The Political Persuaders by Dan Nimmo

Author:Dan Nimmo [Nimmo, Dan]
Language: eng
Format: epub
Tags: Political Science, Political Process, General
ISBN: 9781138537477
Google: QDnSDwAAQBAJ
Goodreads: 56465687
Publisher: Routledge
Published: 2020-02-18T12:24:36+00:00

The Art and Science of Political Polling

To serve as a reliable basis for planning and executing campaign strategy a professionally conducted poll must provide an accurate reflection of constituency opinions. In the brief history of the polling industry, the procedures of opinion-surveying have been refined until its most cautious practitioners describe polling as an art and the most confident researchers label it a science. To possess a full grasp of the strengths and limitations of the product client-candidates purchase at such a high price, we need to examine these procedures.

The sample. The first step in a properly conducted political survey is to select a group of individuals for interviewing whose opinions accurately reflect those of all adults of voting age in a given constituency. The sample need not be large, but it must be representative.

The average citizen who is only casually acquainted with political polling is reluctant to accept the notion that pollsters can derive an accurate approximation of what an electorate thinks by talking to 1,500 adults (frequently the maximum size sample used by private firms). Yet pollsters do precisely this. Governor George Romney withdrew from the New Hampshire presidential primary after his pollsters delivered pessimistic forecasts based upon surveys from 700-800 potential voters. Certainly there are possibilities of inaccuracies in small samples. But the refinement of polling procedures is such that, if the sample is properly drawn, it is possible to reduce the range of sampling errors so that in ninety-nine cases out of a hundred the results of a survey of 1,000 properly selected adults will be within 4 per cent of accuracy. For example, let us assume that in a nationwide survey of 1,000 adults 40 per cent of respondents say they approve of the way the president is handling his job. The chances are ninety-nine out of a hundred that somewhere between 36 and 44 per cent of American adults hold this opinion. If our sample size had been 4,000, our range of error would have been 2 per cent, permitting us to estimate that 38-42 per cent so evaluate the president’s performance. Had our sample consisted of 17,000 adults, the error range would be 1 per cent. But there is a point at which it ceases to be profitable to reduce sampling error by increasing sample size. Sample precision increases with the square root of the number of persons interviewed; to cut sample error in half, the size of the sample must be increased four times, or the square of the error. At a minimum cost of $10 per interview, a 1 per cent error range would be a $170,000 luxury; at those prices both pollster and client are willing to tolerate a 4 per cent error in a survey costing $10,000. Another important point is that the size of the population under survey has little bearing on the size of the sample required for a given degree of precision. If we seek a 5 per cent error, we are wise to use a

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.