BSC isn't done with testing assemblies with that rig, but they say testing any one assembly is a laborious process. I'm pretty comfortable that wet-sprayed cellulose and properly caulked sheathing & framing would blow away any of the shimmed-leakage assemblies described too, with or without a membrane vapor & air barrier.

I expect there will be a much broader publication on this project forthcoming- this was just the preliminary view. The assemblies tested were all clear-walls too (13.6$% framing fraction) with no windows or doors, which would be yet another set of leaks to consider. While we'd all like them to test exactly what we're building, the possible combinations are endless.

The the only eye-popping result in their findings outlined in the 3-pager (http://www.buildingscienceconsulting.com/presentations/documents/TM%20Research%20Project%20Summary_rev.pdf) is the last line:

"Conventional energy models (i.e. those that account for air leakage energy using Q=mcdT) may over predict the
negative energy impact on walls that have a significant interaction effect (e.g. air moving through insulation)."

The heat-exchanger effect present with fiber insulation mitigates a large fraction of what had been presumed to be pure losses in simpler models(!).