This thesis conducts cross-country analyses using data from all inhabited
continents to examine the support of common expectations based on either Neoclassical
Economics or popular beliefs. The first two chapters use SACMEQ data
from sub-Saharan Africa.
The first chapter argues that changes in class size trigger a number of mechanisms
affecting how the pupils’ household, school leaders, teachers and peers behave.
These behaviours are highly context-specific and may counterbalance or exacerbate
one another. It finds that the main threat to a pupil’s achievement is sharing the
teacher with more peers, but that household behaviours can mitigate or even
outweigh this threat.
The second chapter examines the conditional correlation of observable teacher
characteristics and pupil achievement. It argues and demonstrates that previous
research using the same data does not sufficiently address the teacher-pupil
matching problem and that lacking to do so leads to very different conclusions. The
chapter categorises the available observable teacher characteristics as proxies for
either subject-matter or pedagogic competency and examines their
complementarity by adding interactions between the individual proxies of these
two competencies. The evidence suggests these two competencies are substitutes in
six of ten countries.
The third chapter uses OECD TALIS 2013 data to explore the connection between
teachers’ workload and their job satisfaction. It applies a production function
approach that combines both Top-down and Bottom-up approaches. It finds that
the effect of teachers’ workload measured in hours is negligible. But evidence of the
effect of teachers’ perceptions of their workplace from the English sub-sample
provides clear evidence that the workplace matters.