I would argue both that your study was scientific. and the results are statistically significant. :)
The reason I said un-scientific was because there are many uncontrolled variables in my setup. Such as entrance direction, entrance height and surround environment (some where on barns, others on trees). Also, I believe geographical location plays a big part in bee behavior. For instance, the deep south has much more absconding than here in the north. Is it because there are many more nesting places available, or perhaps because they have more time to prepare for a milder winter to survive so they can be more choosy. There is the opportunity to learn from any study, just be cognizant of the environmental differences.
I hear what you saying, all good points, and not to beat a dead horse :deadhorse: ....BUT!
My point was you had a controlled sample size and controlled a number of variables, swarm trap size, general location, etc. Granted you didn't control all variables, I'm not sure any scientific study can in researching this particular subject. No doubt, some studies control more than others, therefore warrant more merit.
But for clarification, while certainly not perfect, I think your study meets an acceptable minimum to qualify as scientific. To me, a non-scientific study would be if I put a poll on this forum asking people based on their experience, what works better, 20L or 30L swarm traps. Now, that would be equivalent to these "meaningless" phone or text in polls that news agencies run all the time, which are truly "un-scientific' and TOTALLY without merit.