This post is also available in: Deutsch
It is Sunday and we are in Bremen in the middle of the election campaign. Local, regional and European elections are scheduled for May. For this reason, the so-called Sunday question is being discussed more and more frequently and the results of surveys are being discussed. How do polls on elections come about? I’ll answer that question today on my blog.
Opinion polls before elections
Opinion polls before elections are carried out by professional institutes. Usually they are also referred to as the Sunday question. So one can compare the Sunday question with the “Gretchen question” from Goethe’s Faust. The answer to both questions should be formulated in good conscience. What exactly is the Sunday question? If there were elections next Sunday, which party would you vote for?
For example, opinion research institutes determine a certain number of people from different population strata for a survey beforehand. This should make the survey representative and meaningful at a later date. There is a specific method for evaluating the survey. It is called statistical extrapolation and ends with the prognosis of the election outcome.
The institutions have various methods for determining the intention to vote. One of these is that respondents are given a choice of several parties on the electoral list to answer their questions. The interviewee must then decide on a possible answer on the basis of this specification. These personal surveys are carried out by home visits or street surveys.
In another method, respondents are asked to spontaneously name their favourite party from the gut. For cost reasons, surveys are usually carried out by telephone. Election polls usually take place shortly before elections. The publication of the results can or should influence the electoral behaviour on the actual election day.
Hit rate of forecasts
However, the past has shown that election forecasts often have a considerable margin of error. The honesty of the interviewees plays a role here. Because not all answers reflect the real own opinion.
Sometimes the interviewee is still undecided, sometimes the interviewee doesn’t want to come out and say something wrong, and sometimes he changes his mind until the election. Therefore, opinion poll forecasts are in principle less accurate than weather forecasts.
Clients for opinion polls often come from the media or from the party landscape itself. The intention of opinion polls and their publication is, among other things, to raise the awareness of voters. The aim is to achieve the highest possible turnout.
According to a text analysis this text is 31% subjective. The most important words are: opinion poll, Sunday question, interviewee, survey, forecast, party, interviewee
A computer would sort this text into category politics.