Statistically speaking, the anthropomorphic shape is probably not that rare. Sure, there are sometimes many ways of accomplishing something biologically. However, the precurser features of an organism have a direct path of development. This tends to eliminate alternate possibilities. Additionally, it has been found that any differences between organisms tends to be enviromentally and efficiency based. The case for a high frequency of analogous structures (different features, but with the same function) seems to be diminishing. It's rare and probably statistically less likely, even on intersteller scales. Those who have thought otherwise tend to eventually find they weren't looking hard enough for the advantage of that particular feature difference.