Peer review is absolutely essential to any contemporary scientific undertaking. Peer review allows anonymous, unaffiliated scientists in the same area to look through the methodology of the study and determine if indeed all the proper controls were put in place, the statistics were done properly, and that the conclusions were supported by the data.
No human procedure will ever be perfect, of course, and peer review will almost never catch any very serious attempt at falsifying the data. For that to be caught, what is also needed is repeatability—that is, that each experiment be repeated by different, unaffiliated teams of people. Any serious result will almost inevitably be repeated.
The notion that peer review stifles true scientific discovery is a canard of pseudoscience. Any true phenomena of any interest at all to anyone will be published, none moreso than those which cause scientific revolutions. Of course, in the case that the experiment is particularly surprising, the peer review team may well (with reason) ask for more rigorous controls. This follows Carl Sagan’s famous aphorism that ‘extraordinary claims require extraordinary evidence’.