Evaluating the Effectiveness of Responsible Gambling Messages: A Rapid Evidence Assessment

Abstract

To minimize the harms associated with gambling, an array of responsible gambling (RG) messages has been developed to raise awareness of the risks of problem gambling and encourage safer gambling behaviors. However, evidence is limited as to the utility of RG messages to promote positive gambling-related beliefs and behaviors. In the current paper, we report the results of a Rapid Evidence Assessment (REA) of empirical research on RG messages. We identified 3200 unique articles published between 1890 and September 2024 using search terms related to RG messaging. Eighteen articles (containing 20 unique studies) met our inclusion criteria. Two general themes emerged: 1) RG message preferences among players and 2) RG message effectiveness. Specifically, players prefer self-appraisal messages, which were more effective in promoting RG behaviors compared to informative messages. Messages content also needs to be segmented (i.e., low-risk players prefer different types of messages than high risk players, such as highlighting player quizzes for low-risk players and helplines for high-risk players). Lastly, RG messages should be presented dynamically (e.g., pop-ups on an Electronic Gaming Machine: EGM). Results suggest a need for the gambling industry to adopt targeted, evidence-based RG messaging, as well as a need to engage in integrated knowledge mobilization, to more effectively promote RG. These findings underscore the importance of tailoring RG messages to player risk levels and preferences while leveraging dynamic delivery methods to maximize their effectiveness in promoting safer gambling behaviors and reducing harm.

Gambling is a popular recreational activity enjoyed by millions worldwide (Williams et al., 2017). Although most people do not gamble in excess, it is estimated that approximately 0.12–5.8% of American players develop a gambling problem, with other countries showing comparable rates (Potenza et al., 2019). Responsible gambling (RG) strategies have been developed to prevent the development of problem gambling and promote safer gambling attitudes and behaviors (Ladouceur et al., 2017). Amongst these strategies, RG messages act as a key touchpoint between operators and players, helping to increase awareness of available RG tools and encouraging players to gamble within their means (Gainsbury et al., 2018). They are often prominently displayed in various gambling contexts, such as gaming floor signage within land-based casinos, pop-up messages on electronic gambling machines (EGMs), and notifications on online gambling platforms. However, questions remain about their overall effectiveness in influencing player attitudes and behaviors.

Despite a growing body of research on RG messaging, most of it has been narrowly focused on pop-up messages delivered via EGMs (Ginley et al., 2017). Less attention has been given to the broader landscape of RG messaging, including the most effective design, content, and delivery methods for promoting safer gambling practices across diverse player demographics and gambling contexts. To address this gap, in this paper we present the findings of a Rapid Evidence Assessment (REA) of the empirical research on RG messaging.

The Evolution of RG Messaging

RG messages are communication tools designed to promote safer gambling practices and mitigate harm by informing players about gambling risks, providing strategies to maintain control over their gambling behaviors, and raising awareness of available support services. Early RG messages were heavily influenced by public health campaigns, such as those targeting tobacco and alcohol use, and focused on increasing player awareness of the risks associated with gambling and the probabilities of winning (e.g., “The odds are not in your favor”; Caillon et al., 2021). The assumption was that such messages would increase player knowledge, subsequently leading to safer gambling behaviors (Blaszczynski et al., 2011).

An early approach to RG messaging was promoting this educational content via signage, pamphlets, or videos to inform players about how games worked (Wohl et al., 2010, 2013). Over time, technological advancements have expanded RG messaging to include dynamic forms of communication, such as being embedded in EGMs or online gambling platforms (Monaghan, 2009). This allows for just-in-time messaging in which players can be exposed to RG information prior to initiating a playing session as well as when players are at a critical juncture in their play, such as when a pre-set limit on play has been reached. Moreover, these messages have become more complex and have been adapted to encourage self-appraisal and personal responsibility (Gainsbury et al., 2015).

Despite their widespread adoption, RG messages face significant challenges in achieving their intended goals. Research that has evaluated the utility of pop-up messages on EGMs have yielded mixed results. Some findings suggest that messages that prompt players to take breaks or reflect on their gambling behavior can reduce session length and expenditure (Monaghan, 2009; Wohl et al., 2013). However, other studies have found that these effects are short-lived, with players resuming their gambling behavior shortly after dismissing the message (Ginley et al., 2017). For instance, Newall et al. (2023) found the presence of the safer gambling message “When the fun stops, stop” had no significant protective effect on participants’ gambling behavior. These findings highlight a broader challenge in RG messaging—many campaigns rely on simplistic slogans that lack empirical validation and fail to resonate with diverse player demographics, which limits their potential to reduce gambling-related harms (Marko et al., 2023).

