GENDER GAP AND EDUCATIONAL DEVELOPMENT FOR
RURAL HOUSEHOLDS' CHILDREN IN SECONDARY SCHOOLS
IN OYO STATE, NIGERIA
. RAHJI M.A.Y ** AND FAKAYODE S.B. *.
**Department of Agricultural Economics, University of
Ibadan,Ibadan.
* Department of Agricultural Economics and Farm Management,
University of Ilorin, Ilorin.
Abstract
The study examined school enrolment and gender gap for rural household
children at the secondary levels. A multi-stage sampling technique was used
in data collection. Probit model analysis was used in analyzing the data set.
Evidence from the analysis indicated that more boys were enrolled than girls.
The factors considered differently affect the children. Father's education
variable is significant for boys. This variable is marginally significant for
girls. The mother's education variable though positive in both cases is not
significant. The probit model predicted a gender gap of 18.72. The results
indicated that most of the gap is due to differences in the ways households
perceive male and female children. There is a preference for boys over girls
in secondary school enrolment. Based on the findings of this study, incentives
for the enrolment of girls were recommended. These include: differential fees
or free tuition, and increased public subsidies for female education at this
level.
Key Words: School enrolment, gender gap, rural children, Oyo State, Nigeria.
Introduction
Enrolment in schools represents the largest component of the investment in
human capital in most society (Schultz, 2002). Education, schooling and
human capital development are often used interchangeably in the literature.
The human resources of a nation are considered to be the engine of growth of
that country. These must however be adequately developed and efficiently
utilized. Education bestows on the recipients a disposition for a life-long
acquisition of knowledge, values, attitudes, competence and skills (Aliu,
2001). Hence, rapid socio-economic development of a nation has been
1
observed to depend on the calibre of human capital in that country.
Education is thus central to the development process.
The pivot of the educational policy thrust of the Federal Government of
Nigeria is the Universal Basic Education (UBE) scheme launched in 1999. It
is meant as a strategy to reduce illiteracy in the country and to enable the poor
to be educated out of their poverty. The decision to extend the scheme to the
junior secondary school is a welcome development. One, it is in line with
international standard. Two, primary education is no longer considered
adequate in a fast changing high information technology (HIT) world.
The decline in economic activities since the early 1980s has made education a
luxury to many Nigerians especially in the rural areas. Enrolment in
secondary schools is of policy significance and relevance in Nigeria. This is
because parents are known to invest in their children according to sex, birth
order or natural endowments. Under such, situation, children are valued by
parents for different reasons and girls and boys are not exact substitutes.
According to the Central Bank of Nigeria (CBN), 2000), total enrolment of
boys and girls into secondary schools in 1999 stood at 6.095million. In 2000,
this number increased by 5.0 percent to 6.4 million. The percentage of girls in
secondary schools in 1999 was 45.0 percent about 2.743million. This
increased to 46.0. about 2.944million in 2000. The total number of secondary
schools increased by2.0 percent from about 8113 in 1999 to 8275 schools in
2000.According to CBN (2000), the gender gap in literacy rate at the rural
level between boys and girls was 18.3 percent in favour of the males overall.
However, in the age group 10-14 years (secondary school ages) it was 6.5
percent in favour of boys. This indicates that there is a gender dimensions to
educational attainment and development in Nigeria.
There is a large literature on education in developing countries. The focus in
some cases is on the market-determined value of education. This is seen as
the major regulator of individual demand for education (Handa, 1996). There
is, however, an equally important non-market component to the demand for
education especially for rural children (Colclough, 1994). These non-market
factors manifest themselves through households` characteristics that affect
the time and opportunity cost of schooling.. There is thus the research need to
assess the rural households` demand for schooling for its children. The aim
is to ascertain the factors influencing school enrolment in the rural areas..
There is also the research need to ascertain the gender gap in school
enrolment for secondary age children in the area of study. The aim of this
attempt is to contribute to the literature and knowledge in terms of : who is
2
schooled in the rural Nigeria? What is the role of household characteristics in
children's enrolment? Are the factors affecting enrolment the same or are
different for male and female children? What is the nature of the gender gap
in access to education by rural children? These queries motivate the quest for
empirical verification and quantification of the issues raised with respect to
Oyo state, Nigeria.
The general objective of this study is to examine secondary school enrolment
for rural households` children in Oyo State, Nigeria. The specific objectives
are
to
identify the factors influencing secondary school enrolment of rural children,
estimate the gender gap in school enrolment of secondary school age rural
children, assess the effects of some policy factors on school enrolment,
examine the gender dimensions to secondary school enrollment in the area of
study, and make recommendations based on the findings of this study.
Methodology
Oyo State is located in south -western Nigeria. It lies within latitudes 7o 5' and
9o 10' N of the equator and longitudes 2o 38' E and 4o 35' E of the Greenwich
Meridian. Farming is the main occupation of majority of the people in the
state. The state is strictly agrarian and rural based.
