NCPA - National Center for Policy Analysis

Polling: Tell Me What I Want To Hear

December 17, 2002

Public policy advocates often hire pollsters to survey the electorate in support of an agenda they seek to promote. In almost every case the polls they commission reveal answers they like. Critics argue this is because pollsters often prompt respondents by wording questions in such a way to get desired answers.

A recent poll of Canadians on funding a major health care initiative supports this argument:

  • The poll reported that 82 percent approved a report's recommendation for C$15 billion in additional health care spending over three years.
  • This held even if that spending took the entire projected budget surplus.
  • But when the polling firm spelled out that little money would be left for other priorities, support suddenly fell by a third.
  • Indeed, the poll also found that 88 percent saw no need for new taxes to pay for health care

And yet, the same poll also found only 6 percent of those polled were "very familiar" with the report. Half the respondents knew little or nothing about the report on which the spending recommendation was based.

Similar affirmative responses were received in polls about increased defense spending, a $260-million upgrade of the Calgary Olympic facilities, ratifying the Kyoto accord, etc. In every case, pollsters found general support for an idea, based on little information with no information on the tradeoffs of more spending.

How credible are polls that ask people wish-list questions without asking for tradeoffs? The reason people don't make trade-offs when polled is because pollsters don't ask them to.

Source: Jeffrey Simpson, "When They Poll the Know-Nothings," Toronto Globe and Mail, December 11, 2002.


Browse more articles on Government Issues