Explainer: Measuring Disadvantage at a Household Level
Written by Paul Koshy and Richard Seymour, NCSEHE Research Fellows
A key issue in Australian equity policy in higher education is the measurement of socio-economic disadvantage among students at all points of contact with the university: at initial access, in participation and performance, and/or in outcomes such as completion and post-graduation activities.
A recently released research paper by the Australian Bureau of Statistics (ABS) outlines a rationale and approach for constructing a household-level index measure of socio-economic disadvantage. The approach uses data from the General Social Survey (GSS) and results in an experimental measurement index.
On page 25, the authors note that “the interest in finer level indexes emerges from the fact that area based measures of disadvantage may not necessarily capture the disadvantage of those living within the areas.” This distinction is important to Australian higher education.
Currently, students are ‘assigned’ a socio-economic status (SES) ranking of ‘low’, ‘middle’, or ‘high’, upon enrolment, on the basis of an estimate of the SES of the geographical area in which their permanent residence is situated. Previously, this SES status was determined according to the postcode of the residence in question, but now SES is determined via the smaller Statistical Area Level 1 (or ‘SA1 area’).
The SES of an SA1 area is determined using its national ranking on the basis of the Socio-Economic Index for Areas (SEIFA) Index. SEIFA itself is based on ABS Census data of households in the given area. SA1 areas located in the bottom 25% of all Australian SA1 areas using the SEIFA index are nominated as low SES. Accordingly, students who live in areas defined as the bottom 25% are subsequently considered low SES students.
A recent study published by us and our NCSEHE colleague, Mike Dockery, uses data from the Household, Income and Labour Dynamics in Australia (HILDA) survey to undertake a similar exercise. We wanted to determine the socio-economic characteristics of households that underpin university enrolment. Students enrolled in university were classified into SES quartiles on the basis of the probability of attending university. The results were then compared to the students’ SES classification on the basis of the postcode of their residence.
Our research found that students are consistently misclassified under area measures, with the number of low SES students likely to be considerably smaller as a result. Given this finding, the development of a household measure may provide a more refined measure of socio-economic disadvantage on the basis of an individual student’s household characteristics.
The experimental index constructed by the ABS researchers using GSS data shows that it provides something of a valid measure at the household level. However, it does also require further testing in terms of variable selection, validity and the treatment of missing variables. The report is an interesting one and the research certainly very welcome, as its findings indicate that household measures of disadvantage could be constructed. A finer definition of socio-economic disadvantage would enable policy makers, and subsequently universities, to better target outreach, access and support resources.