While much research has been conducted on the influence of individual and family characteristics on social exclusion, very little has examined the role of community and neighbourhood factors. The effects of neighbourhood characteristics on young people's educational outcomes at age 15, 17 and 19 years in Australia, using data from the Longitudinal Surveys of Australian Youth (LSAY), are investigated in this report. The results indicate that outcomes are likely to be much the same for two students with similar individual and parental characteristics attending the same school but living in neighbourhoods with different levels of socioeconomic status. Mentoring efforts that help shape aspirations of young people at an early age could have a high payoff in terms of post-school outcomes.
About the research
While much research has been conducted on the influence of individual and family characteristics on social exclusion, very little has examined the role of community and neighbourhood factors. This project is considering the differences in education and training outcomes in areas of social advantage by comparison with areas of social disadvantage, taking the contribution of these neighbourhood factors into account. This report contains the results of the quantitative aspect of the study, using data on individuals and their families from the Longitudinal Surveys of Australian Youth (LSAY) and the 2006 census. A second part to the project is a qualitative study exploring the influence of access to high-quality education and training and other community infrastructure on education and training outcomes. The results of this study, in which regions in Victoria and South Australia are compared, will be available in 2014. This work is one of three projects undertaken by the Centre for the Economics of Education and Training (CEET) at Monash University, as part of its three-year (2011—13) research partnership with the National Centre for Vocational Education Research (NCVER) exploring the geographical dimensions of social inclusion and vocational education and training (VET) in Australia.
- The socioeconomic status of a neighbourhood is an important characteristic in explaining variations in student outcomes, but residential turnover, the composition of households and the multicultural nature of the neighbourhood also play a role.
- The characteristics of schools make an important difference, but in reality data for many of these (for example, school leadership and teacher quality) are either not readily available or the characteristics are not easily measurable. The effects of a neighbourhood are sometimes difficult to separate from the impacts of schooling because of the correlation between the two.
The authors argue for caution when inferring the significance of the relationship between neighbourhood characteristics and student outcomes if the model estimating such a relationship does not contain the appropriate controls for school effects.
Prior aspirations are important in predicting the final post-school destinations of young people. The results suggest that mentoring efforts that help to shape the aspirations of young people at an early age could have a high payoff, in terms of their post-school outcomes.
Managing Director, NCVER
Understanding the factors that drive socioeconomic inequalities in educational and training outcomes in Australia is important for designing appropriate policy responses. One key aspect is being able to identify the importance of neighbourhood socioeconomic characteristics in accounting for observed differences in outcomes.
In this study we investigate the following two questions:
- Whether post-school outcomes (in terms of early school leaving and participation in further education and training) for students from disadvantaged areas are similar to those for their counterparts from advantaged areas?
- Which socioeconomic characteristics of the local area contribute most to inequality in students’ post-school outcomes?
The study uses data from the 2003 cohort of the Longitudinal Surveys of Australian Youth (LSAY), which includes rich information on the characteristics of young people who were 15 years old in 2003, the school they attended and their experience of transition from school to post-school destinations. The 2003 cohort was drawn from the same sample of 15-year-olds in Australian schools who participated in the 2003 Programme for International Student Assessment (PISA), conducted by the Organisation for Economic Cooperation and Development (OECD), which means a much richer dataset is available for analysis. These data are augmented by the characteristics of neighbourhoods at the postcode level derived from the 2006 Australian Census to enable investigation of the effects of neighbourhood characteristics on young people’s schooling outcomes.
The data enable us to determine whether neighbourhood characteristics are important for student outcomes, after controlling for individual, parental and school effects.
The report focuses on the effects of four neighbourhood characteristics — neighbourhood socioeconomic status, residential stability, household type and ethnic diversity — measured at the postcode level on student outcomes at ages 15, 17 and 19 years.
Various models are estimated to provide detail in the way neighbourhood characteristics affect student outcomes. In the first set of models, only neighbourhood characteristics are included as control variables. The second set includes controls for individual characteristics (for example, sex, place of birth, Indigenous status, whether only child in family etc.) and parental characteristics (labour force status, occupation and education level of father and mother, family ethnicity etc.). The third set includes school fixed effects to control for the average observed and unobserved differences in the quality of schools (for example, class sizes, quality of teachers, facilities and peers etc.) and in the differences in the ‘quality’ of the student intake (for example, student ability, parental background etc.). It is uncommon in the literature to find studies that have simultaneously controlled for neighbourhood and school effects in the same model. The fourth set of models, which are estimated only for student outcomes at age 17 and 19 years, includes controls for the prior attributes of a student (achievement, attitude, aspirations and application) measured at age 15 years.
