1351.0.55.061 - Research Paper: Assessing the Quality of Linking Temporary Visa Holders Administrative Data to the 2011 Census, Mar 2017  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 20/03/2017  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All


Temporary entrants are an important and growing part of the Australian population, with the latest available figures from the Department of Immigration and Border Protection (DIBP) showing close to 1.2 million temporary visa holders in the country as at 30 June 2016. This includes over 400,000 international students and approximately 170,000 skilled worker visa holders.

International students have a significant impact on the Australian economy through their contribution to the education export industry, while the availability of temporary skilled workers enables businesses to access international labour markets for skills and technical expertise. Increasingly, temporary entry is becoming a pathway to permanent residency. As such, the policy importance of this group is growing, and data is needed to provide a sound evidence base to adequately shape and evaluate relevant immigration and settlement policy. Data integration presents the opportunity to help meet this need for more information on temporary entrants.

In 2014, the National Migrant Statistics Unit of the Australian Bureau of Statistics (ABS) investigated the feasibility of probabilistically linking records from the 2011 Census of Population and Housing to administrative data on temporary visa holders held by DIBP. The linkage was deemed feasible and results were published in the research paper Assessing the Suitability of Temporary Migrants Administrative Data for Data Integration (ABS cat. no. 1351.0.55.053). Following on from this, the ABS has proceeded to link the temporary visa holders’ records to the 2011 Census records.

An innovative linking strategy was developed to link the Temporary Visa Holders (TVH) dataset with the 2011 Census. In addition to standard geographic and demographic information, this strategy exploited comparable employment and education variables to improve the likelihood of finding unique links amongst a population characterised by a high degree of homogeneity. A dwelling indicator was also utilised to assist in linking secondary applicants.

At the completion of the linkage process, more than 240,000 of the 513,000 TVH records had been linked to a Census record to create the Australian Census and Temporary Entrants Integrated Dataset. This equates to a linkage rate of 48%. Given the characteristics and temporary nature of this population of interest, this is considered a good result.

The two main reasons for unlinked TVH records were found to be that either the person was not recorded in the Census or that the quality of information on the TVH and/or Census datasets was insufficient to adequately link the record. It is likely that factors such as language barriers and uncertainty about whether to participate in the Census may have contributed to a large number of temporary entrants not being included in the Census, or providing insufficient or inaccurate information.

The linked dataset is characterised by a degree of under and over-representation of some migrant subpopulations within the international student and skilled worker temporary entrant populations. For example, temporary visa holders from China and those aged between 18 and 25 years were found to be particularly under-represented. Hence, inferences made from the linked dataset about these subpopulations may be biased if appropriate adjustments for the unlinked components are not made. As such, a weighting strategy to account for under- and over-representation of key subgroups in the dataset and support appropriate statistical analyses has been developed as part of this study.

Information contained on the TVH dataset was generally of high quality across most variables, with the exception of address information. The quality of address information varied considerably and in some instances was missing. For international students in particular, many of whom share quite generic characteristics, detailed geographic information gleaned from their address can be especially important in finding high quality, unique links. A combination of missing and poor quality geographic information can seriously affect linkage outcomes, especially in the absence of a unique identifier common to both datasets. Name information was also not used in the study. In the future, however, the potential for using anonymised name information as an additional linking variable could mitigate the issues caused by less than optimal address information and offer substantial improvements in linkage outcomes.

This paper concludes that with informed use, the linked dataset can be used for analysing the social and economic characteristics of temporary entrants by their visa and Census characteristics for the very first time. Two case studies are presented to demonstrate this. A dataset containing information representative of 240,000 temporary migrants is a significant step forward in filling a data gap that is well recognised. This dataset has the potential to inform a number of policy and research questions including the economic contributions of temporary visa holders, their geographic distribution, and the extent to which temporary workers are in jobs commensurate with their skills. Following on from these promising results, it is recommended that the linkage be repeated utilising the 2016 Census, and that the resulting linked and confidentialised dataset be made available for research and analytical purposes.