It’s awful.

Look at this:

Thanks the Economist
Thanks the Economist

For ages, there was a massive discrepancy between telephone polls and online polls. And it’s part of the problem with newspapers reporting on how the polls are shifting, because what are they actually talking about though?

They’re basically taking the average responses of two groups of people:

  1. People with landlines who still answer their landlines; and
  2. People who actually participate in online polls.

And would we call either of those groups especially “normal” or at all representative of the average voter?

Notes from the Guardian:

Twenty years ago, telephone polling companies would draw 7,000 random residential telephone numbers; these would yield 2,000 completed interviews. Now they must draw 28,000 numbers. Response rates have collapsed from 30% to 7%.

The problem for online companies is different; but, by definition, they can poll only those willing to join online panels. Pollsters therefore have to extrapolate from the people they can reach to the increasing numbers they can’t reach.

At this point, all that anyone can actually say is the odds are 50:50. But this is a yes or no vote, so the safest place to be on this was always going to be “50:50”.

The Telegraph this morning:

Screen Shot 2016-06-22 at 12.48.31 PMBloomberg this morning:

Screen Shot 2016-06-22 at 12.49.52 PM

The FT this morning:

Screen Shot 2016-06-22 at 1.00.47 PM

The Guardian again:

Screen Shot 2016-06-22 at 1.07.33 PM

So here we sit, a day away, and it’s all on a knife-edge.


Happy Wednesday.

Rolling Alpha posts about finance, economics, and sometimes stuff that is only quite loosely related. Follow me on Twitter @RollingAlpha, or like my page on Facebook at Or both.