Foreword: In 2006, the Chrysler LX platform had no tuning and no datalogging tools available. What passed for a good test then is nothing like you would be able to do now. Still, for the time these were the most extensive tests out there...
These tests were conducted with what I called a "Frankentake", which was essentially a partial copy of someone else's idea: a K&N Typhoon modified to take a much larger filter. I altered the original idea in such a way as to require the removal of the heat shield, leaving the filter completely exposed to the engine bay. I did this in order to lower the filter down into a known cold air pocket in the engine cavity.
Was the heat shield expendable? Would the cold air pocket more than counteract the benefit of that shield? Cursory examination said 'yes'. That conclusion would be regarded as crazy unless a thorough, clear and reproducible test was conducted.
These test results and others that followed have influenced several commercial intake designs for the LX and LC platforms ... and spawned endless technical arguments ever since.
Relevant Pre-Existing Modifications
The tests themselves were conducted on multiple days under different conditions and are linked on the menu at left.
Testing Conclusions: If you perform some simple modifications to allow air inflow and heat outflow in the engine bay, amazingly, a heat shield is of minimal value.