What is raw data

Why do raw data deviate from the representative results?

The raw data is the simple clicks. In theory, multiple votes are also possible here. In contrast, only the voting results of registered users are included in the calculation of the representative data. In addition, these are weighted according to the population average of Germany, that is: each population group is taken into account according to its percentage of the total population (for nationwide surveys).

If raw data and representative data deviate significantly from one another, there can be various reasons: For example, it can mean that many voters have not registered. Or that there was a bias in certain factors. For example, because a lot of voices came from a certain region. However, we only consider each region according to its proportion of the total population. This ensures that the opinion of a Cologne citizen, for example, does not count more than the opinion of a Frankfurt woman. A simple example: 80 percent of men take part in a survey, but only 20 percent of women. So that the voices of men are not disproportionately significant (because the gender distribution is fairly even in Germany), the voices of women have to be taken into account a little more and those of men accordingly less so, so that in the end just as many women as men to speak. This process is called weighting.

The more controversial the opinions of different population groups on a topic (e.g. supporters of different parties or different age groups), the higher the likelihood that raw data will differ significantly from representative data. For topics where all Germans have a more or less similar opinion, the raw data will more likely correspond to the representative results.

Further information on the subject can be found in our methodological explanations.