What is the purpose of "peer-review" in science?

To check method and data, and thus weed out mistakes, overstatement and fraud. Unfortunately, evidence says peer review also arbitrarily weeds out legitimate science because the hypothesis and or conclusion don't match the paradigm(s) of the reviewers. (1) Sometimes the reverse is also true, were a journal will let bad science through simply because it matches the beliefs of the reviewers. In fact, peer review is not much better at this point than a coin flip. I agree with the need for peer review, so the current state of it needs to be addressed somehow.

Peer review helps ensure that sound scientific methods are used and reasonable conclusions based on available data are made in a paper. Generally, the editor of a scientific journal will send all submitted articles out to three different scientists that work in a similar or related field, for review. Sometimes papers get turned down completely, often they get sent back for some modification or even a few more experiments, and sometimes they are accepted as is.

Sometimes the reviewers will even run a few experiments themselves, especially in the case of some really unexpected results (like the cold fusion fiasco - many labs tried to repeat the experiments and figure out what was going on).

The general public almost never reads the original, peer-reviewed journals.

The purpose of peer review is to have another, fresh set of eyes to review data or findings to improve the reliability of a particular conclusion.

Sometimes there is something unique in the sample set up, processing, or analysis that would lead one individual to come to a different set of conclusions than another. The next step is to the reconcile the different sets.

In my industry, I deal with hydrogeologist. There is a generally accepted, peer-reviewed model which predicts how groundwater moves through the subsurface. That, however, is based on a large number of assumptions from rate of transport to boundary conditions. While they are plugging numbers into an accepted model, the peer-review must be in the assumptions and calibration of the model to experimental data (pump tests).

The same must be true of climate change models. This is the reason why the theory of "global cooling" observed in the 1970's never gained any steam, because peer review didn't validate it. Although, you talk to those with a pre-adjenda against climate change, and they incorrectly equate and trivialize the widely accepted and peer reviewed analysis of global warming with a failed, rejected theory of global cooling. It's like saying "cold fusion" is impossible just because one set of scientists had flawed data.

This is one of the impacts of a failed scientific education for most American students.

to submit the theory to others who are in a position to verify it, or disprove it. Although having been verified by a peer review does not necessarily mean that it is incontrovertable, only that the information used and the method of determination were sound. It does not provide a means for determining if all the available data was used (which leads to skewed conclusions) whether the data was accurate (like much of global warming data wasn't due to improper placement of thermal equipment) or if the theorists has an underlying agenda.

To keep everybody honest by allowing qualified people to check your work.

That having been said, you can't honestly assert that politics doesn't play a role in scientific research...

Scientists depend on grants and government funding to pay for their research.

They're competing with other scientists for a limited amount of funds.

And politicians are more likely to fund projects that deal with "fashionable" areas of research, or that are going to reach conclusions that will support and give justification for their own political agendas.

And so researchers deliberately choose or direct their research projects in ways that are most likely to get that funding and follow the political winds of trendy and "junk" science, often as not. And if they fudge the results or slant their interpretation of the data to favor those trends, they're more likely to get more funding in the future to expand upon that previous research.

Okay Cap'nT, I again find this a very interesting question and I have another point of view to spew forth. Peer review is only good when the people who fund their projects do not have input. Such as most of the scientific information that is funded by the government, leaves me scratching my head and asking myself if the information they provide is not slanted in the directions of the current political regine.

This is either the most intellectually stimulating question ever posted on Y! or the largest, stinkiest pile of BS ever to be squeezed out of a bull's âss.

Push 1 for answer #1, 2 for answer #2.

Peer review is supposed to ensure that a study's findings are accurate and/or reliable. Unfortunately, in practice peer review is as reliable a method as it would first appear.

Global-warming believer.

Great answer, Intelex.

Bad Robot nailed it and deserves best answer. It's a meritocracy, not a democracy. Submit flawed research and the peer review teams kick it out.

Peer Review exists in the medical field too...to review deaths of patients and such - to rule out foul play, etc.