While this BBCnews report about how earth-like planets may be much rarer than originally thought at first seems disheartening (well, to those of us who'd rather humanity not be the pinnacle of cosmological evolution), there's a fundamental flaw in the study, that of a "sampling error".

We've only detected weird, "hot Jupiter" systems because that's all we can detect. Or, rather, all we could detect when the paper was written (I think astronomers are getting better at detecting other kinds nowadays). Detection of these sorts of solar systems is performed by watching a star "wobble" over time. A solar system containing a close, massive planet spinning like a circular saw blade around its sun is much easier to detect than a solar system without one. The sun in a "hot Jupiter" system, when compared with one without, wobbles like a weeble in an earthquake.

So, while it is true to say "the planetary systems scientists have found so far are very unlikely to harbor earth-like planets", it does not follow that "earth-like planets are therefore extremely rare in the universe".

It's such an obvious flaw that we must be looking at poor editorial practices on behalf of BBCnews rather than a fundamental problem with the paper itself. As always, absence of evidence does not equate to evidence of absence, and the peer review process surely would've caught this. On the face of it, looks like BBCnews engaged in a bit of press-release-regurgitation, something I didn't expect from them.