Page 193 - Contributed Paper Session (CPS) - Volume 4

CPS2173 Felicien Donat Edgar Towenan Accrombessy et al.
Hence, evidence of the descriptive analysis and PCA informed the
regression. The theory and empirical literature on the determinants of
students' academic school performance is relatively abundant. A number of
factors (inputs) combine a series of activities, practices and conditions to
produce school output. Academic performance studies identify three
categories of variables as factors of the production function in education. The
first category of inputs relates to the individual characteristics of pupils:
gender, age, cognitive structures (intelligence, motivation, self-perception),
etc. Then, the second category of inputs concerns the variables related to the
family environment of the pupils such as: the level of education of the parents,
the availability of capital goods and educational material within the household
(computer, dictionary, textbooks, ...), the language used at home by the family,
the size of the family, the child's participation in domestic or rural work, etc.
Finally, the variables related to the school context consider the characteristics
of the teacher (gender, training, motivation, ...) and those of the school such
as class size, equipment, pedagogical practices and organizational
characteristics. Sociocultural variables, particularly those related to the family
environment, influence the child's academic success (Duru-Bellat 2003, Diallo
2001, Fuchs et al 1999). Estimates show that school results are better for
children whose parents are educated. The same conclusions are reached in the
case of Haiti. On the other hand, in the case of Morocco, Hijri et al, 1995 show
that this relation is not significant. This result can be explained by the fact that
women with a high level of education were easily engaged in professional life
and entrusted the care and support of their children to housekeepers, often
without any level of education. In Benin, based on the 2014 survey data and
linear descriptive modeling, PASEC assessed success factors at the end of
primary schooling. As a limit, this study does not consider family and other
individual variables that may be involved in explaining these differences. For
example, the availability of capital goods and teaching materials within the
household, the language of the family, the size of the family, the skill level of
students entering PS, the amount of time spent for homework, student
engagement in learning, etc. are also key factors.
Our analyses focus on PASEC data collected in 2004 and 2014 on student
learning outcomes in primary education in Benin according to standards that
facilitate comparison between CONFEMEN countries. Since a multiplicity of
factors acts simultaneously on school performance, we retain the variables to
which the literature attaches great importance and that the PCA put emphasis
on, including the three categories of variables mentioned above. The existence
of two levels of analysis within the model poses the problem of non-
compliance with the two essential assumptions of the ordinary least squares
(OLS) approach, namely the independence of observations and
homoscedasticity (Snijders and Bosker, 1999 and Bressoux, 2007). An attempt

182 | I S I W S C 2 0 1 9