Sample size calculation using G*power Analysis

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
jedisummary
PLS Junior User
Posts: 1
Joined: Sat Oct 06, 2018 4:58 pm
Real name and title: jedisummary

Re: Sample size calculation using G*power Analysis

Post by jedisummary »

Hosam wrote: Thu Nov 26, 2015 7:48 pm However, if you want to use Gpower you could get sample size of 138 respondent as the following (please correct me if I'm mistaken)

F tests - Linear multiple regression: Fixed model, R² increase
Analysis: A priori: Compute required sample size
Input: Effect size f² = 0.15
α err prob = 0.05
Power (1-β err prob) = 0.95
Number of tested predictors = 5
Total number of predictors = 5
Output: Noncentrality parameter λ = 20.7000000
Critical F = 2.2828562
Numerator df = 5
Denominator df = 132
Total sample size = 138
Actual power = 0.9507643
Thanks for the link.
danurahardja
PLS Junior User
Posts: 2
Joined: Tue Dec 24, 2019 9:53 am
Real name and title: Danu Rahardja Singgih

Re: Sample size calculation using G*power Analysis

Post by danurahardja »

can anyone show me a tutorial on how to use the program? i a currently doing my thesis, with 2 independent variable, 1 mediator and 1 dependent
astratom
PLS Junior User
Posts: 1
Joined: Fri Feb 21, 2020 7:10 am
Real name and title: astra tomy
Contact:

Re: Sample size calculation using G*power Analysis

Post by astratom »

thanks for the help guys
chenchenchen
PLS Junior User
Posts: 5
Joined: Mon Mar 01, 2021 1:56 am
Real name and title: chentianzhi student

Re: Sample size calculation using G*power Analysis

Post by chenchenchen »

Hello all,

I have a reflective-formative type model and there are two first order latent variables. I am interested in testing the weights of the two first order latent variables to second order latent variable. I will use the disjointed two stage approach to model analysis. Should I calculate the sample size in the second stage by using n=2 for the number of predictors and the t-test method? Or should I calculate the sample size in the first stage by using n=20 for the number of predictors and the t-test method? (one of the first order latent variables have 20 indicators and another have 9 indicators.)

Thank you.
navya66
PLS Junior User
Posts: 7
Joined: Thu Jul 09, 2020 1:36 pm
Real name and title: Navya Jain , Research Scholar

Re: Sample size calculation using G*power Analysis

Post by navya66 »

Hi everyone,
In continuation to the discussion here, I had a query.
When calculating minimum sample size in GPower for a study that has one tailed hypotheses, I have used the following settings (based on 2nd order, 2 latent variables predicting another 2nd order latent variable, all reflective):

t-test
Linear multiple regression: Fixed model , single regression coefficient
A priori analysis
One tailed
Effect size 0.05
Alpha error prob. 0.05
Power 0.80
No. of predictors = 2
This gives me a minimum sample size of 126

As part of my analysis, I have done a group specific analysis (2 groups with greater than 126 sample size) as well as analysis of the total dataset (greater than 300 sample size).
Am I correct to assume that I have followed the correct settings for GPower and I can say that the sample size of the study is sufficient based on the above GPower A priori analysis?
jmbecker
SmartPLS Developer
Posts: 1281
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Sample size calculation using G*power Analysis

Post by jmbecker »

Your sample size is sufficient to detect effects with an effect size of 0.05 with 5% error probability and 80% power in each group. However, that does not mean that you are able to detect differences between the groups. That would be another different power analysis.
Dr. Jan-Michael Becker, BI Norwegian Business School, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
navya66
PLS Junior User
Posts: 7
Joined: Thu Jul 09, 2020 1:36 pm
Real name and title: Navya Jain , Research Scholar

Re: Sample size calculation using G*power Analysis

Post by navya66 »

Yes, I have understood your point. As I mentioned, I have done separate analysis for each group as well as an overall analysis, I have not done any comparison of the two groups.
Thank you Dr. Becker and wish you a very happy new year :)
rathor1072@gmail.com
PLS User
Posts: 16
Joined: Wed May 18, 2016 6:59 pm
Real name and title: shekhu

Re: Sample size calculation using G*power Analysis

Post by rathor1072@gmail.com »

There is another method to calculate power
Aguirre-Urreta, M., & Rönkkö, M. (2015). Sample size determination and statistical power analysis in PLS using R: an annotated tutorial. Communications of the Association for Information Systems, 36(1), 3.

Has anyone tried this?
michaelaskew
PLS Junior User
Posts: 1
Joined: Wed Jul 26, 2023 5:43 am
Real name and title: michael askew

Re: Sample size calculation using G*power Analysis

Post by michaelaskew »

To perform a sample size calculation using G*Power analysis, follow these steps:

Choose a Statistical Test: Determine the specific statistical test you plan to use for your study, such as t-test, ANOVA, regression, etc.

Select Effect Size: Decide on the effect size you want to detect. This is often based on previous research or your expectations.

Choose Significance Level (α): Set the significance level, typically denoted as α (alpha), which represents the probability of making a Type I error (false positive). Common values are 0.05 or 0.01.

Select Power (1 - β): Decide on the desired statistical power, often denoted as 1 - β (beta). This is the probability of correctly rejecting a false null hypothesis (i.e., avoiding a Type II error). Common values are 0.80 or 0.90.

Indicate Number of Groups (if applicable): If your study involves multiple groups, indicate the number of groups you will be comparing.

Specify Other Parameters: Depending on the test, you might need to input additional parameters, such as the correlation between variables, degrees of freedom, etc.

Run G*Power: Input all the above information into the G*Power software. It will calculate the required sample size for your study.

Interpret Results: G*Power will provide you with the recommended sample size based on your inputs. https://coinflipper.net/
Post Reply