Great study & article. Perhaps incense is not the most scientific test, but it's effective enough for this setup.
Does anyone know if there is raspberry pi add-on or device that acts as an "air quality" sensor, or similar? It would be interesting to monitor this a bit more scientifically in real-time in a variety of environments.
Lastly, I think I've seen this here before, or on reddit. I was inspired to do something similar myself - I use a 19" box fan and a purple 3M MERV furnace filter, 20" square. Except the rounded corners they are almost exactly the same size. Some white duct tape around the edges to seal/hold in place. Total cost is about $40 usd for the fan & filter. I would anticipate a similar performance to the leaky MERV's, as mine is filtering more air as a ratio to total air flowed by the fan.