysis

kini wrote: Fri Sep 08, 2023 1:41 pm Dear Alluno online,
I want to run a multi-group analysis in smart pls. For this, I would like to do an apriori estimation of the minimum sample size requirement for each group using G*power3.1. Which test should I use in G*power? Any guidance on this would be appreciated.
According to my research, there are different methods and tests for conducting a multi-group analysis in PLS-SEM, such as the confidence intervals, the partial least squares multigroup analysis (PLS-MGA), the parametric test, and the Welch-Satterthwait test. Each method has its own assumptions and advantages, and you may need to choose the one that best suits your research question and data characteristics.

However, a general approach for estimating the minimum sample size requirement for each group is to use the one-way independent samples ANOVA test in G*power3.1. This test allows you to compare the means of two or more groups on a continuous outcome variable. To use this test, you need to specify the following parameters:

The type of power analysis: You should choose “A priori: Compute required sample size - given alpha, power, and effect size”.
The statistical test: You should choose “F tests: ANOVA: Fixed effects, omnibus, one-way”.
The type of effect size: You should choose “f”, which is the standardized mean difference between groups.
The input parameters: You should enter the values for alpha (the significance level, usually 0.05), power (the probability of detecting an effect, usually 0.80 or higher), and f (the effect size, which can be small, medium, or large, according to Cohen’s benchmarks5).
The output parameters: You should enter the number of groups that you want to compare, and the software will calculate the total sample size and the sample size per group.
You can find some examples and tutorials on how to use G*power3.1 for one-way independent samples ANOVA here, here, and here.