I commend you on the work you have done here, but ...you made a crucial statistician mistake.
Your data collection process is flawed. I don't have the data to prove it, but I think it is pretty clear (as someone has already mentioned) that interaction does play a part in the design mission availability - i.e. what missions you are running/ran in the current cycle determines what missions are made available to you the next cycle. Since you didn't run any missions, it is improbable that your data is conclusive to prove your hypothesis.
Chances are really very very good that the decision matrix limits the number of missions for resources you "need" (based on popular choice) by triggering missions for resources you don't 'need' (your unpopular choice) beyond a simple ratio matrix. There are a few reasons why they would do this, all are pretty much a factor of limiting the resources any particular crafter has available to him'/her. This, effectually ...
...facilitates interaction with others. By limiting resources, a crafter can't (theoretically) be self-reliant past a certain point. This forces him/her to go to others/GTN to get the resources they need.
...slows the process. If a crafter has unlimited access to resources, it would make that craft skill trivial by design.
...makes those resources (and the items that can be made with them) have a value. This is more-so true with recipes, but affects raw resources as well.
And, to be clear, every mmo in existence with a crafting option has some facet of these implemented.
In the end, I think - at best - you may be able to conclude that the decision matrix re-initializes on instance reset - logging out/in, changing instances, etc - if you had a significant enough variation in your lists to prove it ...because that is essentially all you tested for.