A comprehensive study published in the scientific journal Science Advances reveals that researchers’ personal political positions have a significant influence on the results they arrive at.
The researchers George J. Borjas and Nate Breznau have conducted an experiment that shows how ideological blinkers shape modern social science.
The study addressed one of the most contentious topics in public debate, namely the question of whether immigration affects support for public welfare schemes. A total of 158 researchers, divided into 71 different teams, took part in the experiment. All participants were given exactly the same dataset and were tasked with answering the same question.
Neither objective nor neutral
The results from the various teams diverged in all directions and ranged from strongly negative to strongly positive effects. The most striking finding was the direct correlation between the researchers’ own attitudes towards immigration and the conclusions in their reports.
The research teams that had stated in advance that they were positive towards immigration largely found results that supported this view. They often concluded that immigration has a positive effect on support for the welfare system. On the other hand, the teams that were critical of immigration, in almost all cases, found negative effects.
Climate demonstrations such as this one in Berlin, Germany, in 2025 help to shape researchers’ own convictions. Photo: AP/Ebrahim Noroozi/NTB
Model selection as a tool for bias
The study shows that this bias does not necessarily stem from deliberate cheating or manipulation. Instead, ideology enters into the research design itself. Researchers choose different statistical models and variables that align with their expectations and political convictions.
This means that the very method selected functions as a mechanism for producing the desired answers. The teams that were positioned at the extremes of the political spectrum often chose model specifications that produced the most extreme outcomes, in line with their own standpoint.
Moderate researchers have the highest quality
The experiment also included an anonymous peer review in which the various models were evaluated by other researchers. Here it emerged that the ideological extremes delivered work of lower quality. The most moderate teams received the highest support from their colleagues.
This finding challenges the notion that academia is a neutral arena for the pursuit of truth. It means that climate models from researchers who are clearly influenced by an ideological bias towards the green transition must be taken with a pinch of salt.
It also means that studies on the Covid vaccine must now, in retrospect, be viewed with far greater scepticism.