« Profs paying for research? | Main | Good Roads Everywhere »

How to Create Question-Order Effects in Surveys

Here’s a great piece on why it’s so dangerous to rely on a single-item measure of public opinion, especially without understanding what other questions preceded that item in the survey.

[Hat tip to Eric Lawrence]

Posted by Lee Sigelman on October 25, 2008 08:58 AM |

Comments

Brilliantly funny! I have emailed the link to my students and to many colleagues.

Posted by: Steve Schier | October 25, 2008 11:29 AM

Have always loved Yes, Minister :-)

Posted by: nimh [TypeKey Profile Page] | October 25, 2008 09:37 PM

And I, in turn, tip my hat to Joey O’Mahoney’s mother.

Posted by: Eric L. | October 26, 2008 03:56 PM

It really IS times that Yes, (Prime) Minister becomes part of the curriculum in North American schools. One of my professors used to take half an hour from every lecture to show us selected episodes.

Posted by: Dubi Kanengisser | October 26, 2008 09:24 PM

I’m glad you enjoyed it. I’ve used it with my students because it neatly illustrates the point.

Posted by: Dr Jo O'Mahoney | October 27, 2008 07:19 AM

OK, this clinches it….I have to shell out to get both series on DVD now…I’m not sure how I’ve managed to teach thus far without it.

Posted by: Matt Jarvis | October 27, 2008 02:58 PM