Another thing that destroys good science is FRAUD. It is estimated that a large percentage of scientific papers contain made up data. This is probably not surprising considering that most science is done by graduate students who depend on getting useful results to get their Ph.D. or Masters degrees. Graduate school is grueling, with long hours, years of often dead end work and low pay. The goal is to get your degree and move on to a University position, govt or industrial lab. Insurance against fraudulent data is reproducibility of your results by others. This is rarely attempted because there is no one free to do it. Everyone in the lab is slavishly working on their own projects. If a new student enters the lab to continue the work of one who graduates and cannot reproduce the results, it is often attributed to a learning curve or the new guy is “doing something wrong”. Sometimes the fraud is blatant. I worked in the lab of a top guy in his field. One day he comes in and shows us a paper that was published in an Indian Journal. It was identical to a paper he had published a decade earlier, only the names were changed. As far as I know nothing happened to the fraudster. This is reportedly pretty common in certain parts of the world. In another example I was on a team that flew to a guy’s lab that had published a paper reporting the successful development of a process we were interested in using. When we got to his lab and asked about it, he admitted that no one in his lab could reproduce the results, and that it was done by a grad student from Turkey who left the country after he gradated. He pretty much admitted the results were fake.
Research is harder to fake in an industrial lab where you are working with others to develop a product or process. The piece of the project you are working on is handed over to someone else and if it doesn’t work it’s instantly discovered.
Thank you for your personal examples of where you have seen science fraud. You are correct that this is pretty common. John Ioannidis, who is a Professor of Medicine at Stanford University, has documented that almost half of all research that is published in scientific journals can’t be replicated.
“.... Insurance against fraudulent data is reproducibility of your results by others. This is rarely attempted because there is no one free to do it. ....”
Or fund it!
Also, there’s no ‘reward’ in doing it! You will not get a graduate degree - PhD, etc. for doing ‘reproducibility work’ even though it’s absolutely vital for science to move forward. The current system produces ‘papers’ that are likely not meaningful and that’s about it.