Okay, you’re talking about a single metastudy (a study of studies if you will) that was commissioned by groups that were certain that the the theory that human-caused global warming was wrong and hired a well known skeptical and legitimate climate scientist to show that the scientists in support of the theory were cherry-picking studies that supported this conclusion.
So, essentially the task of the study was to look at about 2,000 recently published papers in the field of climate science and determine if their conclusions were consistent with the theory that global warming was caused by human activity. The scientist running the study expected only a fraction of the papers would.
But that’s not what he found. After spending a few years reading studies and their conclusions, he found that 97% of the published literature was consistent with human-caused global warming. That was overwhelming.
As for the other 3%, all of them had their own idea of what had led to the results they found. None of them were consistent with human-caused global warming, but each one had their own idea of what their study supported. None of the remaining 3% supported each other at all - if one of those dissenting studies had been right, every other study in the metastudy had to be wrong, because none of the others (even the dissenting ones) supported its conclusion.
It gets worse. A lot of scientists took a closer look at those 3% of studies and found most of them were fundamentally flawed in their methodology. In other words, the study’s author was trying to prove something and found evidence that supported it, but tossed out the evidence that didn’t. Some were just sloppy, others simply came to a conclusion that was supported by the data, but turned out to be wrong.
The metastudy was so overwhelming that the skeptic who prepared it admitted he was wrong about his skepticism and agreed that the underlying theory was right, although it might need some adjustment.