Gambling in Suburban Australia

Research Report – March 2019

2. Method

2.1 Overview

This exploratory place-based study investigates gambling consumption in two sites in suburban Melbourne, Australia. Each site comprises a geographic area consisting of six suburb clusters specially developed for the purposes of this study.

Site 1 was selected for high levels of EGM gambling availability and consumption. This area incorporates six suburbs in western Melbourne within the City of Brimbank, an area of Victoria's highest EGM losses, as well as one suburb in the City of Maribyrnong. The suburbs selected were: Sunshine, Sunshine North, Sunshine West, Ardeer, Albion and Braybrook.

Site 2 was selected as it has around-average levels of gambling consumption compared to the rest of Victoria, and is a similar proximity to the city to Site 1, as well as a similar size (geographical and population). Site 2 comprises a cluster of six suburbs located in eastern Melbourne within the City of Whitehorse, including Box Hill, Box Hill South, Box Hill North, Blackburn, Blackburn South and Blackburn North.

The Box Hill district comprises part of the only 'dry area' in Victoria, where hotel, bar, club and retail liquor licences have been substantially restricted, and licences must be approved through a poll of local residents conducted by the Victorian Electoral Commission (Victorian Commission for Gambling and Liquor Regulation [VCGLR], 2017). Given that it is a prerequisite for EGM licence holders to have a liquor licence, this provides a unique environment for a study of this nature. This arrangement allows for great collective community agency relating to the licensing of unhealthy commodities in Site 2.

In total there were 11 EGM venues in these areas, with eight venues in Site 1 and three in Site 2 (see Appendix A).

The study adopted a range of qualitative and quantitative methods. The report uses data obtained from the following components of the study:

  • Secondary data were used to develop a community profile of each site:
    • Australian Bureau of Statistics (ABS) 2016 Census population and housing and socio-economic indexes for areas (SEIFA) data
    • Victorian Commission for Gambling and Liquor Regulation (VCGLR) EGM data
    • review of socio-historical information relating to each site.
  • We carried out site and venue observations (n = 11 venues) and compared activity in these venues against venue Responsible Gambling Code of Conduct (CoC) documents.
  • A total of 159 people participated in interviews or focus groups. These included:
    • semi-structured interviews with people who gambled and have experienced gambling harm and significant others (e.g. partners, siblings and children of people who gambled) (n = 64)
    • focus groups with the general resident population: English language in Site 1 (n = 12) and Site 2 (n = 12); people with a disability/carers (n = 3); and Vietnamese language focus groups in Site 1 only (n = 38)
    • semi-structured interviews and focus groups in both sites with gambling professionals from venue, treatment and policy and regulation areas, community and social welfare organisations, and local government professionals (n = 30).

The data sources and methods are described in more detail in Appendix A. Appendix B provides study materials.

2.2 Recruitment of study participants

A short survey was developed for the purpose of screening and recruiting potential participants for in-depth interviews with people who gambled and their significant others, and focus groups with other local residents in the general population (described in Appendix A). This survey collected information about usual recreational activities, gambling attitudes and participation, EGM venue visitation, and socio-economic and demographic information (see Appendix B).

Survey administration

The survey was programmed using LimeSurvey™ and a website was developed to support the study. The survey was piloted and minor changes were subsequently made. The online survey completion time averaged 12 minutes. To facilitate the participation of those less comfortable in the online environment, the survey was also adapted to a paper version that could be completed by a researcher face-to-face or over the telephone via a toll-free number.

Survey promotion

All local residents (aged 18+) in each site were eligible to participate in the survey. A variety of approaches were used to promote the study and to recruit participants (see Appendix A for details). Residents were encouraged to complete the survey through the award of $100 supermarket vouchers for randomly selected respondents (= 50 across two sites).

Advertisements were placed in local newspapers over several weeks in each site as well as online via Twitter, Facebook and Gumtree. A study flyer was distributed to household letterboxes (see Appendix B). Posters and flyers were circulated throughout each study area with support from local councils and services. Local media outlets were contacted with information about the study, with one newspaper in Site 1 publishing a feature story about the study, and a local radio station conducting an interview with a researcher in Site 2 (no equivalent station existed in Site 1). Local Gambling Help Services provided support in the form of dissemination of study promotional materials, referrals of a small number of people who gambled to the study, and the use of interview rooms. The local government and a number of community services within each of the sites also helped by promoting the study and our recruitment information though their networks and providing interview rooms.

Further promotion was conducted directly by attending local community groups and activities including: local cultural groups, the Men's Shed, knitting groups, community lunches, local markets and festivals, neighbourhood houses, non-government organisations and faith-based organisations in each site.

