That was the deciding factor in the votes on Frontex and Co

It would have to go to and fro with the devil for the percentage of yes votes in two votes at the community level to be absolutely identical. In some cases, however, the most recent voting results come very close to the results of earlier votes. This is shown by a comparison of all submissions since 1981 with the latest referendums on “Lex Netflix”, on the transplantation law and on Frontex financing.

Correlating voting results can be used to determine which population groups were able to be mobilized and which were not. A correlation measures the strength of the statistical relationship between two variables. In this case, this variable is the yes vote share of all 2120 municipalities in the relevant templates. By comparing it with previous votes, it is possible to work out what was decisive in the decision-making process of the citizens.

Which votes are missing at the top

The first surprise was how strong the dependency of the results of the templates on “Lex Netflix” and on the transplantation law, which were little related in terms of content, was. Because the content of the two templates seems to have nothing to do with each other. But in communities where support for one bill was high, the same was true almost across the board for support for the other bill.

What was missing at the top of the list were, for example, in the case of the Transplantation Act, votes that were related in terms of content, such as those on the Reproductive Medicine Actwhich was accepted on June 16, 2016 with 62.4 percent.

According to the political scientist Michael Hermann, one explanation for this could be that many earlier health bills lacked a liberal counter-argument. “And that’s exactly what we had with the Transplantation Act,” says Hermann, “namely that each individual is allowed to dispose of his body and that the state has no place in it.” This argument apparently caught on with those who voted no.

Even with the “Lex Netflix” one searches in vain for related votes among those with the strongest linear correlation. There is a media law that was rejected in February, certainly a certain dependency. However, it is significantly lower than, for example, the voting on the acquisition of civil rights by third-generation foreigners; this federal resolution was rejected by the people on September 26, 2004.

“But the colorful bouquet of correlating votes has its logic,” says political scientist Hermann. The civil rights of foreigners or income replacement in the event of maternity are left-progressive social issues. And such a socio-political expression was apparently also decisive in the yes to the film and transplantation laws. In both votes, the front line clearly runs between left and right political orientation. That is why the dependency of the Yes shares of the two votes is so high.

The third vote from last weekend falls outside of this framework: the one on Frontex financing. Your correlation list is peppered with templates that are very close in content. “That does not surprise me. Foreign policy often reveals its own voting patterns,” says Hermann. In fact, out of the 5 voting results most similar to the one on Frontex, 4 concern foreign policy. And the fifth, that of fighting terrorism, also has a foreign policy dimension.

What is also noticeable, however, is that the correlation with other votes is much less pronounced for Frontex than for the other templates. “In fact, this original was an oddball. Because on the one hand it means opening up to Europe, but on the other hand it means a more restrictive migration policy – ​​i.e. opening and closing at the same time,” says Hermann. This constellation meant that national conservatives and bourgeois liberals formed the clearest front line. The broad bourgeois support for Frontex funding also explains why the bill did less well in the core cities than in the agglomerations.

A total of 361 popular initiatives were examined

For the analysis of the correlations, all votes since 1981 were taken into account; a total of 361 popular initiatives, referendums or decisions on federal laws. It was calculated how well the results matched from community to community. This does not mean that the yes share of communities has to be exactly the same, but that the difference in the yes shares from community to community remains the same.

An example: the Zurich municipality of Aeugst am Albis narrowly rejected the transplantation law with 49.9 percent yes votes, and accepted the film law with 52.4 percent. The municipality of Affoltern am Albis, also in the canton of Zurich, accepted the transplantation law with 53.2 percent; in the case of the film law, the yes share was 56.4 percent. In the two municipalities, the approval for both templates is 3 percent apart. The voting results on the Transplantation Act and the Film Act of these two communities therefore correlate in this difference.

The method is used by political scientists after voting in order to get closer to the relevant political issues of the campaigns conducted. However, this is usually only done at canton or district level. Here, the corresponding calculation was prepared for all municipalities with the help of computer code, in order to then compare the voting results with earlier templates at the push of a button.

Data: For the study, the real-time data on the voting day on federal voting templates from the federal open data portal were used. The data includes all templates since 1981 and can be viewed here.

method: The Pearson correlation method, also known as product moment correlation, was used to calculate the correlation. It denotes a measure for the degree of linear context between two at least interval scaled characteristics, which does not depend on the units of measurement and is therefore dimensionless. It can take values ​​between +1 and -1. With a value of +1 (or –1), there is a completely positive (or negative) linear relationship between the characteristics under consideration. When the correlation coefficient is 0, the two features are not linearly related at all.