de.mpg.escidoc.pubman.appbase.FacesBean

Post

 
 Vis
  Gender Disparities in Educational Outcomes
Item is

Ophav

 Ophav:
Alsing, Julie1, Forfatter
Jones, Edward Samuel 2, Vejleder
Tilknytninger:
1Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7001              
2Økonomisk Institut, Det Samfundsvidenskabelige Fakultet, Københavns Universitet, København, Danmark, diskurs:7014              
skjul Ophav
Vis Ophav

Indhold

Ukontrollerede emneord: Gender disparities, Education, East Africa, Fixed effects
 Abstract: In East Africa enrolment rates have increased remarkably and a shift in attention towards educational attainment has therefore emerged within recent years. In many countries large increases in enrolment have come at a cost in quality, and although many children today are enrolled in school it is more questionable whether they actually learn.
This thesis explores the two dimensions, equality and education, and investigates whether gender disparities in educational attainment pose a risk in East Africa. More specifically, whether these exist on overall levels and whether they are persistent for a number of subgroups of the population.
The analysis is based on existing relevant literature , human capital theory and a survey data set named Uwezo. Data contain almost 350,000 observations for children in Kenya, Tanzania and Uganda and have information on child-, parental-, household-, village-, and school characteristics. The surveyed children have been tested in both literacy and numeracy and a unique feature is the fact that data are non-school based and therefore hold test scores for both enrolled and non-enrolled children. With this data structure a common issue in educational studies, selectivity in schooling choices, is avoided.
Acquisition of cognitive skills will most likely be dependent on unobservable genetic ability and unfortunately, this is problematic to measure. A standard OLS model is non-applicable as it will provide non-consistent estimates due to the possible correlation. To correct for genetic ability the analysis adopts sibling data and applies a fixed effects model. Assuming that genetic ability is family specific applying a household fixed effects model will ensure consistent estimates of gender effects in learning as it exploits within family outcome differences for girls and boys.
The empirical analysis indicates significant gender differences in obtained test scores in all three countries on overall levels. Girls do on average achieve higher scores than boys however as these results are in line with the theory for children’s early cognitive development they do not necessarily indicate issues of gender disparities. Results furthermore indicate poverty disparities in children’s performance and an analysis of gender effects in these poverty disparities is carried out.
Results reveal significant gender disparities for children from families with high levels of poverty. Within non-poor families the overall result still holds however in poorer families the effect changes and boys perform better than girls on average, indicating gender disparities. This result is confirmed by a regional investigation of gender disparities driven by poverty.
Lastly, the heterogeneity analysis finds a significant, though relatively small gender disparity in educational attainment driven by age. For the younger part of the sample girls still achieve significant higher scores than boys however this effect changes around age 15 after which girls achieve significant lower scores than their counter boys.
A number of factors may influence validity of obtained results. The most severe threat in relation to internal validity is the assumption of similar genetic ability for siblings. It is conceivable that child specific elements affect ability and if this is the case the household fixed effects model will not handle all of the unobserved ability. External validity may be affected if the sample used is not representative for the population and investigating this issue suggests inconsistencies, although relatively small, between the sample and the population.
Concluding on this empirical analysis it seems reasonable to suggest that gender disparities driven by especially poverty is an issue in East Africa, most significantly in Kenya and Uganda. Results encourage a shift in educational-, and gender policies to target girls in poorer families specifically rather than girls in general.
skjul Indhold
Vis Indhold

Filer

Navn:
Master(11).pdf (Hovedtekst)
Bemærkninger:
-
Tilgængelighed:
Offentlig
Mime-type / størrelse:
application/pdf / 1020KB
Copyright dato:
2013-12-06
Copyright information:
De fulde rettigheder til dette materiale tilhører forfatteren.
skjul Filer
Vis Filer

Basal

Bogmærk denne post: https://diskurs.kb.dk/item/diskurs:59171:1
 Type: Speciale
Alternativ titel: Evidence from East Africa
Alternativ titel: Kønsforskelle i Uddannelsesresultater
Alternativ titel: Beviser fra Østafrika
skjul Basal
Vis Basal

