About & Methodology

2023 Being Chinese in Australia: Public Opinion in Chinese Communities reports the results of a national survey of 1200 adults in Australia who self-identify as of Chinese heritage. The survey was conducted between 27 September and 10 December 2022. The Social Research Centre (SRC) conducted the non-probability online survey using panel providers Multicultural Marketing and Management (MMM) and The Online Research Unit (ORU).

Chinese-Australians constitute a population that is dispersed and diverse, including recent immigrants from multiple countries with different native languages who may have difficulty completing a public opinion survey in English. Increased attention paid to Chinese-Australians by both the Australian and Chinese political classes may have also made some potential respondents more reluctant to cooperate with a survey request. Collectively, these characteristics present significant challenges to anyone wishing to survey this population.

The questionnaire was developed by the author and then refined in consultation with the SRC. The order of questions in the questionnaire was different from the order presented in this report.

The survey respondents included citizens, permanent residents and visa holders (excluding tourist visa holders) who identified as having Chinese ancestry. A total of 1721 panel members engaged with the survey and of those, 1200 (69.7%) completed the survey. Quotas based on country of birth and age were set to minimise the over-representation of younger, highly acculturated, second-generation respondents.

The sample design was representative of the Australian population of adults over the age of 18 who reported Chinese ancestry as either their first or second response to the 2021 Australian Bureau of Statistics (ABS) Census question. The profile of survey respondents was compared with ABS TableBuilder 2021 benchmarks across age, gender, location and country of birth to determine the extent, if any, of response bias.

To achieve the desired quotas, the SRC provided MMM with seven sets of links to recruit participants via six different forms of social media. Prospective respondents were asked to supply information such as gender, age and country of birth via email to allow MMM to screen for those in scope before being provided the survey link. Where a respondent had completed the 2021 survey, they were excluded by MMM. ORU were sent their own unique set of links and promoted the survey via their own panel and targeted panel members that had been pre-screened for country of birth (see Table A).

Respondents had the option to complete the questionnaire in English, Simplified Chinese or Traditional Chinese. A very high proportion chose to complete in a language other than English, with 56.5% selecting Simplified Chinese, 8.6% selecting Traditional Chinese and 34.9% completing the survey in English.

Given approximately 5.5% of the Australian population identify as of Chinese ancestry in the 2021 Census, randomly recruiting a sample of 1200 Chinese-Australians would have required interviewing and screening more than 20,000 Australians. This approach would not have been practical, and so the SRC worked with non-probability responses to build a representative sample.

To address potential bias in the sample, weighting was applied. The design weights were adjusted so that they matched external benchmarks of key demographic parameters likely to be correlated with the survey outcomes and the likelihood of response. The survey collected a range of characteristics about the respondent for which population totals can be obtained from the ABS. Having this range of data meant that we could try different model covariates with a view to aligning the weighted sample as closely as possible with the population totals for available characteristics. The final choice of covariates was determined from two considerations: minimising the bias and maximising the effective sample size (as measured by weighting efficiency).

The variables that best met the criteria for inclusion in the weight were age by education, gender, language spoken at home other than English, country of birth, and state or territory of residence (see Table B). The overall weighting efficiency is 86.1% and a margin of error of 3.0%.

To address data quality and the results being potentially biased, an IP address check was built into the survey to reject any address that was outside Australia. Quality checks around verbatims, speedsters and nonsensical or contradicting responses were also implemented.

The question regarding ‘Confidence in global leaders’ was administered differently in 2021 and 2022. In 2022, online respondents were able to select ‘never heard of the person’ or ‘don’t know’ on the first response screen. In 2021, respondents were only offered these responses if they declined to answer at the first response screen. This resulted in a significant difference in the number of respondents who said they had ‘never heard of the person’ when compared to 2021.

A question about the partnership between Australia, India, Japan and the United States (the Quad) was added to the questionnaire in 2022, with ‘not sure’ and ‘never heard of the Quad’ responses being displayed on a second screen. This differs from how data were collected for the same question in the 2022 Lowy Institute Poll (‘not sure’ and ‘never heard of the Quad’ responses displayed on the same screen). This explains the differences seen in the proportion who answered ‘not sure’ and ‘never heard of the Quad’ between surveys.

Table A: Social media channels by demographics

* Denotes small base <30

Table B: Covariates used in model for establishment weights, with population distributions

---

Acknowledgements

Several questions in this report were modelled on those developed by other polling organisations, including the Pew Research Center, OmniPoll and Scanlon Foundation. Fieldwork was managed by Karly Day and Tina Petroulias of the Social Research Centre (SRC). Dr Benjamin Phillips, Andrew Ward and Jack Barton of the SRC provided design and weighting advice. John Davis of OmniPoll provided independent consulting and reviewed the questionnaire and earlier versions of the report. Natasha Kassam, formerly of the Lowy Institute, contributed to the design of the questionnaire. Ian Bruce of the Lowy Institute updated and refreshed the website interactive with contributions from Brody Smith. Richard McGregor and Sam Roggeveen of the Lowy Institute read and commented on earlier versions of the report, as did Anthony Bubalo. Jack Sato of the Lowy Institute provided data checking assistance. Clare Caldwell of the Lowy Institute copyedited the report.

This report is part of the Lowy Institute’s Multiculturalism, Identity and Influence Project, funded by the Australian Department of Home Affairs. Responsibility for the views, information or advice expressed in this report is that of the author. The contents of this report do not necessarily reflect the views of the Lowy Institute or the Australian government.