SCI199Y: Short project 3: due November 28
Background: What makes a good poll
The article in the Globe and Mail that was handed out on October
10 is a very clear, if brief, account of what makes a good poll.
In Statistics: concepts and controversies, the following poll
guidelines are listed:
- What was the population? That is, whose opinions were being sought?
- How was the sample selected? Look for mention of random
sampling, probability sampling.
- What was the size of the sample? It is even better to give the
margin of error.
- How were the subjects contacted?
- When was the survey conducted?
- What were the exact questions asked?
The author goes on to say
The codes of ethics of the American Association for Public Opinion Research
(AAPOR) requires disclosure of this information. The major opinion polls
always answer these questions in their press releases when announcing
the results of a poll. But newspaper editors have the bad habit of cutting
out the paragraphs containing these facts and printing only the lead
paragraphs announcing the results. Worse, new methods for conducting
telephone surveys have brought a proliferation of state or regional polls
run by newspapers or broadcasting stations. ... If a politician, an advertiser,
or your local newspaper announces the results of a poll without complete
information, be skeptical.
Assignment
- Find (and submit a copy of) a convincing
newspaper or magazine article about a poll.
Explain why you think it is convincing.
- Find (and submit a copy of) a newspaper or magazine article that
strikes you as possibly misleading. Explain why you find it so.
Hints: The New York Times is a pretty reliable source of
good poll information, but so also are the Globe and Mail
The Times, and the Boston Globe. Magazines often report
results of readers' letters as if they were a scientific poll ("a sample
of our readers writes", "the results of our self-selected sampling indicate"...).