Outcomes at age 15 years
The determinants of four student outcomes at age 15 years are investigated: achievement; attitudes (towards schooling); aspiration (to complete a vocational education and training (VET) or university qualification); and application for school (time spent on homework).
The neighbourhood characteristics generally have significant effects on all student outcomes, except student attitudes, even after individual and parental factors are controlled for. The results suggest that students living in less prosperous neighbourhoods are more likely to have aspirations for undertaking vocational education and training.
When school fixed effects are introduced into the model to control for the variation in the characteristics of the school that students attend, the neighbourhood effects are generally no longer significant. This suggests a strong correlation in the observed characteristics of the neighbourhood where the student lives and the observed and unobserved characteristics of the school the student attends.
Neighbourhood stability exerts an independent positive influence on students’ achievement, even after controlling for school fixed effects. A neighbourhood with low residential stability signifies a high transient population and a low level of home ownership. In a neighbourhood with high residential stability, the turnover among the students’ peer group in the neighbourhood as well as in the school is likely to be less, which will mean students will have a less disruptive social life and will be able to focus more on learning tasks.
Many individual and parental factors have significant independent influences on student outcomes after controlling for neighbourhood and school fixed effects, many of which have also been reported elsewhere.
Outcomes at age 17 years
The study also investigates the factors that affect whether a student at age 17 years continues to be engaged with some type of education and training (school, vocational education and training or university) or not. In the full model, which includes observed individual and family characteristics, school fixed effects and student attributes measured at age 15 years, neighbourhood characteristics have little influence on outcomes. Any positive effect from living in an advantaged area appears to work through the fact that these areas generally have ‘better’ schools, which promote further education and training. All student attributes measured at age 15 years however exert significant independent effects on the outcome.
Outcomes at age 19 years
Finally, the study considers the factors that affect student outcomes at age 19 years — VET outcome, university outcome or non-study outcome. As at age 17 years, neighbourhood characteristics are not significant in predicting outcomes at age 19 years in the full model.
Prior aspirations are important in predicting the final post-school destinations of young people, with VET aspirations significant in predicting a VET outcome and university aspirations for a university outcome. The results suggest that mentoring efforts, which help to shape the aspirations of young people at an early age, could have major implications later in terms of better post-school outcomes. The result is of particular relevance for young Indigenous children, who often lack appropriate adult role models. Prior achievement scores and attitude to school have positive independent effects on university outcomes but negative effects on VET outcomes.
After controlling for individual and parental influences, the study found that four neighbourhood characteristics have small but significant effects on student outcomes. While neighbourhood socioeconomic status is probably the most important of these characteristics, the other three also have a role in explaining the variations in student outcomes.
The effects of neighbourhood characteristics almost all disappear when school fixed effects are introduced into the model. The results underline the difficulty in inferring the significance of neighbourhood characteristics in predicting student outcomes when the model estimating the relationship does not contain appropriate controls for the school context.
The school effects in our models are fixed and capture all the differences, observed and unobserved, in school characteristics. But because the characteristics of neighbourhoods where students live are generally mirrored in the composition of schools, these are also captured by the school fixed effects. In the models reported here neighbourhood characteristics are, therefore, probably mediated through school fixed effects. Separating the neighbourhood effects from school effects is thus complicated and requires detailed measures of both the school and neighbourhood contexts. School-level measures, including school leadership and teacher quality, that provide information on the quality of the school could be useful in this respect, but are not always available or readily measurable.
The school fixed effects, however, allow for the identification of neighbourhood effects from differences in outcomes between students at the same school who live in different neighbourhoods. The results thus indicate that outcomes are unlikely to be significantly different for two students with similar individual and parental characteristics attending the same school, but living in neighbourhoods with different levels of socioeconomic status.
A possible conclusion from this research is that the inequalities in student outcomes may be reduced by a better allocation of resources to schools. The evidence for this is, however, indirect and does not provide guidance on what aspects of school quality and resources are most likely to make a difference. Changing the ways principals and teachers are allocated to schools and ensuring a school’s student intake is not disadvantaged through the selection practices of other schools could also help to reduce disparity in student outcomes.
The analysis provides information for the average case. A closer examination of the data shows that a number of neighbourhoods with below-average socioeconomic status have above average student outcomes, and vice versa. While the current data allow us to identify these neighbourhoods, they are inadequate for identifying factors that may explain student outcomes in these neighbourhoods. In the second part of this research project we use qualitative methods to identify possible critical factors that may be responsible for such divergent results in a selected number of neighbourhoods in Victoria and South Australia.