First design
Site theme
The next segment is taken from a bankruptcy that deals with the relentless pressure to publish that scientists face and the corresponding explosion of fraud that this strain creates. Fraud can take many forms, from “serious fraud” of pure falsification of knowledge, to many types of “soft fraud” that comes with plagiarism, manipulation of knowledge, and conscientious choice of strategies to achieve the desired result. The more fraud grows, the more the public loses acceptance as true in science. Solving this challenge requires a basic change in the incentive and praise structures that scientists work on. A difficult task, no doubt, but not very unlikely, and one that I will strongly appreciate.
Modern science is hard, complex, and built on titles and many years of hard work. And the science in vogue, almost everywhere, is based on calculation. With the exception of a few (and I mean very few) hardcore theorists who insist on writing things down with pen and paper, there is an almost absolute guarantee that for every paper in any clinical box you can read, a computer was involved in a step of the process.
From reading bird droppings to galaxy collisions, modern science owes its lifestyle (and its patience) to the computer. From a computer sitting on an ignored table to a giant device filling a room, “S. Transistor” is expected to co-author nearly all of the 3 million journal articles published each year.
The sheer complexity of science and its reliance on traditional software renders one of the front-line defenses against hard-pressed fraud useless. This defense is peer review.
Worse, many software codes used in science are not publicly available. I repeat this because it is a bit far-fetched to imagine: there are millions of articles published every year that depend on computer software to download the results, and that software is not within reach of that purpose. Other scientists can read up on this to see if it’s legitimate. Or not. We simply have to accept it as true, but the word “accept as true” is way down the list of scientists’ priorities.
Why don’t scientists make their code available?It all boils down to the same explanation for why scientists don’t do much that can be limited to the clinical process: there’s no incentive. In this case, you don’t get any h-index issues for publishing your code on a website. You only get to publish articles.
This bothers me greatly when I write peer-reviewed articles. How should I judge the accuracy of an article if I can’t see the entire process? What’s the point of searching for frauds when the PC code of the published result can be molded to give the desired result and no one will notice?
I’m not even talking about intentional computer fraud here; It’s even challenging to stumble upon basic errors. If you make a mistake in an article, a referee or editor can spot it. And science is better prepared for it. If you make a mistake in your code. . . Who reviews it? Whenever the effects look good, it will publish them and the reviewer will settle for them. And science is faring even worse.
The science becomes more and more complex over time and relies more and more on cushyware code to make the engine work. This makes fraud of difficult and comfortable types less difficult to achieve. From the use of sophisticated equipment that you perceive slightly but use to obtain the desired result, to simply faking it, science is getting more and more done.
Join the Ars Orbital Transmission email to receive weekly updates in your inbox. Register →