| dc.description.abstract | This study was aimed at investigating the role of socioeconomic
and cultural factors and isolate the potential
determinants of sex mortality differential during the childhood
period at the national level in Kenya. The major objective was to
explore the variation of mortality by sex under different
categories of selected socio-economic and cultural variables from
the 1993 Kenya Demographic and Health survey data (KDHS 1993) I and
then asses and isolate the potential determinants of the sex
differences in mortality.
The socio-economic and cultural variables investigated
include: level of education of mothers place of residence of
mothers marital status of mothers current type of employment of
mothers
region of residence of mothers I ; the religious
affiliation of mothers I and the ethnic origin of mothers I • Each
variable was splitted into different categories.
The Trussell variant of the original Brass method was used to
determine infant and child mortality levels whereas multivariable
regression method was employed to isolate the potential
determinants of mortality at the childhood period by sex.
In general the factors investigated were found to contain
large variation in infant and child sex differences in mortality
rates. Excess female child mortality rates were observed among
'children of mothers: with no education; residing in urban/areas;
who are working away from home; who are widowed or divorced;
residing in Nyanza and Central provinces; with Catholic religious
affiliation; and, from Nilotic ethnic origin. Among the categories
Nyanza province recorded the highest excess female child deaths
whereas Eastern province recorded the highest male child deaths.
Although the direction of childhood mortality change among the
various socio-economic groups was found to be the same for males
and females, the magnitudes of the change were found to vary by
sex.
In the multivariate analysis the proportion of variance of the
dependent variable explained by the selected socio-economic and
cultural variables was found to be relatively small; and this was
attributed to, the non linearity of some relationships, the high
random variation that might exist while using a mortality indicator
as dependent variable and the non inclusion of some important
variables related to individual child risk of death. However, the
F-test values of the analysis confirmed that the model was
statistically significant. Education and ethnicity variables were
found to explain most of the observed variation. In all, level of
maternal education, type of employment of mothers', marital status
of mothers' and the ethnic origin of mothers' were found to be
important determinants of the proportion of children dead of both
the sexes. In addition urban place of residence was found to be a
potential determinant of the proportion of male children dead
compared to rural place of residence. Primary and secondary and
above level of maternal education and other Bantus ethnic origin of
mothers' were found to have reduction effect on child mortality.
Children of mothers working away, working at home, widowed and
divorced have relatively higher mortality rates when compared with
the selected background reference categories for both the sexes.
Urban place of residence, in the case of male children was. found to
have a reduction effect on child mortality.
In view of the results of the demographic and statistical
analysis this study recommended that policy makers and planners
have to integrate gender issues in development planning so as to
facilitate equal opportunities for both the sexes at any level of
socio-economic strata and In any cultural set-up. It also
recommended deep and comprehensive studies to be conducted in this
area. | |
