Because the author took Clausen's consideration seriously, they were able to uncover their mistake and correct it.
Furthermore, since peer review is fallible, I don't believe Clausen's identical mistake making it through peer review proves the process any less important.
Perhaps Clausen's criticisms were not well intended, but wouldn't any respectable scientist demand rigor before seriously considering such a hypothesis? After all, the burden of proof is not on Clausen.
Had the data come from someone with a proven track record in the field, he may have found it worth his time to peer review it himself.
Appeal to authority is not always a fallacy. It would have been fallacious if he was outright denying or accepting a hypothesis as being the logical conclusion based on one's standing. All that I read from it, however, was that he had a hard time believing it, and would not spend the time to look into the data personally, but that it should be reviewed.
I don't think it's illogical to take more seriously the hypothesis of someone with a proven track record of good research.
Are you describing logic/science/debate as you believe it to be or as you believe it should be? If the latter, are you suggesting ending peer review?
http://www.researchgate.net/profile/Mike_Oaksford/publicatio...
http://www.pfeifer-research.de/pdf/pfeifer08rh.pdf
I can see how Clausen comes off as snotty. But his response is pretty neutral, even if he praises peer review as a filter beyond its actual merits. He sees a lack of appropriate effort to establish the claim from a nobody, why should he not be skeptical of such claims in general? It's the same heuristic people like Aaronson make on being highly skeptical of the latest P==NP or P!=NP proof uploaded to arxiv.
The valuable thing isn't individual peer-reviewed papers, it's bodies of peer-reviewed works.
"Peer-review" as most people think of it -- the vetting of papers by a few reviewers before publication -- is the start of the process, not the end. Peer-review is what happens after your paper is published and your peers -- the people who read the papers you publish -- read your paper and discuss your paper and try and succeed or fail to replicate your results. A journal isn't "peer-reviewed" because a couple of schmucks read the paper before publication; that's just the scientific equivalent of Fizz-Buzz to avoid wasting everyone's time with obvious crap. It's "peer-reviewed" because it is widely and critically read by a large number of practitioners in a field.
You should never regard an individual paper uncritically just because it is published. Later, in light of a larger body of research, you might regard a paper as particularly insightful or seminal.
To generalize a little more: science isn't about single pieces of data or single experiments. Science is about the process. Experiment/data -> idea -> experiment/data -> refined idea -> ... until the idea and the data match to the greatest possible degree (a temporary state of affairs, until science in other areas makes it possible to refine measurements even more tightly.)
A single paper, even one with solid methodology, extensive data, and clear patterns, is just one part of science. Replication, refinement, methodological improvements, identification of weaknesses, etc. are what make the whole institution of science so powerful.
He deduced that this "strongly suggested" they were not toxic. He was wrong because the plants contain a non-alkaloid compound that is toxic to mammals (and presumably not bacteria). So there is a lesson here about what kind of conclusions you can draw from scientific data.
Krakauer's excerpt of Cole's article was poorly done. If you go to the source article, Clausen says,
“First, it seems like a rather remarkably lucky shot in the dark for a person to suspect the presence of a specific toxin on so little evidence," he wrote. "Couple this with the observation that ODAP is only reported in certain plants of the genus lathyrus which is not the genus hedysarum and the conclusions are even more remarkable (though not impossible). I would be much more convinced if I was reading the report from a credible peer-reviewed professional journal such as the Journal of Natural Products or equivalent. I have often heard a key indicator of bad science that is highly suspect is having its results published in the popular media."
...
"Again, I don't make any claims that the report is wrong since I have no data to analyze, but I am skeptical and will remain so until I see a better forum for the results to be published in." [1]
which to me sounds reasonable and, as a professional scientist, sounds just like what I (and I expect, my colleagues) would say. The peer review process isn't perfect, but it is consistently superior to the crap that is scientific journalism appearing outside of peer reviewed journals. Here's why, and why I don't understand why you would hold Krakauer as an example of citizen science.
Krakauer made (by his own admission) a "rash intuitive leap" [2] that an unidentified alkaloid contributed to McCandless' death, based on one of McCandless' journal entries and unpublished preliminary research by Clausen. As an aside, if one wants to piss off a scientist, that's a good way to go about it. When Clausen completes his research and finds the mystery alkaloid theory doesn't work, Krakauer intuits that a mystery toxic mold must have been involved, a theory he announces on Oprah and NPR. This second theory doesn't hold up (as was demonstrated by Clausen, again). Despite your conjecture, it sounds like Clausen actually investigated Krakauer's theories, up to a point. Krakauer then latches on to ODAP, which was close but still not there. His fourth theory may be correct, but Krakauer disseminated a lot of incorrect information along the way. I wonder, did he publish retractions as a professional scientist would?
I'm not sure why you would hold Krakauer's methodology as "citizen science FTW", but I'm interested to hear.
[1]: http://www.adn.com/article/20130917/krakauers-wild-theory-mc... [2]: http://www.newyorker.com/books/page-turner/how-chris-mccandl...