Player (Mis)Perceptions of RG Messages

A significant challenge in RG messaging is player perception. Many players, particularly those at lower risk for gambling problems, view RG messages as irrelevant to their own experiences (Houghton & Moss, 2024; Morvannou et al., 2020; Thomas et al., 2015). For instance, Hing (2003) found that recreational gamblers often ignore RG signage because they perceive them as directed at individuals with severe gambling problems. More recently, Gaudett et al. (2024) found that players' reluctance to engage with RG programming is, in part, due to the belief that RG programs are designed for players who have a gambling problem. This lack of resonance underscores the critical need to identify RG messaging strategies that capture the attention of recreational players while remaining relevant to those at higher risk.

Challenges also exist in terms of how those at risk for developing a gambling problem and those who are currently living with a gambling problem perceive RG messages. Players tend to find them unhelpful or stigmatizing (Miller et al., 2018). For example, messages emphasizing personal responsibility (e.g., “Set your limits”) can alienate players at risk for developing a gambling disorder (Miller & Thomas, 2018). Although operators may perceive such messages as promoting positive play, players may interpret the message as suggesting that gambling problems are a result of individual failure. Consequently, although well-intentioned, players do not always interpret messages the way operators intend (Bernhard & Preston, 2004). This disconnect between message content and player experiences highlights the importance of designing messages that are both empathetic and tailored to the diverse needs of players. Addressing this gap requires a deeper understanding of how message content and framing influence player perceptions and their engagement with RG.

Consequently, an issue may be the framing of RG. Positively framed messages emphasizing the benefits of safer gambling behaviors are more likely to resonate with players than negatively framed messages that highlight the risks of gambling harms (Rothman et al., 2003). For example, according to the Positive Play Model (Wood et al., 2017) a message like “Set a budget to enjoy gambling without regrets” may encourage proactive behaviors, whereas a message such as “Overspending can ruin your life” may evoke defensiveness or resistance, particularly among players who do not perceive themselves as vulnerable (Gainsbury et al., 2018). Identifying framing strategies that balance relevance with appeal is crucial to developing messages that resonate across player demographics and motivate safer gambling practices. That is, there is a need to consolidate findings from the existing literature to better understand what types of RG messages are most likely to engage players. Exploring how message content, framing, and delivery impact perceptions can help guide the development of evidence-based messaging strategies that resonate with a diverse range of players and promote positive play.

Can RG Messages Promote Responsible Play?

Initial research on the effectiveness of RG messages focused on the utility of static messages, such as signage in casinos (e.g., Schrans et al., 2004). Although these messages can provide valuable information about RG tools and resources, they often lack the immediacy and interactivity needed to capture players’ attention and drive action (Hing et al., 2015). Online gambling platforms have sought to address the limitations of RG messages by integrating them directly into the gambling experience, using personalized notifications to alert players when they exceed predefined time or monetary limits (Gainsbury et al., 2018). These tailored approaches have shown promise in increasing player engagement with RG tools, but more research is needed to determine their long-term effectiveness.

The content of RG messages also appears to play a crucial role in their effectiveness. Messages providing specific, actionable guidance (e.g., “Take a break every 30 min to stay in control”) are more effective than vague statements like “Gamble responsibly” (Hing et al., 2015). Additionally, messages that encourage self-appraisal, such as “How does gambling make you feel when you win or lose?”, can prompt players to reflect on their behavior and consider adjustments (Glock et al., 2013). However, players often fail to read pop-up messages presented to them (Hollingshead et al., 2019), potentially due to message fatigue (i.e., a negative psychological state marked by feelings of monotony and disengagement caused by repeated exposure to messages advocating the same health focused behaviors, McCullock et al., 2025).

These mixed findings underscore the critical need for a systematic examination of RG messages to determine what content, framing, and delivery methods are most effective in promoting positive play. By consolidating and synthesizing existing research, this REA aims to identify evidence-based messaging strategies that can overcome challenges such as message fatigue, increase player engagement, and lead to meaningful behavioral change. Understanding these factors is essential for developing RG messages that resonate with diverse player demographics and sustain their effectiveness over time.

Knowledge Need: The Case for a Broad Scoped Rapid Evidence Assessment

Although several reviews have examined RG messaging, most have narrowly focused on pop-up messages or personalized notifications (Bjørseth et al., 2021; Monaghan, 2008). Such reviews have shown moderate effects of pop-up messages on gambling behavior, but considerable variability in outcomes. Broader reviews of RG tools and interventions (e.g., McAuliffe et al., 2021; McMahon et al., 2019) have further identified issues such as publication bias, inconsistent study quality, and a lack of robust evidence for many messaging strategies. However, these broad reviews often give limited attention to the content of RG messages and rarely examine how message effectiveness varies across different player demographics and gambling contexts. Moreover, previous reviews tend to silo message delivery methods without critically evaluating message framing or content, neglecting the influence of player characteristics such as risk level, age, and play style. They have also overlooked opportunities to design messages that promote positive play rather than merely addressing problem gambling, resulting in gaps remaining in the literature.

