The percentages come from looking at all studies, papers, research, etc. and determining the number one one side or the /i?
When the administrators of research funding withhold future grants from scientists who publish papers questioning some aspect of the current global warming scenario, while giving additional funding to scientists who publish papers supporting it (or claiming some global-warming tie-in to whatever phenomenon they're examining), the count becomes skewed. This is political action, not science.
This happened in the '70s with research into medical effects of the popular "recreational" drugs - before such research was effectively banned. Among the resuts were a plethora of papers where the conclusions obviously didn't match the data presented and a two-decade delay in the discovery of medical effects and development of treatments. Only NOW are we finding evidence that PTSD might be aborted by adequate opate dosages in the weeks immediately following the injury, or that compounds in marijuana may be a specific treatment for it - as they are for some forms of epilepsy and may be for some cancers, late stage parkinsons, and so on.
The same happens when the editors of a journal and their selection of reviewers systematically approve and publish only research supporting the current paradigms, to the point that scientists with contrary resuts must find, or create, other journals or distribution channels (which can then be smeared as non-authoritaive, creations of the fossil fuel industry, right-wing politicans, or conspiracy nuts - and their articles LEFT OUT OF THE COUNT). Again, this is politics, not science.
Then there's the question of the methodology of the count itself. What is counted as "support for" versus "opposition to"? What does it count as a scientific paper? Were well-established research methods used? Was it reviewed? By whom? Was it done by scientists with no established position on the issue, by scientists supporting one side, by pollsters, by an advocacy group, by politicians? (Hell, was it done at all? Truth is the first casualty of politics, and fake polls are one of the commonest murder weapons.)
For an instance: How would you interpret the study behind the Scientific American article that seems to indicate:
- Planetary temperatures have tightly tracked a function of three orbital-mechanics effects on the earth's orbit and axial orientation - up to the time of human domestication of fire.
- That occurred as the function was just starting to inflect downward into the next ice age.
- The deviation amounted to holding the temperature stable as the function slowly curved downward. (Perhaps a feedback effect - more fires needed for comfort in colder winters?)
- This essentially flat temperature held up to the industrial revolution, when the temperature began to curve upward, overcoming the gradually steepening decline of the function.
- If this deviation is the result of burning fossil fuels, they are expected to run out in about 800 years - after which the temperature might crash toward the "Ice age already in progress" as the excess carbon is removed from the atomsphere by various processes, or simply be overwhelmed by the orbita