Why an Observatory ?

Why consider the field of higher education and research ?

Nowadays, equality is a core value of universities. For this reason, discrimination is an unacceptable practice that institutions are committed to fighting. However, this has not always been the case.

Created from the 13th century onwards, French universities excluded women and men who were not Catholic. Generally speaking, we can consider that access to universities, which today is based on academic criteria, was initially based on social status [Goastellec 2020]. It has therefore taken many transformations for people of all genders and from all backgrounds to be able to access higher education, in law and in reality. It has taken even more to enable them to occupy dominant positions in this system, to decide on its modes of organisation and governance, and to participate in the construction and dissemination of knowledge. However, universities are not only reflections of the societies of their time, they also generate hierarchies and can help modify or reinforce them.

It is therefore necessary to examine higher education through the prism of equality, both diachronically and synchronically, using various disciplines, to enable institutions to implement the values of equality and universality being claimed. On the one hand, facts have to be established and on the other hand, they have to be analysed. Understanding the mechanisms of inequality production and measuring discrimination appear to be requirements to progress in the transformation of universities on the one hand, and for the latter to be able to play a role of social transformation on the other.

At first glance, in the world of higher education and research, discrimination is difficult to conceive. It goes indeed against the values of equality and universality that are the basis of the university project. Moreover, discriminating restricts the recruitment space and deprives one of skills: it is therefore an irrational behaviour, contrary to the interests of one's training and one's institution.

However, several facts show that the higher education and research environment is not impervious to the risk of discrimination.

  • The report of the High Council for Equality published in November 2021 shows that sexist and sexual violence in higher education and research is a massive and trivialized phenomenon, but one that is underestimated and misunderstood, with the code of silence and impunity remaining the rule.
  • The subject is taken seriously: according to the recent report of the Ministry for Higher Education, Research and Innovation (January 2022), each of the 181 higher education institutions has adopted an equality plan, including the fight against discrimination (in accordance with Article 80 of the law on the transformation of the public service of 6 August 2019).
  • Tests have already shown that discrimination also exists in the public sphere, especially in access to public employment (L'Horty, 2016; Challe et al., 2018; Petit et al., 2020).
  • Other experimental studies have shown that teachers fall victim to their stereotypes about students' chances of success in their assessments (Alesina et al., 2018; Papageorge et al., 2020).
  • Although examination boards are legally established for all courses, in practice course leaders often make their decisions in isolation, especially for master's level courses with minimal class sizes.

Discrimination: a problem of objectification

Of all forms of violence, discrimination is one of the least visible. From the victim's point of view, it takes the form of a refusal of access to a given resource (a job, housing, a public service, etc.); the causes may be multiple (and not necessarily identifiable). From the point of view of the person who discriminates, discrimination may be unconscious. Even if it is conscious, the discriminator has no interest in revealing his or her behaviour, as it is a crime.

Direct measurement on pre-existing databases is almost impossible:

  • Business sources or public statistics inform about differences in situation, inequalities, but not differences in treatment in a process of access to a resource on the basis of a prohibited criterion
  • Most prohibited criteria are not observable in existing databases (origin, religious beliefs, sexual orientation, disability, etc.)
  • It is difficult to achieve measurements all things being equal using application data (e.g. civil service competition data). Moreover, they do not cover all prohibited criteria (Bréda and Hillion, 2016; Greenan et al., 2019).

Thus, observing discrimination requires specific equipment.

Measuring discrimination: several sources

There are three types of discrimination monitoring systems that are systematic in scope.

Two of them are based on the experience of those discriminated against:

  • Spontaneous reports (e.g. annual reports of the Defender of Rights; listening units; 3928, www.antidiscriminations.fr)
  • Solicited reports: victimisation surveys (e.g. INED-INSEE Trajectoires et origines survey; ACADISCRI project)

The third one consists of directly observing the behaviour of discriminators:

  • Testing method: audit by couple (real people); test by correspondence (fictitious people). The correspondence test has gradually become the reference method in the international scientific literature on discrimination (Baert, 2017; Bertrand and Duflo, 2017; Neumark, 2018).


Testing: advantages and disadvantages

Surveys or complaints have an advantage: samples are potentially representative. However, there are several limitations:

  • Discriminated people are not always in the best position to witness discrimination: this is a non-observation bias.
  • When they are, they may not wish to report themselves as discriminated against: this is a non-reporting bias.
  • Even if they do report, the feeling of discrimination is not always an objective measure of discrimination, especially because one is not reasoning "all things being equal".

Testings have several advantages. The measure is fully controlled by the evaluator. The effect of the characteristic to be tested can be isolated. The measure is reliable, unbiased and « all things being equal ». However, there are also limitations:

  • The measure is partial: on the labour market, a small number of occupations are generally tested.
  • The measure is punctual (over a given period).
  • The measure is localized in a given territory (most often the Île-de-France Region).

This is peculiar to all experimental measurements. To overcome these limitations, testing have to be done again and again. It is also necessary to combine approaches and to reiterate them over time: this is the objective of ONDES.