Given access to the data, or more ideally the ability to rapidly batch query, one could ask does being a google owned or rev-share site affect rankings? And if so, up or down? These are the questions statistics is made to answer.
edit -- and it doesn't even necessarily need be hardwired. For example, google could have more visibility into site interactions on google owned or tooled (GA) sites, and that could tend to lead to lower serps. Ie just the presence of certain features, only available to google when google owns or tools a site, could tend to produce lower ratings no matter what those features are. This is, I would argue, dirty pool for what is effectively a monopoly. And statistics can answer these questions.
Lower latency to the user should. If, ceteris paribus, google sites are lower ranked then yes, that should be illegal. ie consider regression and compute I[google] + site_speed; after controlling for site speed, being a google owned property shouldn't matter. It's fine if google owned sites tend to be faster and hence are more lowly ranked.
I specifically meant more along the lines of google can view your navigation through GA tooled sites, and there is a wealth of information that can be derived from that.