Education of Disadvantaged Children in OIC:
The Key to Escape from Poverty
73
attendance in school for 6-11 year olds and for finishing 5 years of education for 12-15 year olds.
For attendance in school for 12-15 year olds, household head’s education contributes the
most to inequality while for finishing 8 years of education it is the location. Overall, the
combination of the region of the household and location of the household (urban/rural) explain
more than 40 percent of the variation of the D-index for each indicator. Regional and locational
inequalities seem to contribute more to overall inequality than household wealth, except
for finishing 8 years of education. For this indicator, household wealth becomes as important as
region of the household but its contribution is still less than the contribution of location of the
household. In general, the impact of ethnicity on inequality is less pronounced compared to
circumstances like region and location. Lastly, the number of children in the household and
gender of the child are circumstances that have almost no impact on inequality.
Figure 28 Shapley decomposition for 2015
Source: Authors’ calculations using DHS 2015
Therefore the HOI and Shapley analyses corroborate some of the findings in section 2.1. i.e. gender
is not a main determinant of access to schooling while region and location are. However, while
poverty/wealth was shown to have a negative correlation with access to school rates in 2.1, the
Shapley results show that household wealth has a lower an impact on inequality in education than
region/location/education level of parents.
Probit regression results
The DHS regression results confirm that all circumstances turned out to have significant
negative association with education outcomes, in varying degrees. Living in rural areas (as
opposed to living in urban areas), living in regions Diourbel, Fatick, Kaolack, Louga and Matam (as
opposed to living in Dakar), having a household head that has no education (as opposed to having
a household head with a higher education degree), living in a household with 5 children or more
(compared to households with 1 or 2 children), living in a household that is in the 1st, 2nd, 3rd or
4th wealth quintile (as opposed to being in the 5th – the richest- quintile), being Wolof, Poular,
Mandingue, Diola or non-Senegalese (as opposed to being “other”) and lastly being a boy are all
significantly and negatively associated with (i)(ii) attendance in school for 6-11 and 12-15 year
olds and (iii)(iv) completion rates of 5 and 8 years of education. A detailed analysis of the results
for each indicator can be found in the Annex 2 but in summary:
22,1
20,8
19,6
21,5
0,4
1,4
1,0
0,8
10,4
6,7
17,4
15,9
21,1
26,4
21,2
22,7
17,9
21,5
23,8
21,3
6,7
8,9
5,0
6,2
21,4
14,4
12,0
11,6
0 10 20 30 40 50 60 70 80 90 100
Finishing 8 years of education
Finishing 5 years of education
Attendance 12-15 year olds
Attendance 6-11 year olds
location gender ethnicity region household head's education # of children in household wealth