All Brian and I are saying is that what happened on the test panel is different due to one of the many variables you can experience when painting. One thing that I think you can rule out is silicone being the problem. If silicone was that concentrated that it wouldn't allow paint to stick, you would have had more fish eyes in the surrounding area than you could imagine. Another being sanding dust...can't see that being the issue for many reasons.
I've been painting for many years (over 30 years) and trying to solve problems for techs in body shops for nearly as long. Every once in a while you do run into a situation that defies all reasonable explanation for the information you are given. There is an explanation out there, what it is I can't give you a 100%, yep this is it answer. I ran into a situation similar to yours about 18 years ago, beat our heads into the wall for weeks, couldn't figure it out. A few weeks go by and an apprentice took me aside and asked if opening and closing the bay door while somebody was painting could be a problem, (Cold Winter day, no air make up unit to control booth temperature). We recreated this situation with a 2 minute door open. Temperature inside the booth dropped by 45 degrees and took over 20 minutes to recover. This shocked the paint but it seemed to adhere...for about 10 days. We pressure washed the test panel and the paint blew off. Will this happen every time? I don't know, as in your case, to many variables. If everything was exact in this trade, we wouldn't need forums like this, every paint job would be perfect and all would be well. It's these things that make it all interesting and enhance the learning experience.