Dr. Garrison,
I am using Structural Equation Modeling and through the analysis to improve model fit and work with the sample, admittedly smaller than ideal (n=99), my modified model worked much better when I removed Cognitive presence from the model, yet the theoretical framework and survey used was based on CoI. Have you seen other papers that are of quality that are able to use some of the model, but not the entire thing?
To explain this limitation our sample size was a contributing factor and secondly this is a mastery type course of introductory computer skills where learners come in at a wide variety of previous experiences.
Thank you,
Johnathan Yerby
Georgia State University
