Determinants of Women’s Empowerment in Kenya: A Logistic Regression Approach.

Abstract

This study uses ‘basic individual and household characteristics’ to estimate the probability of a
woman's empowerment. Specifically, it will use the logistic regression model to analyse the
determinants of women's empowerment to understand the relationship between various factors and
the likelihood of women achieving empowerment. The study will utilise the 2014 Kenya
Demographic Household Survey (KDHS) data. The various domains of empowerment will be
considered. The domains identified in the Exploratory and Confirmatory Factor Analysis will be
assigned equal weights, which will then be distributed equally among each of their constituent
indicators. A woman will be considered to be empowered if her score is more than half the total
weighted score and not empowered otherwise. In the second analysis stage, basic household and
individual characteristics of women (independent variables) will be used to estimate the likelihood
of achieving empowerment using logistics regression. After fitting the model, post-estimation
analysis will be carried out. Specifically, the Hosmer-Lemeshow test will be used to measure the
model's goodness of fit. The study found that the odds of a woman being empowered are 175 times
higher if she has ever attended school compared to a woman who has never attended school if all
other factors are held constant. Women who have attained the primary level of education are much
less likely (98% less likely) to be empowered than women who have attained university education.
Women who listen to the radio at least once a week are 2.51 times more likely to be empowered
than those who do not listen to the radio. Women in the richest wealth quintile are 2.03 times more
likely to be empowered than those in the poorest. Women falling in the poorer wealth quintile
category are 27% more likely to be empowered than women in the poorest, with all other factors
constant