Control variables/measurement error with multiple responses

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
swcsschm
PLS Junior User
Posts: 1
Joined: Tue Feb 03, 2009 11:43 am
Real name and title:

Control variables/measurement error with multiple responses

Post by swcsschm »

Dear all,

I would like to use SmartPLS for evaluating causality in team relations. Two points remain somehow unclear:

1. How do I model control variables in SmartPLS? Are they just moderators (which might then inflate my sample size calculated by G*Power?)?

2. I get survey responses from team members (n>=1 per team). Some latent variables are subjective evaluations of an “objective” state of team relations (i.e. there can only be one such status per team (e.g. it is objectively high, medium or low) but different team members might disagree in their valuations => measurement error). Is there any standard way to treat team-member variations of the evaluation of these “objective” latent constructs in the same team? Should I take an average of these or eliminate the teams with too high variations on the “objective” variables?

Any help would be greatly appreciated.

Best Regards,

Christian
Post Reply