Science of examinations and competitions, docimology, of “dokimè” (test) and “logos” (science) finds its origin in the work on the validity of the notation systems of the French psychologist Henri Piéron. It was in 1922 that he proposed the concept and launched research around the results of the primary school certificate. These would be popularized by his book Examinations and docimology published in 1963 , followed in 1971 by Gilbert Landsheere ‘s Précis de docimologie .
These founding works evoke, in an original way, the behavior of “examined and examiners” in order to make us aware of the biases and uncertainties that weigh on the ratings and evaluations. However, the concern is far from new. The very invention of the note is the fruit of a long history . While the United States moved towards assessment via MCQs in 1910 , in France they preferred to keep the classic assessment of exams with written answers.
How do you judge an exam paper? What about the relevance of the marks in case of repeating the exam or changing examiners? Answering these questions refers precisely to the founding work of docimology. They show, from an analysis of variance, that the part of the explanation of a note would be linked, up to 40%, to the student’s skills. As a corollary, 60% would result from the identity of the examiner.
If the method used by these precursor works is open to criticism, they nevertheless underline all the ambiguity surrounding the justification of a note. Henri Piéron will even go so far as to say that “to predict a candidate’s mark, it is better to know his examiner than himself! »
As for the psychologists Laugier and Weinberg, they will try to determine the number of corrections necessary to achieve a score that is “fair”. In doing so, mobilizing the Spearman-Brown formula , they will come to the conclusion that in philosophy, for example, 127 correctors would be needed to achieve a fair score.
This controversy over the difficult marking of copies is regularly found in the front line during the corrections of the baccalaureate . A research from 2008 will even go so far as to evoke a kind of lottery of marks in the baccalaureate .
In fact, beyond the controversies, docimology raises a major question. Should the marks of a class, as is very often the case, correspond to a Gaussian distribution , i.e. with a few “weak” pupils, a few “strong” pupils and the vast majority of pupils in the “average”, like the tyranny of Normal Law that reigned in Japan between 1955 and 2000?
A dilemma of size refers indeed, almost systematically, to the average of the marks which inevitably induces a form of chance. Too high an average could mean a poor assessment of the level of the students, while the opposite could suggest too strict marking criteria. But, if the majority of students have a mark between 8 and 12, what is the point of using a grading scale of 0 to 20?
Beyond the averages, common sense requires taking into account, with more attention, the standard deviation, that is to say the range of the scores, from the lowest to the highest. In doing so, it is easy to imagine that a teaching subject could have a decisive influence on the general average when the standard deviation of the marks would be higher than that of the other subjects taught.
Ultimately, what is the purpose of a note? Is it a question of appreciating the level of competence and knowledge of the pupils at a time T, or of using the notation as a tool in order to classify, if not to filter the pupils, in the idea of achieve an optimal allocation of talent in society?
Author Bios: Nadir Altinok is a Lecturer, IUT of Metz, UMR BETA at the University of Lorraine and Claude Diebolt is Research Director at CNRS, UMR BETA at the University of Strasbourg