Here is a link to a suggested process for testing model props (with minimal equipment, and focus on results, not theory) - Its been around for several years and may have some relevant data reported.

http://www.sciencebuddies.org/scienc...ero_p018.shtml