Ignoring the fact that I had already tackled this point in one of my previous entry under "Imagination without knowledge is ignorance waiting to happen", I will try to address this issue by invoking a terrific essay by Daniel Koshland that appears a while back in Nature. I'm doing this because I had to make references to this article several times already. I want to make a clear and concise review of what he wrote so that next time another argument on "peer-reviewed journals are stuck-up", I'll either point to this entry, or do a quick cut-and-paste.
Here's a bit of a background. Dan Koshland is a biochemist. He was an assistant scientist at Brookhaven Lab in the 1950's and worked in the study of muscle enzyme. What he discovered and concluded was that the prevailing, accepted theory at that time that was the work of Emil Fisher was wrong. Koshland's "fit-induced" theory was going up against a well-known "lock-key" theory of a very famous and respected biochemist.
He decided to try to get it published, but with a lot of trepidations...
"Realizing that a young scientist was largely identified as promising on the basis of one or two good papers - and could be largely destroyed on the basis of one or two bad papers - I was aware of the risks I was taking"
Naturally, the journals that he submitted his work to were skeptical. The type of comments he received were along the line of
"The theory of Emil Fischer has been a cornerstone for 100 years and will not be overturned by the ideas of an obscure young biochemist from a young national laboratory."
He described his angry reaction to the journals that rejected his papers, which is understandable. But here is where he made a point in which the claim of many quacks that science only care about upholding the current idea gets totally smashed:
"I realize now that a new theory is likely to meet resistance, but it should, if based on GOOD experiments, receive sceptical encouragement if science is to remain in balance. Non-confirmists are necessary for progress in science, just as mutations are necessary for progress in evolution. However, there must be constraints to select good mutations from bad mutations. Too many mutations block evolution, as error-prone straints of bacteria have proved. So non-conformist thinking in science must be encouraged to make progress, but restrained to prevent anarchy. In science, it is PEER-REVIEWED JOURNALS and granting agencies that provide such balance."
Before we forget, note that this came from a "non-conformist" himself who had to struggle to get his idea published and accepted. I find that I put more validity and credence to an opinion of someone who HAS gone through what he or she is writing about, rather than from someone who just READ about it.
Koshland went on to say...
"The trouble is that journals can easily become too conservative, because editors find it easier to reject the unusual than to take a chance on the unthinkable...... The existence of multiple journals provides the final safeguard against too much conservatism and is the ultimate reason that science is more receptive to non-conformity than any other segment of our society."
The one good thing about science, and experimental science in particular, is that if it is based on valid evidence, SOMEONE is bound to eventually verify it. So his theory did get published, and is now a standard explanation in textbooks of biochemistry.
I think that last part that I quoted is something many people outside of science do not understand or are aware of. And in this aspect, I will focus on physics in particular. While the most prestigious publications for physics are Nature, Science, and Physical Review Letters, there are MANY other physics journals that are equally important. There are also many more "lower-tier" journals in which even "dubious" papers are published. The fact that notorious papers such as the Fleishman and Pon's "cold fusion" paper and the Podkletnov's "antigravity" paper could appear in peer-reivewed journals are CLEAR evidence of this multi-tier system would allow even such unorthordox ideas to appear in print.
So it then begs the question that one would want to ask a quack: "If Fleishman and Pons paper could appear in a peer-reviewed journal, why can't yours?" Could it be THAT bad? After all, there is no longer the excuse that peer-reviewed journals would ONLY publish things that only agree with the current idea.
 D.E. Koshland, Jr., Nature v.432, p.447 (2004).