Opinion Polls and Volatile Electorates: Problems and Issues in Polling European Societies by Matt Henn

Author:Matt Henn [Henn, Matt]
Language: eng
Format: epub
Tags: Political Science, General
ISBN: 9780429834837
Google: 5F8PEAAAQBAJ
Goodreads: 43329358
Publisher: Routledge
Published: 1998-08-01T00:00:00+00:00

Table 7.3 Error Effects in the 1992 British General Election Polls (%)

Sampling Variability -

Methodology ?

Differential Registration 1.2

Late Changers 2.5-3.3

Nondisclosure 2.0

Total 5.6

Actual Error 8.5

Unexplained Error 2.9

Source: MRS Report, 1994

However, this leaves approximately a 3% margin of discrepancy to explain in the polls, or about 40% of the error. Following the 1992 polling experiences, a number of recommendations were made which advocated various technical and logistical innovations to reduce error in opinion polls. These included:

If pollsters continue with quota sampling methods they should use as many 'partisan' quota control variables as necessary (MRS, 1994), that they use appropriate indicators as measures of these variables, and that they have access to current and comprehensive base data on which to identify the correct distribution of categories within each variable for every sample point selected.

The constituencies selected for the pollsters' samples should also be given careful consideration in terms of such factors as: the historical distribution of voting patterns (to check on electoral volatility); marginal status; propensity for tactical voting; charismatic political personalities; strong peripheral party candidates.

So as to reduce the problems of non-disclosure and undecideds in the voting intention question, pollsters should use secret ballots. Such a technique could be developed further by using various weighting procedures to emphasise the views of committed 'core' voters, and reduce the impact in polls of the uncommitted.

Modelling either with political weighting, or with the use of alternative questions as voting predictors. This may contribute to a clearer understanding amongst the pollsters both of the strength of feeling on political issues and party preferences, and of revealing clues about the likely voting behaviour of the don't knows and won't says, and estimating and accounting for differential turnout.

Finally, whether or not quota samples themselves should be replaced, and pollsters revert back to random methods.

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.