Is there more of it than you think? This article seems to think so.
The data presented in the article were restricted to the US health and biological science, so it simply "extrapolated" into the rest of science. Now, even if this is true, that there are "... three incidences of fraud for every 100 researchers...", one needs to evaluate two important things as far as physics is concerned:
(1) were these part of important, high-profiled work and publications?
(2) did it get through to the stage where it was being used as a valid source?
The two instances mentioned in the article (the Schon and Ninov debacles) certainly fit #1, but this is out of how many high-profiled work? The thing we need to keep in mind here is that these high-profiled work, published in major journals, are usually highly scrutinized. So one certainly cannot get away with a lot of things, much less, fraudulent results, since someone is bound to try and reproduce your results. This is why it is still a major puzzlement why Schon did what he did, as if he didn't think someone was going to try and reproduce his results. So these important works in physics usually are not affected by frauds, because it is just way too difficult to get away with it.
Now, certainly such things are less scrutinized in lesser journals. But most of these very seldom become significant work that we rely on. And this brings us to #2. How many of these frauds actually made it "out" and into the stages where they were applied and used? I don't know of any. Why? Because if it is fraudulent, it would not work, no matter how much you try! That was the problem with the Schon results, no one could reproduce it! It could not get out of the scrutiny part and into the applied/application part.
The point here is that fraudulent work in physics hasn't made it out too far for it to mislead or fool enough people. (I'm distinguishing this from outright crackpottery such as the "hydrino".) If it doesn't work, it doesn't work, no matter how one tries to hide the fraud. Science, and certainly physics, is still a slave to Mother Nature.
Because the PRACTICE of science is a human endeavor, it is why we have peer-reviewed process, and then further scrutiny by others in the field. If not, why bother with all of the circus? The public and the media should be made aware of this, and this is why any results or discoveries should be given the proper "gestation period" for the system to work. The fact that something has been published isn't really the end of the story, but rather the beginning of a detailed, careful scrutiny. It is when the public and the media forget about this is when they are "misled".