The posts that people see on the platform “X” could influence their attitudes toward certain political issues. This was the conclusion of a study published in the journal Nature by a European research team.<\/p>\n
The Science Media Center (SMC) asked experts, including Prof. Dr. Judith M\u00f6ller, to contextualize the study and evaluate its findings and methodology. In her statement, M\u00f6ller explains that the study confirms that media content impacts how issues are assessed but does not demonstrate that it can fundamentally alter political attitudes.<\/p>\n
“This is a very carefully conducted experiment that reveals small yet statistically significant effects. These effects relate to how algorithmic feed settings influence the political issues people consider most important and their attitudes toward relevant political questions. One example of such a question is how people evaluate their government’s performance in the Ukraine war.\u201d<\/p>\n
\u201cIt is typical for media effects to be more likely to occur when it comes to which issues people find important. However, effects on political values or more stable attitudes, such as party identification, are less likely.\u201d<\/p>\n
\u201cI consider the methodology of the study to be very sound. All experiments can be criticized on the grounds that people may behave differently when they know they are participating in an experiment. They do this, for example, to help science. However, the design ensures that the impact on the results is minimal: groups in which the algorithm was changed are compared with those in which it remained the same.\u201d<\/p>\n
Of course, it would be better if the sample were representative. However, even if white women with a high level of education initially tended to use the chronological feed, one thing must not be forgotten: the white women who changed their feed settings were not compared with the general population. Instead, they were compared with those who did not change their feed.”<\/p>\n
“Of course, errors can occur in surveys due to incorrect information. I cannot judge the extent to which this happens. However, YouGov is a panel provider with a high reputation that strives for high data quality.\u201d<\/p>\n
It is striking that those who had to change their feed were significantly more likely to drop out of the experiment. This suggests that the process was not entirely straightforward from a technical standpoint, or that it was too inconvenient for participants.”<\/p>\n
“The fact that media content has an impact on the topics we consider important is a well-known finding in communication sciences. This also applies to other media, such as TV or newspapers. The lack of effect of social media on affective polarization has also been noted many times (affective polarization: (negative) feelings toward those who think differently, e.g., members of another political party). In my view, both findings are also transferable to Germany.”<\/p>\n
“To my knowledge, the effect of social media use on conservative political attitudes\u2014for example, toward the war in the Ukraine\u2014has rarely been observed. In the German context, it is important to note that only five percent of the population uses X for news. First, it would have to be examined whether there is a bias in favor of conservative topics and attitudes in Germany.”<\/p>\n
“I found the part of the study that dealt with what was visible in the feed to be very interesting. The study states: ‘The algorithm promotes political content and, within that category, prioritizes conservative content […] . The algorithm demotes accounts of traditional news media and promotes those of political activists […] . Posts from entertainment accounts are 9.1 percentage points (21.5%) more likely in the algorithmic feed […] .’ The algorithm promotes political content and prioritizes conservative content within that category […]. The algorithm demotes accounts of traditional news media and promotes those of political activists […]. Posts from entertainment accounts are 9.1 percentage points (21.5%) more likely to appear in the algorithmic feed […] (editor’s note from SMC).”<\/p>\n
\u201cI consider the downgrading of news organizations to be particularly problematic in view of the study’s findings. This means that the algorithmic group not only saw more conservative content, but also less content that had been verified and classified by journalistic means.\u201d<\/p>\n