Processing Continuous and Categorical control variables.

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
Alexg
PLS User
Posts: 12
Joined: Fri May 18, 2018 11:34 am
Real name and title: Mr. Alessandro Grillo

Processing Continuous and Categorical control variables.

Post by Alexg »

Hello everyone,

hope you all had a good start of the week!

I have some doubts on how modelling control variables in Smartpls and processing the realtive data! I have been reading tons of posts on the topic still I cannot understand two things:

1) I have AGE which is treated as a continuous variable. It would be also a control variable. When modelling it should I create a FORMATIVE LATENT VARIABLE with one single indicator?

2) Education is a categorical variable and it also will be one of my control variables. Below the 5 values and labels.
1=less than high school
2=high school
3=bachelor's degree
4=master's degree
5=doctoral degree
I created 5 dummy variables on SPSS. Should I now create on PLS another Formative Latent Variable with those 5 dummies as indicators? Or should I create as many latent variables as the number of dummies variables and then link each latent to each dummy as indicator?

Thank you in advance for your help!
Alex
jmbecker
SmartPLS Developer
Posts: 1281
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Processing Continuous and Categorical control variables.

Post by jmbecker »

1) Yes, you need to include this as a single-indicator construct. Single-indicator constructs are neither formative nor reflective. They are just equal to the indicator. Therefore, you will also not find an arrow from or to the construct if you have only a single indicator.

2) Lohmöller (1989) describes some ways to model multi-categorical variables in one constructs using dummies as indicators. However, the explications are quite technical. An easier solution is to use single-indicator dummy constructs if you only want to control for the effect of the education variable. (Make sure that you have one dummy less the number of categories; the one that you omit is your reference category).
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
Post Reply