To me all of those "reasons" are purely style choices (with the exception of tab control being easier for a beginner initially - which is why I provided a sample VI).
Using a tab control in this manner may be quicker to implement initially, but leave the program far less expandable later (say for example he realizes half way through a larger application that he needs to have two of the controls visible at the same time. This may end up in an architecture redesign - especially if he was using the tab control as part of the state machine).
As for using the selected tab as part of a case structure: yes, this is possible, but doing so would basically be a state machine. Standard is to use enum constants so that your states have actual meaning.
I would agree that managing references is a pain to start with, but there are many things to make it far simpler. (Arrays of bundled reference and label for example).
