PLS with small samples
PLS with small samples
Hi, I am working at the moment with some old data (2008) that came from a questionnaire realized to a group of companies (n=15). I have built a path-mode based on the questionnaire variables, but I believe that due to the small size of the sample I should avoid using PLS because it will probably create unstable results.
I would like to know if anyone has any suggestion, besides collecting more data which is not possible =(, regarding the way I should analyize my data.
Thanks for your help.
I would like to know if anyone has any suggestion, besides collecting more data which is not possible =(, regarding the way I should analyize my data.
Thanks for your help.
I also heard about this rule of thumb. But how to cite it in the papers? As I am new users of SmartPLS, is it possible to assist me? Thanks in advance.Hengkov wrote:Hi,
Rule of thumb sample size in PLS 10x path/predictor in model. If you have simple model (just one independent and one dependent) N = 15 it's OK.
Regards,
Hengky
samble size rule of thumb
Hi
Refer this article
Hair, J. F., et al. (2011). "PLS-SEM: Indeed a silver bullet." The Journal of Marketing Theory and Practice 19(2): 139-152.
_Robinson_
Refer this article
Hair, J. F., et al. (2011). "PLS-SEM: Indeed a silver bullet." The Journal of Marketing Theory and Practice 19(2): 139-152.
_Robinson_
-
- PLS Expert User
- Posts: 51
- Joined: Tue Nov 04, 2014 5:27 am
- Real name and title: Dr. Bradley Wilson. Senior Lecturer in Advertising. School of Media and Communication. RMIT University.
Re: PLS with small samples
Hello,
rules of thumb are really quite useless. It depends with you have enough power to detect small and large effects respectively.
If you have small effects sample size will need to be larger to detect the smaller effects.
Wynne chin writes about this subject in a chapter with Newsted. Samples sizes in business research.
Brad
rules of thumb are really quite useless. It depends with you have enough power to detect small and large effects respectively.
If you have small effects sample size will need to be larger to detect the smaller effects.
Wynne chin writes about this subject in a chapter with Newsted. Samples sizes in business research.
Brad
Bradley Wilson. Ph.D. Senior Lecturer in Advertising. School of Media and Communication. RMIT University. Melbourne, Australia.
SEE: http://www.researchgate.net/profile/Bradley_Wilson2 http://scholar.google.com.au/citations? ... AAAJ&hl=en
SEE: http://www.researchgate.net/profile/Bradley_Wilson2 http://scholar.google.com.au/citations? ... AAAJ&hl=en
- Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Re: PLS with small samples
Many factors play role important to determine sample size in context PLS:
1. Variability of data
2. Confidence interval
3. Design research
4. and much more
Rule of thumb is standard practice to assess the adequacy of the sample size. If you collect data using probability sampling, may be you need to compute sample size using formula.
Best regards
1. Variability of data
2. Confidence interval
3. Design research
4. and much more
Rule of thumb is standard practice to assess the adequacy of the sample size. If you collect data using probability sampling, may be you need to compute sample size using formula.
Best regards
- cringle
- SmartPLS Developer
- Posts: 818
- Joined: Tue Sep 20, 2005 9:13 am
- Real name and title: Prof. Dr. Christian M. Ringle
- Location: Hamburg (Germany)
- Contact:
Re: PLS with small samples
Yes, rules of thumb are quite useful. @Brad: The literature of the nienties is good but a lot of things happend since then ;)
Best
Christian
Best
Christian
Prof. Dr. Christian M. Ringle, Hamburg University of Technology (TUHH), SmartPLS
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
-
- PLS Expert User
- Posts: 51
- Joined: Tue Nov 04, 2014 5:27 am
- Real name and title: Dr. Bradley Wilson. Senior Lecturer in Advertising. School of Media and Communication. RMIT University.
Re: PLS with small samples
rules of thumb should be used after people have tried to estimate sample size for sufficient power to unearth (or not) the effect size they might be expecting.
most of us as reviewers hate to see small sample size as the only reason PLS was chosen.
Brad
most of us as reviewers hate to see small sample size as the only reason PLS was chosen.
Brad
Bradley Wilson. Ph.D. Senior Lecturer in Advertising. School of Media and Communication. RMIT University. Melbourne, Australia.
SEE: http://www.researchgate.net/profile/Bradley_Wilson2 http://scholar.google.com.au/citations? ... AAAJ&hl=en
SEE: http://www.researchgate.net/profile/Bradley_Wilson2 http://scholar.google.com.au/citations? ... AAAJ&hl=en
- Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Re: PLS with small samples
Yes, the number of samples in the PLS should consider the effect size and power analysis as well. Maybe we need to make a special software to count the number of samples in the PLS method or add a feature on SmartPLS to compute it. #just my idea.
Greetings,
Greetings,
-
- PLS Junior User
- Posts: 1
- Joined: Thu Mar 12, 2015 4:23 pm
- Real name and title: Habib Abdula
Re: PLS with small samples
If you have simple model (just one independent and one dependent) N = 15 it's OK. ......
- Hengkov
- PLS Super-Expert
- Posts: 1599
- Joined: Sun Apr 24, 2011 10:13 am
- Real name and title: Hengky Latan
- Location: AMQ, Indonesia
- Contact:
Re: PLS with small samples
Hi,
No, If you consider the effect size and power analysis, sample you need to be around 43 (calculated with G * Power). PLS is not used to excuse the small sample, this is just a misconception. PLS also took samples were rational, but is not as CB-SEM. In addition, if the number of indicators > 15 will not be able to run the model
Greetings,
No, If you consider the effect size and power analysis, sample you need to be around 43 (calculated with G * Power). PLS is not used to excuse the small sample, this is just a misconception. PLS also took samples were rational, but is not as CB-SEM. In addition, if the number of indicators > 15 will not be able to run the model
Greetings,