School Participation in Rural Pakistan: A Situation Analysis Author: Haroon Jamal Factors Affecting School Participation A multivariate analysis is carried out by estimating logistic regression function for school participation of 5-14 age cohort children. All potential correlates of school participation are included in the logistic function to assess the probability and marginal effect on the household decision to enroll their children in school. The logistic function incorporates head of household and spouse characteristics, besides pertinent demographic, social, economic and locational factors. The summary statistics of the logistic regression indicate a good-fit of the model with a high percentage (75 percent) of correct predictions and expected signs of all coefficients associated with variables. Table 3.2 displays estimated coefficients, level of significance and marginal effect with respect to probability to enroll. Model summary statistics are also provided in the table. An important finding of this study is the significant role of female headed households in the decision to send children to school. The variable of female headship is appeared statistically significant with large marginal effect. Similarly, spouse education level is more effective than head of household in influencing decision to enroll. All variables of household wealth (house ownership, RCC roofing, ‘Pacca’ wall structure, household asset scores and livestock ownership) are statistically significant with high marginal effect on probability to enroll. Among these the impact of RCC roofing is relatively substantial. The supply side constraints are represented by distance to primary and middle school, while ‘time to reach public transport’ is used as a proxy for village remoteness. All these proxies are negatively correlated with the enrolment decision. The large negative marginal effect is associated with the variable ‘Time to reach Primary School’. To capture the provincial and regional differences in terms of population and level of rural development, locational (province and South Punjab agro-zones) dummy variables are incorporated in the model. All these appeared significant with signs according to a priori expectation. The negative coefficients with high marginal effect of rural Sind and Balochistan with reference to Punjab confirm the descriptive information presented above. Similarly, variables representing districts of south Punjab, which have relatively high percentage of poverty incidence, are showing large negative impact on the probability for school enrolment. spdc.org.pk/Publication_detail.aspx?sysID=767
Posted on: Tue, 02 Dec 2014 09:45:49 +0000
Trending Topics
Recently Viewed Topics
© 2015