Pew Research is a biased source because CGI Script polling poses an existential threat to their business model.
Pew Research (and pretty much any other credible research institution or academic) are domain experts, understand sampling methodology, and the very-well-known biases which occur from biased and/or self-selecting samples.
Public polling can be exceedingly accurate based on surprisingly small samples (roughly 300 in most national political polls, for example, and even that is generous), so long as the sample is in fact truly random. Oh, and that the responses are not similarly filtered.
One of the most famous cases of a poll which failed due to sampling bias was in the 1948 US Presidential election, and resulted in the Chicago Tribune erroneous headline, "Dewey Defeats Truman". This was drilled hard in my own stats education several decades ago and remains a sharp lesson in the risks of biased sampling. For a good descrition of the error, see:
https://web.archive.org/web/20180823071102/https://textbook....
Pew was simply the first well-articulated reference I turned up. There are numerous others. You've offered nothing other than your uninformed opinion and casual insults and accusations. Please don't do that.