Wednesday 15 Aug 2018 | 18:59 | SYDNEY
Wednesday 15 Aug 2018 | 18:59 | SYDNEY

Capturing people in the polls


Fergus Hanson


24 October 2010 14:44

An article in The Economist earlier this month summarises one of the trickier problems facing American politicians and strategists (and increasingly those in other countries) — can you still trust opinion polls'

The root cause of the crisis of confidence in polling is how to reach the growing number of people without landlines (or Cellphone Onlys as they are called). This group is now estimated to represent around a quarter of Americans (it's a increasing problem in Australia too). It's very hard (and expensive) to track these people down, work out where they are from and, if you can reach them, convincing them to answer a survey.

Although The Economist did not cover it, this problem has driven some clever American innovation, which as far as I am aware, has not yet reached Australia. Knowledge Networks is a good example. They were contracted by the Chicago Council on Global Affairs for their latest poll and, as a quick look at the methodology indicates, a lot of thought went into overcoming this problem. Knowledge Networks has developed a panel to answer surveys online, but instead of allowing people to self select (which is how most internet polling is done, leading to bias), people are randomly selected and if they don't have the internet, they are provided with access.

The other upside of using this approach is the very good completion rate — 66% in the case of the Chicago Council poll. It will be interesting to see if the idea takes off elsewhere.    

Photo by Flickr user JonJon2k8, used under a Creative Commons license.