Page 206 - Contributed Paper Session (CPS) - Volume 3

CPS1996 Ali Hebishi Kamel A. et al.
individual have completed the secondary education or higher and value 0
indicate that individual is not educated or under secondary education.
Education is the main variable in our study since higher levels of education
means higher human capital and thus higher income level.

Binary Logistic Regression
[Z= 10.29154 +0.93101*lnTot_Inc + 3.02868 M_EduSt1 -5.08226 poor1
+1.82298 gander -2.50622 Urbrur1 -0.55383 lnTot_Inc:poor1 +0.37582
poor1:gender1 -0.91931 M_EduSt1:gender1 + 0.28444 lnTot_Inc:Urbrur1
-0.24310 gender1:Urbrur1 + 0.22899 lnTot_Inc:gender1]

π
Where Z = log ( ) where  is completed secondary education, The
1− π
Predicator here is log-transformed of income so we can conclude from the
model that for one unit income increase is associated with an increase in odds
of being completed secondary education than being under secondary
completed. All independent variables have significance level then their
parameters are different from 0. The parameters with significant negative
coefficients decrease the likelihood of that response category (completed
secondary education) with respect to the reference category. Parameters with
positive coefficients increase the likelihood of that response category.
B can be interpreted as the change in probability of being completed
0
secondary education is equal Z, if all independent variables = 0 in the model.
All independent variables in Education level model have statistically significant
relationship with the odds of dependent variable (Education level). Also all
interactions between independent variables are statistically significant.
B represents the difference in the probability of predicted variable
i
(completed secondary education) for each one-unit difference in X , if the rest
i
of independent variables remain constant.
B = 0.93101 represents that for one unit increase in log income value will
1
implies an increase of the log odds ratio of education status completed
secondary education versus under secondary education by 0.93101, if the rest
of independent variables are constant. Also this model reflects the strong
relationship between income and education.

3.3 The relation between ineqality income and disparities education
Is there a relation between inequality income and disparities education?
Disparities are differences. Because we are statisticians, we are interested in
differences that are statistically significant, or statistical disparities. So Disparity
analysis will focus on subgroups according to: Gender, Education status, place
of residence………etc.

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