The question of whether or not pollsters or predictors are skewing their findings to influence outcome of elections shouldn’t be the first question.
The first question should be: Is it possible to influence elections by altering their results?
If the answer is yes, then next question should be: What, if anything, is stopping them from attempting to influence elections?