Analysis of survey responses demonstrated that letterboxing of flyers to all households in Site 1 yielded the highest number of survey participants, and social media the least. In Site 2, an opposite pattern was found with social media promotion, primarily Gumtree advertisements, yielding the highest number of participants, while letterboxing achieved lower numbers of participants. Local newspaper advertisements and promotion through local services and community events were moderately successful in both sites.


Recruitment in Site 1 was undertaken for 37 weeks from March to September 2015 and from November 2015 to February 2016. In Site 2, recruitment ran for 23 weeks from September 2015 to February 2016. A total of 411 completed survey responses were received; 252 in Site 1 and 159 in Site 2. A small number of participants who lived in suburbs immediately adjacent to the study area (22 in Site 1 and 18 in Site 2) were included in this total. Seventeen responses were excluded as participants reported living in Melbourne but not within the study area or adjacent suburbs.4

From this, sample participants were invited for an interview based on their responses to categories of people who gambled intensely or who reported lifetime harms from gambling and their significant others. Local residents who did not gamble at harmful levels were selected to participate in focus groups. We sought a balance of men and women in each site, as well as a mix of younger and older participants.

Study participants

In addition to those described above (recruited through the survey), local Vietnamese-speaking residents were recruited through the networks of a locally based Vietnamese-speaking research assistant in Site 1. Professionals were recruited by direct approach based on searches of services available in the area and existing researcher networks. A summary of the sample is provided in Table 2.1 below.

Table 2.1:Study participants
Type of participant Number of participants
Site 1 Site 2 Total
Person who gambled and has experienced harm/problems 24 20 44
Significant other 12 8 20
Focus group (English language) 12 12 24
Focus group (disability/carer) 3 - 3
Focus group (Vietnamese language) 38 - 38
Professional 19 11 30
Total 108 51 159

Note: Detailed participant data and demographic tables are provided in Appendix A.

Twenty-four people who gambled (16 female, eight male) and 12 significant others (all female) were interviewed in Site 1. In Site 2, a total of 20 people who gambled (16 male and four female) and eight significant others (seven female and one male) were interviewed. A more detailed description of this sample is provided in Appendix A, Table A2.

EGMs were the primary problematic form of gambling for 75% of people who gambled in Site 1 and 70% of people who gambled in Site 2. Of participants who gambled at harmful levels, seven of 24 in Site 1 (29%) and eight out of the 20 in Site 2 (40%) also reported visiting the casino in central Melbourne in the past month. Further detail is provided in Appendix A, Table A3.

2.3 Data analysis and triangulation of results

Descriptive profiles were compiled for both sites. This consisted of the ABS census population in each site, the ABS SEIFA Index of relative socio-economic disadvantage (IRSD) for each suburb, and the VCGLR EGM data for each venue.

All qualitative interviews and focus groups were digitally recorded with consent from participants, and subsequently professionally transcribed. Vietnamese-language focus group recordings were translated and transcribed into English by an accredited National Accreditation Authority for Translators and Interpreters (NAATI) interpreter. These transcripts, along with venue CoC documents and summary site observation notes, were uploaded into NVivo 11™ software. Documents were initially thematically coded to nodes by the authors. Codes were refined, sorted and clustered as analysis progressed (Miles, Huberman, & Saldaña, 2013; Saldaña, 2015). Coding was cross-checked and validated between both authors who frequently discussed the themes to test observations and insights that were emerging from the data.

The final stage of analysis involved the triangulation of methods to test the validity of findings between the study methods. Triangulation of data from multiple sources enhanced the consistency and applicability of the qualitative components of the study (Noble & Smith, 2015).

The authors presented their findings to staff and/or councillors in each local government area prior to publication.

Participant codes

Quotes from participants reported in the results section are coded to provide anonymous context with reference to the site (1 or 2), study categorisation (person who gambled [G], significant other [SO], local resident [LR], Vietnamese local resident [LRV] or professional [P]) and gender [M] or [F].

2.4 Ethical approval

Ethical approval for the study was provided by the Australian Institute of Family Studies Human Research Ethics Committee (ref. 14/27) and multicentre approval was obtained from Cohealth and IPC Health (formerly known as ISIS Primary Care).

To address any local concerns about the study, the authors presented their findings to staff and/or councillors in each local government area prior to publication.

4 The primary purpose of the survey was as a recruitment conduit for interviews with people who gamble and significant others and English language focus groups. However, quantitative data on gambling at the local area level in each of the two sites were also collected and are intended to be the subject of a later publication.