Back

A reader asks: In a time when people have so many tools to limit unwanted and unfamiliar interaction, how can pollsters be sure they’re reporting results from a random, representative sample?

Jennifer Rubin: "Important point. Response rates have plummeted, and #polls now require a mammoth effort even to come up with a sample of a few hundred voters. This phenomenon adds to the difficulty in devising an accurate model to capture expected turnout. Only the most professional and well-funded pollsters do a more credible job in producing a representative sample. It is one of the reasons we should be highly skeptical of #polling from entities without an established track record. Overall, too much weight — and too much political commentary — is devoted to polls of questionable accuracy and even less predictive value."