The first approach to testing these theories out is to test for SEO variables the same way a scientist would conduct research on the feeding habits of a chimpanzee in the wild for example. In order to conduct a proper test in this case, you'd want the chimp and its food in an isolated area where you'd no other variables could affect the chimp's behaviour. But if you do that, you've lost the key aspect to your question. The research is no longer being done in the wild. So all you can do is park yourself deep in the jungle, keep well hidden and observe. You should do the same thing with search engine algorithms. You will have to watch the search results to help you identify new trends that you might need to test.
You should use keen observation skills when analyzing search engine search results. Keep an eye on your search phrases of choice and pay attention when things move around. What did the sites that moved up have in common? What about those that dropped? What about top sites in more competitive sectors? What do they share in common? The answers to all of these questions will provide you with additional pieces of information that you'll use to build the hypotheses you're going to test.
As a search engine optimizer, you're going to have to use similar tactics as the above example to help you gather information beyond what people are discussing in public. But observation alone isn't going to give you all of the knowledge you need to keep up with changes. To acquire more information you'll have to move beyond basic observation and into the realm of testing.