Many people believe that election polls are a reliable origin of information about social sentiment and support for individual parties or candidates. However, as the political scientist from the University of Warsaw emphasizes, Dr Bartłomiej Bishop, polls can be unbelievable if they are not carried out in a liable and transparent manner.
Test methodology and reliability of surveys
The test method, investigating and presentation of results are crucial to the reliability of the surveys. According to Dr. Bishop, for polls to give a real image of social sentiments, all available investigation techniques should be used. The results will be as reliable as possible. Each method has different properties. For example, telephone interview is much more accurate, but it is much harder to get respondents due to the fact that photovoltaic killed telephone research.
Deficiencies of individual test methods
net investigation is carried out easier, but gives over-representativeness to people from younger and better educated groups. Moreover, frequently these studies are carried out utilizing a fixed pool of respondents, which reduces their accuracy. Face-to-face tests are the most costly and have the highest number of refusals, although they are inactive used.
Mixed-mode – solution for reliable polls
According to Dr. Bishop, present we gotta usage a mixed-mode, which is simply a mix of investigation techniques, due to the fact that no of them produce 100% reliable results. Mixed-mode allows you to get more reliable results as it reduces errors associated with individual test methods.
Unbelievable electoral flow tests
any studies of the flow of electorate are besides unbelievable – in peculiar between candidates or parties with small support. investigation showing the movement of voters on a tiny example of respondents is completely unbelievable. Statistically, this cannot be calculated reliably. If there are 600 active voters in the poll about 1,000 active voters, for example Adrian Zandberg will support 2% of respondents, or only 12 people.
Manipulation of poll results
Another aspect where manipulation can happen is the presentation of poll results. According to a political scientist, the unattainable poll poles are a model example of this phenomenon. frequently the candidate who won has a post twice as large as the 1 who has only 10 percent little votes. Excel wouldn't make those results. This is calculated on a strong media message, not to present decently research.
Statistical mistake and reliability of surveys
According to the scientist, attention should be paid to the size of the test sample, which is how many respondents participated in the study. 1000 respondents are an acceptable test attempt, a statistical mistake is then 2-3 pp. The smaller the sample, the bigger the error.
Conclusions and recommendations
Therefore, the polytologist recommends that erstwhile assessing the surveys, attention should be paid to the size of the test sample, the method of investigating and the presentation of the results. In addition, a statistical mistake and undecided voters who are or are not included in the survey should be taken into account. This allows us to get more reliable information about social sentiments and support for individual parties or candidates.
Continued here:
Why can election polls be unbelievable?
