MGA: Differences between age, in different groups

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
MariaLunde
PLS User
Posts: 11
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

MGA: Differences between age, in different groups

Post by MariaLunde »

"I have a question and I hope someone can assist me.
I have a dataset in which I compare two groups, A and B, using a multigroup analysis.
So far, the results have been promising.

However, I also want to evaluate differences in age within each group.
To achieve this, I divided the original dataset into two age-based datasets (+/- 40 years) and conducted a multigroup analysis for the two groups (A and B) in both age-based datasets.

My question is, is this method appropriate? Should I consider the Bonferroni correction in some manner?
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: MGA: Differences between age, in different groups

Post by jmbecker »

In general, if you have a continuous variable using a split (e.g., median split, or any other arbitrary split) is less reliable than using the original continuous variable as a moderator in the model.
In your particular case, you may also end up with quite small groups that have low power to detect differences. Thus, I would always advice to include age as a moderator to investigate the impact of age differences on relationships.
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
MariaLunde
PLS User
Posts: 11
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

Re: MGA: Differences between age, in different groups

Post by MariaLunde »

Thank you very much for your response 😊
My sample size is >4000 customers, so it should be adequate in terms of splitting.

Is this a valid way to handle this research issue`?
I split the dataset into 2 parts based on age (>40 vs <40). Then, I conducted MGA on group A and group B separately, based on each of these age-split datasets. This resulted in a significant finding in group A, on age (< 40 years), on construct Y.

Is it appropiate to report this result, or do i need to include a age moderator in the model?
I am not sure how to include moderator in a MGA analysis...
Post Reply