It is impossible to have truly "objective" polling. There are the questions, there are the answers, and you assess within that context. We have standards and practices that help steer us towards higher quality input/output knowing that nothing is airtight.
This is a nonsense standard of your own making. News, history, doesn't matter. "Just the facts" and "true objectivity" are a noble dream at best (to borrow from Novick) and are unachievable. You do the best you can, account for biases/different narratives, and deliver the results. That's always how it has been. It doesn't make these fields totally arbitrary or worthless as you imply.