0.4% complete a survey. Is that 0.4% representative of the 99.6% who don't have the time or willingness to respond to surveys?

Survey polling has been in question ever since the famous failed forecast that Dewey would beat Truman in the election of 1948.  

Click to enlarge.

The popular explanation for the failed forecast was that polling at that time relied upon calling people on their home phones.  Since phones were expensive and many homes did not have them, such a strategy necessarily overweighted high income individuals in the response pool and the wealthy disproportionately voted Republican (Dewey).  

The problems in sampling a pool of likely voters have grown legion over time.  For the poll to be meaningful, it is absolutely critical that it be random.  And that is a pretty steep hill to climb.  In fact, virtually impossible.  All sorts of post sampling adjustments are made to try and ensure that the responses received match an anticipated profile of the electorate (age, income, race, location, party affiliation, etc..)  

It is not only a very difficult process to ensure that those polled match an expected profile of the electorate.  It is also very expensive.  And the more adjustments are made, the less random is the sample.

From Who in the World Is Still Answering Pollsters’ Phone Calls? by Nate Cohn.  The subheading is Response rates suggest the “death of telephone polling” is getting closer.  Note, this is not Nate Silver of Signal and Noise fame.

There were several figures which struck me.

How do you account for the fact that few people answer? Before I respond, I want to dwell on just how few people are answering. In the poll we have in the field right now, only 0.4 percent of dials have yielded a completed interview. If you were employed as one of our interviewers at a call center, you would have to dial numbers for two hours to get a single completed interview.

No, it wasn’t nearly this bad six, four or even two years ago. You can see for yourself that around 1.6 percent of dials yielded a completed interview in our 2018 polling.

The Times has more resources than most organizations, but this is getting pretty close to “death of telephone polling” numbers. You start wondering how much more expensive it would be to try even ridiculous options like old-fashioned door-to-door, face-to-face, in-person interviews.

0.4%?  If less than half a percent of contacted people are willing to complete a survey, how on earth that that tiny percent of survey takers are in any meaningful way representative of the population at large?

An interesting piece about some of the nitty gritty elements of surveying.

Nothing in it gives me confidence that the people being contacted are in any meaningful way representative of the electorate at large or the electorate that votes.  And all the adjustments that are made to try and correct the lack of representativeness are as likely as not to introduce even more errors.

Polls are good for juicing interest in a campaign but they are less and less likely, individually, to provide a usefully meaningful result in terms of the actual election outcome.