An opinion poll of scientists who work in a relevant field is informative.
These are earth scientists, so this is not a Fallacious Appeal to Authority fallacy.
However, to pin down the validity of the poll we would need more information than is presented in this article. The most important detail we need may be difficult to establish. It is this: To what extent is this field infected with academic politics and the suppression of dissent? Is it likely that people who actually disagree are going along with the majority because they fear losing their jobs otherwise?
I have no idea whether global warming research suffers from a debilitating degree of academic politics or not. All fields have some degree of politics. Indeed, all primate interactions do.
Cold fusion suffers from the worse case of debilitating academic politics in the history of science. A random poll of "scientists" would tell you nothing about cold fusion, but a great deal about the limitations and primate nature of researchers. As I just told a correspondent:
"The only valid poll [of opinions about cold fusion] would set ground rules such as: the respondent must work in some relevant field, and he or she must have read at least 5 papers on cold fusion. With that group of respondents, I am confident that the majority would agree that cold fusion is real. If you limit the poll to people who have read 50 or more papers, then I expect you will find nearly all agree the phenomenon is real."
- Jed
