We can make politics with an algorithm. This is the bias of Sciences Po Bordeaux, which, seventy-five years after its founding, has transformed its method of recruitment. In 2020, the establishment moved from the clic written competition to the in-depth examination of application files on the Parcoursup platform, supplemented by an oral admission.
The Institute of Political Studies (IEP) makes no secret of it: its file selection algorithm is configured to favor scholarship students and those who come from high schools participating in “cordées de lasuccess”, a device for fighting against self-censorship resulting in continuous guidance support from the 4th gradee.
Based on “the deviation from the average”, this algorithmic calculation differs from that which is proposed by the Ministry of Higher Education, called “decision support tool” (OAD), based exclusively on the notes. This choice was born out of the institute’s desire to take advantage of the leeway left to training courses on the platform, but also to answer a question that has become a burning issue, so as not to harm high school students in their applications: how can higher education measure the level of a student when the whole cl has obtained good results?
Unlike the OAD, the Bordeaux algorithm does not discriminate against candidates from high schools where the rating has remained severe, “but it is unfavorable to high schools that practice grade inflation and to those whose student level geneity is such that there is no dispersion of grades within the cl”explains its designer, Professor Vincent Tiberj, a former Sciences Po Paris student who worked in 2001 with Richard Descoings – who was then director of the institute – to launch priority education agreements, a new and imposing entry route. more social diversity in the school.
The result is clear: during the current recruitment campaign, 72% of those eligible would have been the same, regardless of the tool used, and 28% of those eligible would have been retained by the help tool alone. decision-making or by the “in-house” algorithm alone.
“Some 300 students were therefore not selected because we did not use the ministry’s decision-making tool, establishes Mr. Tiberj. I tell myself that these students will have good prospects elsewhere, because the other courses use the OAD. On the other hand, those who were received thanks to the in-house algorithm would not have had good prospects elsewhere. »
You have 59.6% of this article left to read. The following is for subscribers only.