Research was undertaken by Rosa Billi, Christine A. Stone, Paul Marden and Kristal Yeung.
What was the research?
The Victorian Gambling Study is the first large scale, general population longitudinal study of gambling behaviour conducted in Australia. It includes an examination of the number of new cases of problem gambling per year (incidence rate), gambling prevalence (in 2008), pathways to gambling risk categories, and the relationship between gambling risk and health.
Commencing in 2008 with 15,000 Victorian adult participants, data collection followed each year in what is referred to as 'waves'. This fifth report includes findings on the gambling behaviour and health of the 3,701 Victorian adults who completed all four waves over four years.
Four previous reports have been published and can be accessed on the foundation's website.
Initially funded by the Victorian government through the Department of Justice, the study transferred to the Victorian Responsible Gambling Foundation in 2012, and received ethics approval from the Department of Justice Human Research Ethics Committee.
How was the research conducted?
The research involved interviews with participants in four annual waves from 2008 to 2012.
- The prevalence study was a cross-sectional study. It used computer-assisted telephone interviewing (CATI) via random digit dialling of 15,000 adults.
- Published in 2009, Problem Gambling from a Public Health Perspective is sometimes referred by the gambling research community as the 'epi study'.
- Participants were screened for their level of gambling risk via a series of validated questions. The categories ranged from non-risk to problem gambler.
- The study followed participants from wave 1 who consented to further research. Initially, 7,148 adults agreed to participate in the longitudinal study, however 5,003 of these adults participated in wave 2 and 5,620 participated in wave 3. A total of 3,701 adults participated in wave 4.
- Each wave consisted of annual CATI surveys which asked questions about gambling, health and wellbeing.
- A qualitative study was completed between waves 3 and 4 that consisted of face-to-face interviews with 44 participants.
Limitations of the research
- Apart from in its first year, the study was not looking at a random sample of Victorians. Therefore, findings from waves 2 to 4 are not representative of the general Victorian population.
- The study recruited very few numbers of people from cultural and linguistically diverse backgrounds or of Aboriginal or Torres Strait Islander descent. Therefore, there is no analysis available for these populations.
- The analysis of data collected from longitudinal studies takes time to complete and of course changes do occur in the world between the end of the study and findings coming out. Therefore, when findings are published, results may not be as relevant, depending on what may have changed.
- Data collection for this study occurred between 2008 and 2012. During that time and since, lots of changes have occurred in the gambling environment in Victoria. Especially note, there was a big and growing increase in advertising by online corporate bookmakers and the arrival of new devices which can be used to place online bets, like smart phones and tablets. The full impact of these changes is unlikely to be caught in the findings of this study.
What were key findings of the research?
This report includes some results which have been released in the four previous reports from the study, as well as additional findings from wave 1 and new findings from the analysis of all four waves of the study.
Key findings from wave 1, 2008
- In 2008, 73 per cent of adult Victorians participated in gambling. Note that this included a very wide range of activities, including lottos, racing sweeps and even stock market day trading.
- 23 per cent of the population gambled more than weekly; 18 per cent several times per month; and 32 per cent less than monthly.
- The estimated problem gambling prevalence was 0.7 per cent. An additional 2.4 per cent were moderate-risk gamblers and 5.7 per cent were low-risk gamblers.
- 91 per cent of problem gamblers gambled on pokies; 34 per cent bet on races; 25 per cent gambled on table games; and 16 per cent bet on sports. Note that many problem gamblers gambled in more than one way and so were counted in each of these categories.
- 21 per cent of the general population gambled on pokies; 16 per cent bet on races; 5 per cent gambled on table games and 4 per cent bet on sports.
Problem gamblers have much lower social capital, poorer health status and were more likely to report negative life events such as death, divorce, retirement, injury and illness.
Key findings across the study from waves 1-4, 2009 - 2012
Moves in and out of problem gambling behaviour
- The estimated amount of new problem gamblers in a year (incidence rate) was 0.36 per cent. Therefore, a little over one third of problem gamblers in a year had not been problem gamblers the year before and just under two thirds of problem gamblers in a year had been problem gamblers the year before.
