Scientific misconduct happens. Simplifying Sheila Widnell’s definition, scientific misconduct is the deliberate misrepresentation of scientific research. This can be, from most to least blatant, the construction of false data, editing data to achieve desired results, or presenting others’ discoveries as one’s own. Another type is the conflict of interests that financial holdings and the like can create for a scientist, who may consciously or unconsciously promote those interests over the interests of science. The first two types carry a potential for widespread harm, and the last two are more than simply unjust. Given that scientific misconduct is generally about deception, it is hard to be certain how much it actually occurs. The community of science is structured in a way that encourages a defensive cloistering of scientists to themselves in the face of such accusations. Solutions can be found that protect the integrity of the scientific process without undue risk to scientists.
Scientific misconduct has the potential to produce disastrous results throughout the world, both directly and indirectly, and should not be taken lightly. The scientific community works in large part to facilitate the wide diffusion of the results of the scientific process. So any deliterious impact of scientific misconduct can be rapidly spread around the world. And there are serious concerns about the consequences of scientific misconduct. One can start with the attempts to replicate the results of the experiment. This can tie up the resources of many labs in a wild goose chase, wasting both time and money that could be spent on better things. When taking a more technological bent, the conflict of interests and plaigarism can both result in misapplied resources. An undetected, or unchastised plaigarist may be the party who gets funding rather than the colleague from whom the plaigarist took credit. If past deeds are any predictor of future outcomes, this would be a great mistake. Additionally, the entanglement of conflicting interests may lead a scientist to proclaim virtues that do not exist, or deny harms that do. This is most readily apparent in the approval of drugs, but it has relevance outside this narrow view.
In light of the potential damages, it is important to understand how these misbehaviors can arise, and why members of the community might turn a blind eye. Within the scientific community, members tend to specialize into particular narrow fields. Scientific communication tends to occur mostly within these smaller sub-cultures, if for no other reason, then because the sheer volume and technical detail of information to understand makes universal expertise impractical. Thus, those in the best position to be aware of potential misconduct would be individuals most closely identified with the perpetrator. Within these smaller communities, there is a concern for the outside appearance of the field, particularly to funding sources. As funding comes from outside investment, typically government sources, which may fund projects in a different field instead, there is a strong incentive to foster an image of legitimacy for the field. Individuals making accusations of scientific misconduct may seem to threaten that legitimacy, and thus the funding that the science requires.
To maintain the legitimacy of various domains of science, it is important that effective scrutiny take place. While our current system is perhaps overly insular it is quite possible for us to overcompensate. The quality of the science itself is best left to the judgement of experts in the domain, that is science at work, and science generally does work. However, questions of basic scientific misconduct operate by a different set of principles. It is most appropriate for outsiders, who lack the conflict of interests with the field’s advancement to assess the legitimacy of the science. To that end, making data more publicly available, allows for broader scrutiny. Clear, objective, community guidelines also make standards clearer for conflicts of interest. These should be negotiated by scientists, with an eye to protecting their public legitimacy.
The community of science is structured in a way that encourages a defensive cloistering of scientists to themselves in the face of such accusations.
I disagree with this. The reason most misconduct among scientists is discovered is because of other scientists. Empiricism makes it tough to trade lies for long. If something doesn’t happen, it just doesn’t happen, and someone trying to build upon faulty research isn’t going to be able to make it work.
I think that statements like this, which portray scientists as conspiratorial, back-slapping, good ol’boys looking to deceive the public at the expense of truth is extremely damaging to scientists and unjustified.
It is most appropriate for outsiders, who lack the conflict of interests with the field’s advancement to assess the legitimacy of the science. To that end, making data more publicly available, allows for broader scrutiny.
Vast amounts of data are available to the public. In fact, with the Internet, more data is more easily available than ever before. I don’t see that this necessarily helps the public discern between good and bad science, though. How many people feel comfortable interpretting an infra-red absorbtion spectrum? The output of a gas chromatograph? Critiquing a Monte Carlo model? Interpretting the signal on an oscilloscope? Ideally a responsible citizen of a democracy would be capable of doing that intelligently, however, the truth is that most people are incapable of solving a differential equation, and certainly aren’t capable of discerning whether a given set of data obeys that.
I think for that reason, people are suspicious of science and intellectuals in general, because it’s easy to hoodwink them and they know it. That’s no reason to portray scientists as deliberately deceptive. To do that, I think, is a part of the real problem.
I’d been looking for “empiricism” earlier, but couldn’t remember it, thanks.
I did note in there that for the most part, the scientific process works. What I’m talking about is doctored data. This is based, in part, on an article in the AAAS 1991 Science and Technology Policy Yearbook. It had a number of interesting statistics, which stated that most scientists thought they had seen examples of scientific misconduct, but most who had, did not point it out. That and the other stories we were told suggested that there is an old boys network, or, rather, several smaller old boys networks. And for the most part, they’re doing their job. But when accusations of misconduct start flying, the authors claim it reacts much more like a family dealing with alcoholism. Allow me to repeat, with emphasis, misconduct is a rare event.
When I say outside observers, I’m not talking about joe schmoe off the street. First, I’m advocating greater public sharing of data, for instance journals requiring data with submissions. It gives them an opportunity to check for themselves before publishing claims. And in terms of investigations, should it come to that, I’m talking about looking for massaged or constructed bogus data, inconsistent dates, that sort of thing. As such, a statistician and perhaps a police trained investigator would be more in line with what I was thinking of.
Hmmmm…weird sort of link to my work. I work at ICPSR, an archive of quantitative data. I.e., researchers stash their data with us so that other researchers can download and replicate their results…and because long-term preservation of data is such a pain in the ass.
As for misconduct being a rare event…well…(not speaking for work…I just publish Web pages)…take a peek at any of the research that ‘proves’ that obesity equals early mortality. It’s amazing how a study of 200 white nurses can be applies to the general populace…or a how a CDC staffer making a cut and paste error in Excel can change Medicare policy.
The trick, of course, is that lay people can’t evaluate a lot of the research…and the researchers themselves are reluctant to aggravate potential sources of funding (i.e., the CDC).
The thing is, bad science isn’t necessarily the result of misconduct. Biased or inadequate samples aren’t the same thing as falsified or “adjusted” data. Now, if they fail to note their sample, and its shortcomings, that gets into another question. =)