If you look at why papers are found out as fraudulent and retracted it's usually very dumb mistakes, such as the examples from the article, copying data and/or

If you look at why papers are found out as fraudulent and retracted it's usually very dumb mistakes, such as the examples from the article, copying data and/or text from elsewhere, making up numbers that are obvously implausible, and cloning/photoshopping figures or parts thereof, even within the paper. Given how easy it is to avoid these beginner mistakes the percentage of fake data must be so much higher than the actual retractions. Especially from the paper mills where faking data is done by professional fraudsters.