10th Biennial ACSPRI Social Science Methodology Conference in conjunction with the Australian Social Network Analysis Conference (ASNAC)

The focus of administrative data collection for legal assistance services, as with many other sectors, is often about meeting reporting requirements of funding bodies. There are missed opportunities to use routinely collected data to answer research questions. This presentation draws on the Law and Justice Foundation of NSW's work with community legal centres across New South Wales. The project aims to describe how community legal centres statewide respond to legal need, where gaps in provision appear, and how service utilisation compares with available indicators of legal need. Following pilot analyses for a subset of centres, the current phase includes engagement with 32 community legal centres in NSW and unit record data for more than 100,000 services that were provided in 2024-25.

The methodological core of the project is an iterative quality assurance process conducted with each centre individually, in which data is reconciled against other sources and apparent anomalies are investigated. Each centre receives a draft report accompanied by queries about unusual values and questions about ordinary recording practice. Their responses inform a final report for the centre and influence how data is incorporated into a statewide analysis.

Lessons from this process have wide application for research using service data.

First, the capability to produce and use data is distributed very unevenly across providers. Some centres have staff with dedicated data responsibilities, others depend on one person with many competing duties, and several had not previously seen their own data analysed. Supporting providers to collect and maintain quality data is a fundamental step in using service data meaningfully, and one with substantial resourcing implications for projects of this kind. Engagement in this type of project may also increase data capability. Responding to queries requires staff to look closely at their own data and to articulate tacit practices. Capability is therefore not only a precondition for this work but one of its outputs.

Second, quality assurance of this kind may be better understood as knowledge production than as data cleaning. Engagement with providers is essential to understanding what service data represents. Variations in data recording practices between services may be errors but may also reflect differences in service design and client cohorts. Analysis across providers requires harmonisation of data, which risks obscuring these differences. Where harmonisation involves a loss of specificity or nuance, this should inform interpretation of the harmonised data.

Third, understanding how services record data is crucial for the type of meaningful interpretation needed to measure outcomes. This is a live question in legal assistance, where work is currently underway to develop sector-wide outcome measures. Outcomes already recorded cannot be interpreted without knowing how, when and for which clients they were captured. Where outcome monitoring is being designed, current recording practice indicates what could realistically be collected, and the support providers would need to do so consistently. These issues are likely to be familiar to anyone working with frontline service data, whether the interest is in the services themselves, in the populations they reach, or in the data as a resource for wider analysis. Across these lessons, the limits of administrative data in a given context are better treated as research findings in their own right, rather than only as limitations to be noted.