Links

Vis Links

Detaljer

Sprog: English - eng
 Datoer: 2013-10-04
 Sider: -
 Publiceringsinfo: København : Københavns Universitet
 Indholdsfortegnelse: ABSTRACT i
PREFACE iv
LIST OF TABLES, FIGURES AND BOXES vii
PART I INTRODUCTION 1
1.1. Research question 4
1.1.1. Contribution to literature 4
1.1.2. Conceptual clarifications 5
1.2. Limitations of the paper 6
1.2. Reader’s guide 6
PART II EDUCATIONAL SYSTEMS IN EAST AFRICA 7
2.1. Tanzania 8
2.2. Uganda 9
2.3. Kenya 10
PART III ECONOMIC LITERATURE 11
3.1. Schooling in developing countries 11
3.2. Gender differences in early child development 13
3.3. Gender disparities in schooling in developing countries 14
PART IV THEORY OF SCHOOLING 18
4.1. Human capital theory 18
4.2. Hypotheses 21
4.3. Summary of empirical and theoretical background 22
PART V DATA AND DESCRIPTIVE STATISTICS 24
5.1. Uwezo design 24
5.2. Included variables 26
5.2.1. Score measures 26
5.2.2. Independent variables 27
5.3. Descriptive statistics 30
5.3.1. Graphical analysis 34
5.3.2. Test of girls against boys 35
PART VI EMPIRICAL STRATEGY 37
6.1. Empirical model 37
6.2. Possible solution to OVB 41
6.3. Methodology 44
PART VII RESULTS 46
7.1. Gender disparities in educational outcomes in East Africa 46
7.2. Heterogeneity analysis 49
7.3. Sensitivity of results 57
7.4. Summary of analysis 60
PART VIII DISCUSSION 61
8.1. Methodology 61
8.2. Results 63
PART IX CONCLUSION 69
REFERENCES 71
APPENDIX 76
LIST OF TABLES
Table no. Title Page
Table 5.1 Description of variables included 29
Table 5.2 Observations in data 31
Table 5.3 Enrolment rates for primary and secondary combined 31
Table 5.4 Gender specific descriptive statistics for scores 32
Table 5.5 Descriptive statistics for variables included 33
Table 5.6 Gender differences in unconditional sample means 36
Table 7.1 Results for test score estimation 48
Table 7.2 Results for country specific test score estimation 51
Table 7.3 Estimation for test score with interaction for age and poverty 54
Table 7.4 Estimation for test score with interaction for regions 57
Table 7.5 Sensitivity tests for sample choices, sel. results 58
Table 7.6 Sensitivity tests for model choices, sel. results 59
Appendix
Table B.1 Descriptive statistics for test scores, Kenya 77
Table B.2 Descriptive statistics for test scores, Tanzania 77
Table B.3 Descriptive statistics for test scores, Uganda 77
Table C.1 Descriptive statistics for regions 78
Table F.1 Results for regions for table 7.1 and 7.2 81
Table G.1 Results for literacy estimation 82
Table G.2 Results for numeracy estimation 82
Table H.1 Results for interactions with age and poverty, remaining results 84
Table H.2 Results for interactions with regions, remaining results 85
Table I.1 Distribution of regions 86
LIST OF FIGURES
Figure no. Title Page
Figure 2.1 Educational systems 8
Figure 5.1 Score measures 27
Figure 5.2 CDF for overall test score 34
Figure 5.3 Mean of compl. pct. rate for overall test score by gender 35
Figure 7.1 Predictive margins for age interaction, 95 pct. CI 55
Figure 7.2 Predictive margins for poverty interaction, 95 pct. CI 56
Figure 7.3 Regional poverty levels 66
Appendix
Figure C.1 Mean of compl. pct. rate for literacy by gender 79
Figure C.2 Mean of compl. pct. rate for numeracy by gender 79
LIST OF BOX ES
Box no. Title Page
Box 1.1 Millinnium Development Goals 1
Box 1.2 Dakar Framework 2
Box 5.1 Uwezo 24
 Note: -
 Type: Speciale
skjul Detaljer
Vis Detaljer

Kilde

Vis Kilde