Cell phones and polling

By Richard K. Barry

During the campaign, a lot of people were making the point that robocalls didn't call cell phone numbers and therefore were not sampling the kind of people who use cell phones exclusively. Those who were being missed were part of a younger, urban demographic. In other words, they were the kind of people more likely to vote Democratic.

For example, Rasmussen doesn't call cell numbers but say they use an online panel to supplement their polling and presumably capture cell phone users. Typically Rasmussen shows results more favorable to Republican candidates, which, as noted above, hardly seems surprising if they are not calling cell numbers.

Polling guru Nate Silver was on Morning Joe last week talking about this cell phone problem. Again, this isn't news, but one would have to wonder why any polling company, assuming it wants to be accurate, would not modify its methodology to do a better job of reaching cell phone users now and in the future.

I know it's getting tedious, but it does make me ask if we should pay any attention to polls that don't make an effort to call cell phone users at a rate consistent with average national use. Is there anything about this that's hard to understand?