South -western Nigeria has an admirable record in human resource
development in Nigeria. Free Primary Education was introduced into the area
in 1955 by the then regional government. Years later in 1979, Universal
Primary Education (UPE) was launched in the area again at the regional level.
In 1999, the federal government of Nigeria introduced a nation-wide
education programmes called Universal Basin Education (UBE). The efforts
of the past and the current impetus call for the quantification of the effect of
these programmes on school enrolment especially in rural South-western
Nigeria.
A multistage random sampling procedure was used in data collection. The
state had been divided into eight agricultural zones. These are the Eruwa
(Ibarapa), Iseyin, Kishi, Saki, Okeho, Oyo, Ibadan and Ogbomoso. These are
taken as the sampling units and constitute the first stage of sampling.
In the second stage, the list of all the villages within each zone was obtained
from the Oyo State, Ministry of Local Government, Secretariat, Ibadan. For
Ibadan and Saki zones, 6 villages were randomly selected from the list. The
case for Ibadan zone is based on its large population while that for Saki zone
is based on its geographical spread. In each of the remaining 6 zones 5
villages were also randomly selected. Forty two (42) villages were thus
3
selected for the study.
Six (6) farm households were randomly selected in each village as a third
stage of sampling. Two hundred and fifty two (252) households were
sampled. The sampling procedure yielded 461 secondary school age boys and
400 secondary school-age girls.
The model used in this study borrows from the work of Handa (1996). This
is represented as
PrEi  1 X i  
1
2
zi
e

ud 2
2
(1)
and
Zi = bo + b1 X1 + b2X2 + .... + bnXn
(2)
for i = male(boys) (m) or female(girls) (f).The probit model can be
written as
Pr(Ei = 1/Xi) = Φ(XiBi)
(3)
where Ei indicates whether a child of sex i is enrolled in school, Xi, is the
vector of observed individual and household characteristics, Bi is the vector
of coefficients and Φ is the standard normal cumulative density function.
The gender gap in school enrolment is decomposed using the estimated probit
coefficients (Even and Macpherson, 1993). The predicted enrolment of sex,
m, for boys is
P(X m , β̂ m ) 
1 Nm
 φ(X m , β̂m )
N m i 1
(4)
and sex, f, for girls is
P(X f , β̂f ) 
1
Nf
Nf
 φ(X , β̂
f 1
f
f
)
(5)
The gender gap (GAP) in enrolment predicted by equation (3) is the
difference in the predicted enrolment rates for the boys and the girls.
GAP  P(X m , β̂ m )  P(X f , β̂ f )
4
In this specification, girl- children are the reference group. This is why the
predicted enrolment rate for the girls is used in the decomposition procedure.
The gender gap is decomposed into three components. One is an explained
portion of the gap (Exp). This is due to the difference in the explanatory
variables. It is obtained by estimating the enrolment rate that girls would have
if they had the boys’ characteristics or explanatory variables. Hence;
EXP  P( X m ,  f )  P( X f , ˆ f )
Two, the unexplained portion of the gap (UEX) is the change in enrolment
that would occur if the probability of female enrolment were determined by
the male children's coefficients.
UEX  p( X f , ˆm )  P( X f , ˆ f )
Three is the residual (RES) gap of the total enrolment gap. This forces an
adding up constraint that is represented as
GAP = EXP + UEX + RES
or

RES = GAP - EXP - UEX
The residual gap indicates by how much the estimated enrolment gap would
change if the male children are used as the reference group (Even and
Macpherson, 1993). Its value is known to be equal in size but opposite in sign
when the reference group is switched.
The inference that can be drawn from the interaction terms are: positive
coefficient on the distance and parental education variables indicate that the
joint effect of the two factors act as substitutes in school enrolment. Positive
sign on the parental education variables signifies that the parents' education
act as complement in school enrolment decisions ( Handa, 1994, 1996;
Thomas et al; 1991; Barrera,1990; and Strauss, 1990).
It is hypothesized in this study that age of head (X1), father's education (X2),
mother's education (X3), schooling cost (X4), log per capita income (X5),
distance to school (X6), gender of household head (X7), coop membership
(X8), household net-worth (X9), which is a proxy for wealth, father's
education, mother's education (X10), father's education- distance to school
5
(X11), and mother's education- distance to school (X12). The selection of these
variables is based on economic theory and as suggested by previous / similar
studies. (e.g. Handa1994; 1996).
