Communication via telephone call
In the realm of political polling, the method of data collection plays a crucial role in drawing accurate conclusions about public attitudes. A recent study in Germany sheds light on this issue, comparing the results of six polling institutes using three different methods: a pure telephone survey, two online, and three mixed-mode surveys.
The crux of the matter lies in the representativeness of the samples. Only surveys based on a random sample can be considered representative of the population. This is because only population-representative surveys allow conclusions to be drawn about attitudes in the population.
In the case of Germany, representative surveys based on random samples typically show a more balanced distribution of political preferences, with established parties like CDU/CSU and SPD receiving significant support. However, online or mixed-mode surveys often report stronger support for parties like AfD, indicating a potential bias towards more extreme or protest opinions in non-random sampling methods.
For instance, Ipsos' representative Sonntagsfrage in July 2025 gave AfD 24%, while CDU/CSU received 26%. In contrast, INSA's online survey in Thuringia in September 2025 showed AfD at 37% and CDU at 25%, illustrating this disparity between sampling methods.
The disparity can be attributed to the differences in the methods of sampling. While a random sample with 60% landline and 40% mobile share was used for the telephone survey, the participants in the online surveys came from registered groups for online studies of various institutes. This non-random selection can introduce bias that cannot be corrected by retrospective weighting.
Moreover, the share of older people in online polls often does not suffice to represent the older age groups, with 10% of eligible voters not using the internet, especially those aged 70 and above. Conversely, the share of older people in pure telephone polls is disproportionately high.
The study also found that participants who voluntarily register for online access panels differ in their political behavior on the internet from those who participate in purely telephone surveys. Mixed-mode methods, which combine telephone and online surveys, have significant content differences due to the lack of random selection in the online part.
The mixed-mode method is not representative for data collection regarding the population. Telephone and face-to-face surveys are more trustworthy due to their samples being based on chance and their contact methods reaching a larger and more diverse group of people.
Even within party affiliations, there are biases, as shown by weighting based on the Sunday question. The age group over 70, making up 21% of the total population in Germany, should be considered more strongly, particularly in political polls due to clear age effects in voting behavior.
In conclusion, the method of data collection significantly impacts the results of political polls. Pure online samples do not paint a representative picture of the population, and mixed-mode methods, while convenient, may introduce biases that undermine their representativeness. Therefore, it is essential to prioritise representative surveys based on random samples to ensure the accuracy and reliability of political polling.
Read also:
- United States tariffs pose a threat to India, necessitating the recruitment of adept negotiators or strategists, similar to those who had influenced Trump's decisions.
- Weekly happenings in the German Federal Parliament (Bundestag)
- Southwest region's most popular posts, accompanied by an inquiry:
- Discussion between Putin and Trump in Alaska could potentially overshadow Ukraine's concerns
