Some years back Al Rabe did some airfoil research and used his pickup truck somewhat like you're using that van. He lived in Texas and used flat roads to make the runs. He was looking for lots "purer" data than you are. But a couple of details should be mentioned.

He discovered that he needed to make back and forth runs and average the data. The idea was to consider the effect of the wind encountered at the time of the test. Now, you're not looking for the same kind of test results, but it is worthwhile to suggest that comparing your runs to each other should make some consideration of the winds. If the wind is at different direction and strength on the run with tab versus the run without, unless you're keeping track of the wind, there's a good chance the comparison will be skewed.

He also found that the test item needed to be in clean air, that the airflow near the vehicle was extremely turbulent. It's not only screwed up by the vehicle's body, but the body shields or exposes the test item depending on the winds. His results suggested that placing the test item close to the side of the car and back from the front of the car placed the item into extremely random and variable turbulence.

It's obvious you're looking for quite a different level of test results than Al was looking for. But it is worth noting what he discovered about testing with a car. His advice was to have the item as forward as possible, as far away from the body as possible, and to make runs both ways on as straight a road as possible. The difference in test results might give an indication of the variability of the test setup's effects or of the test conditions (wind speed and direction that day).

BTW, Al was quite an innovator. Seems he isn't the only one.