Results
The results of the estimated probit models are presented below. The Pearson
chi-square statistics and the default value (P) were used to measure the
goodness of fit of the models. The calculated chi-squares (χ2) of 306.7634 for
boys and 312.3529 for girls are greater than the tabular values of 135.81 and
124.340 at the 1% and 5% levels of significance respectively. The P values of
0.6782 for boys and 0.6038 for girls indicate good fit for the models. The
acceptable default value is P > 0.15. This signifies a good fit for the estimated
models relative to the data used. The default value is a diagnostic statistic it
has nothing to do with the levels of significance of the estimated parameters.
P<0.15 means that the data used did not fit the model and thus it is rejected.
6
Table 1: Probit: Secondary Enrolment: Boys
Variables
Age of Head (X1)
Fathers Education (X2)
Mother's Education (X3)
Estimated Schooling Cost (X4)
Log Per Capita Income (X5)
Distance to School (X6)
Gender of Head (X7)
If Female = 1
Otherwise = 0
Coop. Membership of Head (X8)
Parameters
t-values
-0.1675
(0.1142)
+0.2401**
(0.1193)
0.1067
(0.0830)
-0.3653**
(0.1433)
+0.7528***
(0.2231)
-0.4250
(0.4239)
-0.2034
(0.1609)
1.4667
+0.5716***
(0.2210)
+0.7147***
(0.2358)
+0.4538**
(0.2100)
+0.3061**
(0.1372)
+0.2873
(0.2910)
3.1354
Household Net worth (Wealth) (X9)
Father's Education, Mother's Education (X 10)
Father's Education, Distance to School (X 11)
Mother's Education, Distance to School (X 12)
Intercept (K)
N
= 461
Enrolled = 296
Not-enrolled
= 165
Pearson χ2
= 306.7634
Likelihood ratio= 306.7634
Log likelihood = -147.2825
Default (P)
= 0.6782
Model's Predicted probability = 89.56
7
2.0126
1.2855
2.5492
3.3743
1.0025
1.2641
2.5864
3.0310
2.1610
2.2311
0.9873
-
Table 2: Probit: Secondary Enrolment: Girls
Variables
Parameters
t-values
Age of Head (X1)
-0.2132
(0.1708)
-0.1984 *
(0.1058)
0.1433
(0.0935)
-0.5671***
(0.2172)
+0.6495***
(0.2198)
-0.4108
(0.2857)
-0.3216***
(0.1248)
+0.4570**
(0.2241)
+0.3659**
(0.1436)
+0.261*
(0.1289)
+0.1734
(0.1305)
+0.2053
(0.1649)
2.6847
1.2482
Fathers Education (X2)
Mother's Education (X3)
Schooling Cost (X4)
Log Per Capita Income (X5)
Distance to School (X6)
Gender of Head (X7)
Coop. Membership of Head (X8)
Household Net worth (X9)
Father's Education, Mother's Education (X 10)
Father's Education, Distance to School (X 11)
Mother's Education, Distance to School (X 12)
Intercept (K)
N
= 400
Enrolled = 167
Not-enrolled
= 233
Pearson χ2
= 312.3529
Likelihood ratio= 312.3529
Log likelihood = -115.6236
Default (P)
= 0.6038
Model's Predicted probability = 69.84
8
1.8752
1.5326
2.6110
2.9550
1.4379
2.5769
2.0393
2.5481
1.7541
1.3287
1.2450
-
Table 3: Results of the Decomposition of Gender Gap in Secondary School
Enrolment.
Secondary
Items
Value
%
16.99
Gap due to X's
3.18
Gap due to Bm
14.96
79.91
Gap due to Interaction
0.58
3.10
Total
18.72
100.00
Source: Field Survey, 2003.
NB: Bm is the vector of parameters for the boys.
Table 1 contains the results for the boys. It shows that father's education (X2),
log per capita income (X5), cooperative membership of heads (X8), household
net-worth (X9), father-mother education term (X10) and father education are
positive and highly significant some at 1%, others at 5%. In this model, it is
only the cost of schooling (X4),that is negatively significant at the 5% level.
Table 2 shows that father's education (X2), log per capita household income
(X5), cooperative member-ship of heads (X8), household net-worth (X9), and
father-mother education term (X10) are positive and significant though at
different levels. The cost of schooling (X4), and the gender variable (X7) are
negative and significant at the 1% level. The log per capita household has a
bigger effect on the probability of enrolment for boys than for girls. This
tends to imply that income constraints are more binding for boys than girls.
In examining the impacts of the significant variables on enrolment decisions,
the signs of the parameters are used. A positive sign indicates that higher
values of a variable increase the likelihood of enrolment. Similarly, a
negative sign implies that higher values of such a factor reduce the
probability of enrolment. Over all, father's education variable is significant
for boy. It is marginally significant for girls. The mother's education variable
though positive in both cases is not significant. This finding may be due to
the lower level of literacy attainment by females in the area. Rahji (1996)
found that adult males had an average of 6 years of formal education while
adult female attained an average of 3 years. This result tends to confirm the
9
findings by Thomas et al; 1991 and Handa, 1996.
