Posted to Steve Outing 28 May 1999
You (Steve Outing) write in today's column :
Most online opinion polls are non-scientific, since they typically make no effort to control who are the respondents. The polling industry has coined an acronym that applies to most Internet polls as practiced today: "SLOP," or Self-selected Listener Opinion Polls.I agree with your conclusion that online polls are inaccurate, but I reject your suggestion that this represents a trend for the worse in polling.
In fact, the majority of polls reported in traditional media are similarly inaccurate. And while it would be nice - as you suggest - were online polls to lable themselves as "unscientific", it would be nice were traditional polls to do the same.
Darrell Huff ("How to Lie With Statistics") reports on the many and varied ways in which polls are inaccurate. In my experience, the two major practises employed by pollsters today are unrepresentative sampling and biased questioning.
Consider, for example, a poll recently conducted by the University of Alberta to measure student satisfaction. This poll was splashed through the local media as 'fact' even though the survey methodology was highly suspect. The poll was coducted by asking graduates present at the convocation ceremony whether they were satisfied with their education. Not surprisingly, most of them said yes (more surprisingly, given the occasion, a large percentage said no).
Just a couple of weeks ago I participated in a telephone poll conducted by Environics, a supposedly 'scientific' polling company. While the poll asked a wide range of questions, it focussed on the use of government gambling revenues to support social services.
At no time was the question "should gambling revenues be used to support social services" asked. In fact, I am opposed to this use of gambling revenues, as I am opposed to gambling revenues generally. However, the pollster insisted that I choose between supporting health care, education or social services.
Not surprisingly, the headline a week later read, "Most Albertans support the use of gambling revenues for cancer research."
Television and radio stations have long conducted phone-in polls; these polls are reported on the evening news as eventful, and viewers assume that they have received a fair sampling of the public opinion.
The proliferation of polls on the internet is dangerous to pollsters, not because they are less reliable than those offered by other media, but because their multitude and similarity to traditional polls exposes to the public the illegitimacy of the traditional poll, an event the professional pollster fears more than public opinion itself.
p.s. You should have a 'talk-back' or discussion area for your column. I had to dig a bit to find an email. Lest my reaction to your column vanish in obscurity, I am posting it in the 'threads' section of my site.
SUBSCRIBE TO OLDAILY DONATE TO DOWNES.CA
Web - Today's OLDaily
Web - This Week's OLWeekly
Email - Subscribe
RSS - Individual Posts
RSS - Combined version
JSON - OLDaily
National Research Council Canada
All My Articles
About Stephen Downes
About Stephen's Web
Subscribe to Newsletters
Privacy and Security Policy
Stephen's Web and OLDaily
Half an Hour Blog
Google Plus Page
Huffington Post Blog