- Approximately two thirds of the new problem gamblers in a year were gamblers with a history of lifetime problem or pathological gambling. Therefore, most 'new' problem gamblers were actually people lapsing back into problem gambling after at least one year or more not doing so.
- Having a lifetime problem gambling risk emerged as the strongest predictor of problem gambling.
Demographic and health factors
- Moving over time to at-risk gambling (low risk, moderate risk or problem gambling) was associated with:
- being male
- speaking a language other than English
- having year 10 education or less
- showing signs of alcohol dependence
- lifetime problem gambling risk, and
- anxiety and obesity.
- Being female was a protective factor for the development of higher risk gambling (moderate and problem gambling).
- The presence of any health condition in wave 1 was associated at a later occurrence of high risk gambling (moderate and problem gambling).
- Participants with anxiety were four times more likely to develop problem gambling.
- After adjusting for the effects of smoking and lifetime gambling risk, this was the only health condition associated with the development of problem gambling.
The presence of problem gambling in wave 1 was associated with the development of new onset health conditions in the following year.
Types of gambling
- For all participants, frequency of gambling on electronic gaming machines (EGMs) was associated with an increase in gambling risk in later waves.
- This relationship was 'monotonic' meaning that for each increase in frequency of EGM participation, there was a corresponding increase in problem gambling risk.
- For participants who had a previous history of problem or pathological gambling, this increase was more marked than those without this history.
- Frequency of race betting is also associated with an increased problem gambling risk.
- However, this increase occurs at a particular threshold. This means that there was no direct increase according to frequency of betting, however betting above a certain frequency was associated to later development.
Movements between levels of gambling risk
Between wave 1 and wave 4
- 93 per cent of zero-risk gamblers remained zero-risk gamblers (that is non gamblers and non-problem gamblers)
- 27 per cent of low-risk gamblers remained low-risk gamblers
- 35 per cent of moderate-risk gamblers remained moderate-risk gamblers, and
- 55 per cent of problem gamblers remained problem gamblers.
When considering transitions across the study
- moderate risk gamblers have the greatest probability of transitioning to problem gambling 9 per cent
- the probability of problem gamblers remaining problem gamblers is 71 per cent
- the probability that problem gamblers transition to moderate risk gambling is 22 per cent
- the probability that problem gamblers will become non gamblers is close to zero, and
- the probability that problem gamblers will become non-problem gamblers is 3 per cent.
How this research might be useful?
This study suggests that problem gambling may be a long term condition, and likely to require long term support for recovery.
Movement in and out of the problem gambling risk state suggests that relapse appears to be common.
The study provides valuable insights into the behaviour of a large number of gamblers over time. It provides unprecedented findings for analysis into how gamblers change or don't change their behaviour over time and provides detail about how other aspects of their lives may be affected by their gambling.
Such a study cannot be definitive, however, when contextualised with other research, it can provide a solid and much better base of knowledge for those concerned with addressing problem gambling and reducing gambling-related harm.
In particular, the study suggests prevention and intervention programs should be targeted toward at risk groups. Risk factors include not speaking English as a first language, signs of alcohol dependence, a history of problem gambling, experiencing anxiety, and education up to year 10 or less.
However, specific gambling behaviours and history also increase the likelihood of certain people entering into problem gambling. Moderate risk gamblers and those who have previously been problem or pathological gamblers are particularly at risk.
The presence of other health conditions is associated with progression to higher risk gambling, highlighting the need to work with other health sectors such as primary care, drug and alcohol and mental health.
This study provides a number of findings related to transitions and predictors which have not been undertaken before. These, along with other findings, require replication. Therefore, further research is required to confirm the findings in this study.
Want to know more?
The full report of the study (PDF - 2362.4 KB).
How to cite this research
Billi, R., Stone, C.A., Marden, P. & Yeung, K. (2014). The Victorian Gambling Study: A longitudinal study of gambling and health in Victoria, 2008-2012. Victoria, Australia: Victorian Responsible Gambling Foundation.
GIRO research updates
This summary is included in the September edition of the GIRO research updates.
September 2014 (PDF - 374.9 KB)