Gauging tradeoffs

Reader comment on: Bloomberg's Health Care Poll

Submitted by Fred Van Bennekom (United States), Oct 14, 2010 07:50

Several things intrigued this survey designer. First, if you believe this poll, Pelosi will remain speaker by a slim margin as democrats lead in the generic ballot 42%-40% (with an accuracy of 3.7%). But the pollsters say they weight the result by gender and race, but not by political background. By random chance they may have polled more democrats. However, this question is posed right after the health care questions, so a sequencing bias is certainly in play -- we just can't tell in which direction.

Second, the question on the elements of health care shows, as noted elsewhere, that the respondents weren't forced to make trade-offs, but that's the critical point. Interval rating questions don't force trade-offs. Other question formats (e.g., fixed sum) or a conjoint approach would have forced consideration of the trade-offs.

Note: Comments are moderated by the editor and are subject to editing.

Other reader comments on this item

Comment on this item