Table 3 contains the analysis of the gender gap in enrolment. Out of the
18.72 gap, the explained portion is 3.18 (20%) of the overall. The
unexplained gender gap is 14.96 (79.91%) of the entire gap. The difference
due to residual is 0.58 (3.10%) of the entire gap. This result indicates that
most of the gender gap in school enrolment is due to differences in the ways
households perceive male and female children.
In terms of gender dimensions, the result indicates a preference for boys over
girls in secondary school enrolment of rural household children. It implies
that boys are more favoured to attend schools than girls. The reasons that can
be adduced for this could be found in cultural/ traditional norms and beliefs
of the people. These, see girls as belonging to their husbands' household later.
Boys are seen as the heir-apparent on whose shoulders rest the responsibility
of continuity of the lineage.
The father -mother education variable is positive and significant in both
cases. This factor thus acts as complement in enrolment of the children. It
thus has a significant complementary impact as a determinant of secondary
school enrolment. The father education- distance to school variable is
positive. It is significant at the 5% for boys. It is not significant for girls. The
positive sign indicates that the variable acts as substitutes in the decision for
school enrolment.
Conclusion
The study identified and estimated the gender gap in secondary school
enrolment in Oyo State, Nigeria. The results indicated a preference for boys
over girls in school enrolment. The difference in enrolment rates for the sexes
was found to be due to the ways the households’ perceive and treat male and
female children. It is hoped that the results of this study will be useful to
scholars as well as policy makers in formulating measures that will correct
the inequality in school enrolment in the area of study in particular and
Nigeria in general.
Recommendations
Gender gap was confirmed in favour of boys in the enrollment of secondary
school age children. Most of this gap is explained by the ways rural
households treat male and female children.
There is a need for awareness campaign to off set this preferential treatment
within the household. The importance of educating the girl-child and its
likely contribution to the development of the country should be the focus of
10
campaigns. In Nigeria the population of females is greater than for males.
There is a need for subsidizing female education. Differential fees or free
education for girls as recently done in the Gambia is recommended. Hence,
there is the need to understand more precisely how households respond to
different inducement.
Cost of schooling is negative and significant. The implication is that the
probability of enrolment decreases as this factor increases. Increase in the
supply of secondary schools for girls by deliberate government policy is also
recommended. The distance to school variable is negative in the models.
Hence, the longer the distance the lower is the likelihood of enrolment.
Distance to the nearest school is used as a measure of access (Casterline,
1987). Schools should be located as near as possible to the users if the rate of
enrollment desired is to be attained. The shorter the distance to school,
ceteris paribus, the higher is the probability of enrolment for rural household
children.
11
References
Aliu, S, (2001). "The Competitive Drive, New Technologies and
Employment: The Human Capital Link". A Paper presented at the
Second Tripartite Conference of Manpower Planners. Chelsea
Hotel, Abuja.
Barrera, A. (1990). "The Role of Maternal Education and Its Interaction with
Public Health Programmes in Child Health Production" Journal of
Development Economics, Vol. 32, pp 69-91.
Casterline, J.B. (1987). "The Collection and Analysis of Community Data" in
Assessment (ed) Cleland J. and Scott, C. Oxford University Press pp
882 - 905.
CBN (2000). Annual Report and Statement of Accounts 31st December,
2000.
Colclough, C. with Lewin, K. (1993). Educating All the Children: Strategies
for Primary Schooling in the South. Oxford Clarendon Press.
Even, W. and Macpherson, D. (1993). "The Decline of Private Sector
Unionism and the Gender Wage Gap" Journal of Human Resources,
Vol. 28. No. 2 pp 279-96.
Handa, S. (1994). "More Evidence on the Role of Maternal Education in the
Production of Child. Height" Department of Economics, University
of the West Indies - Mona.
Handa, S. (1996). "The Determinants of Teenage Schooling in Jamaica: Rich
Vs Poor, Females Vs Males. The Journal of Development Studies,
Vol. 32. No. 4 pp 554 - 580.
Rahji, M.A.Y. (1996) An Analysis of Farm Household Economic Behaviour
in Oyo State, Nigeria. Unpublished PhD Thesis Dept of Agricultural
Economics, University of Ibadan, Nigeria.
Schultz, T.P. (2002). "Why Governments should Invest More to Educate
Girls" World Development, Vol. 30 No.2 Pp 207 - 225.
Strauss, J. (1990). "Households, Communities and Preschool Children's
Nutrition Outcome: Evidence from Rural Cote D'Ivoire" Economic
Development and Cultural Change, Vol. 38, Pp 197 - 234.
Thomas, D. Strauss, J; and Henriques, M.H. (1991). "How Does Mother's
Education Affect Child Height? Journal of Human Resources,
Vol.26. No. 2 Pp 183 - 211.
12