More complex modelling
6. More complex modelling
6.1. Introduction
Having reviewed the evidence at each level of our conceptual model, we now turn to
full ecological models of development. As outlined in the introduction, ecological
models view development as occurring within multilevel, multi-layered contexts.
Thus, as we have shown, while there are direct links from parents’ level of education
to distal family factors, key characteristics of the family, and proximal family
processes and in turn to child development, there are also important and complex
pathways within this conceptual model.
These complex pathways for the effects of education have not been extensively
modelled in peer-reviewed literature, although a number of authors are starting to
estimate such pathways. However, recently there has been an accumulation of
research looking at models of complex pathways for the influence of income on
children’s development and the processes and mechanisms that moderate and mediate
these effects. Therefore, to demonstrate the nature and utility of more complex
modelling techniques we use the example of income and review three papers that
have investigated the pathways through which income influences children’s
development. We also draw out the implications of this evidence for the role of
education. Future work at the WBL will also contribute to this research project.
Complex pathways such as this are primarily explored using structural equation
modelling (SEM) because it is able to investigate multiple mechanisms of influence
simultaneously. SEM allows for concurrent tests of all the associations between
constructs and the direct and indirect associations of all predictors can be assessed
while taking into account a variety of control variables. Note however, that these
SEM models do not get at causality in an econometric sense but decompose
covariation into active and inactive elements providing indications of meaningful
pathways for effects.
6.2. The model of mediating process for the effects of income
In section 5 we reviewed evidence showing the well established association between
distal factors such as income and the developmental outcomes of children and
adolescents. However, several authors such as Yeung et al. (2002) argue that there is a
substantial gap in the literature linking distal factors such as income and children’s
development in terms of identifying and understanding the processes and factors that
mediate this relationship. They posit that this gap limits our ability to answer
questions such as why income matters, when it might matter most, and why it has a
stronger effect on children’s cognitive achievement than on behaviour.
In relation to income researchers have started to examine the pathways through which
economic deprivation operates to disadvantage children of poor families. This
74