Addressing these shortcomings required a more integrated and player-centered evaluation of RG messaging strategies. The current REA aimed to contribute to the literature in three ways. First, although RG programs are widely available, player engagement is low (Gaudett et al., 2024; Nelson et al., 2008). Many RG messages (e.g., “When the fun stops, stop”) emphasize personal responsibility without offering actionable guidance or reflecting the needs of diverse player groups, thereby limiting their effectiveness (Miller & Thomas, 2018; Benhsain et al., 2004). Moreover, such messages are not evidence-based (see van Schalkwyk et al., 2021). Synthesizing evidence on effective messaging can inform frameworks that better resonate with players and promote meaningful engagement. Second, most RG messages rely on generic slogans and display a helpline number, which are often ignored (Hing, 2003; Hollingshead et al., 2019; Schrans et al., 2004). A one-size-fits-all approach fails to address the different motivations and risk profiles of players. Tailored messages that reflect specific demographics and behaviors are more likely to be read, understood, and acted upon (Bennett & Glasgow, 2009; Wood & Williams, 2009). Third, there is a need to develop RG messages that not only reduce the prevalence of problem gambling but also promote positive play. Industry trends increasingly emphasize safer play over personal responsibility (Davies et al., 2022), and successful messaging now focuses on relevant language, engaging tone, and actionable promote, such as self-appraisal prompts or positively framed content highlighting the benefits of behavior change (Glock et al., 2013; Rothman et al., 2003). This REA aims to identify the types of messages most likely to encourage safer gambling practices and to guide the development of evidence-based strategies that support broader player engagement.

Overview of the Current Review

The purpose of conducting the current REA was to identify and synthesize research on RG messaging to guide prevention, harm reduction, and industry practice. An REA is a structured, transparent, and reproducible method for reviewing evidence that applies systematic review principles—such as comprehensive searching, study selection, and quality appraisal—within a compressed timeline (Thomas et al., 2013). REAs are particularly useful in situations where time or resources are limited, and when stakeholders require timely yet credible insights to inform decisions. While REAs do not aim to capture all available evidence like traditional systematic reviews, they maintain methodological rigor by predefining eligibility criteria, using multiple reviewers for screening, and applying standardized frameworks for data extraction and synthesis. We chose to conduct an REA because it offers a practical balance between comprehensiveness and efficiency, producing high-quality, evidence-based insights that are well-suited for addressing time-sensitive or practice-oriented research questions. As such, the REA method provides a robust foundation for informing RG messaging strategies in both policy and practice contexts.

Given the applied focus and policy relevance of RG messaging—particularly in light of increasing regulatory interest and implementation across jurisdictions—we chose to conduct a REA to provide timely, structured, and practice-oriented insights. While REAs are not exhaustive, they offer a transparent and systematic alternative to full systematic reviews when timely synthesis of heterogeneous evidence is required. Our approach allowed for the inclusion of a diverse set of study types (including grey literature), which is often excluded in conventional systematic reviews but critical to informing real-world practice in this domain.

Methods

We conducted a comprehensive database search (i.e. PsycInfo, ProQuest Dissertations and Thesis, PubMed, Scopus) spanning from 1890 to September 2024 to identify relevant articles. We condensed records captured by the search and analysed them using Covidence, a web-based collaboration software platform that streamlines the production of systematic and other literature reviews (Veritas Health Innovation, 2022). Searches consisted of article titles, abstracts, and keywords to identify studies that related to the effectiveness of RG messages (i.e., studies relating to RG messages and common synonyms such as safer gambling and positive play). Search terms were pre-registered on OSF: https://osf.io/pjtcf/?view_only=add1d5c384c5436a9f3e1c890ab31034.

Inclusion and Exclusion Criteria

We included records if they (a) were original (primary), peer-reviewed research (e.g., journal articles, dissertations/theses), (b) were available in English or could be translated into English, and (c) evaluated the effectiveness of RG messages. Both quantitative and qualitative studies were eligible to ensure a broader capture of the literature. Records that did not meet these criteria were excluded (Table 1). Two key exclusions were: 1) studies using undergraduate students and 2) community samples that included people who did not gamble. Although laboratory-based research with student participants can provide valuable theoretical insights, findings from these samples may not generalize well to broader gambling populations due to differences in gambling behaviors, motivations, cognitive maturity, and impulsivity (Gainsbury & Blaszczynski, 2011). Second, we excluded studies involving only community members with no history of gambling, either online or in person. RG messages are designed for individuals who currently gamble or are at risk of gambling-related harm, and including non-gamblers would limit the applicability of findings to the intended target population. These criteria ensured that the synthesized evidence would better reflect real-world gambling behaviors and inform the development of effective, player-centered RG messaging strategies.

Table 1.

Inclusion and Exclusion criteria

Include	Exclude
Paper is original (primary) research • Journal article (peer reviewed) • Dissertations/Theses (can be unpublished)	Paper is not original research • Books and chapters • Review papers • Secondary analyses
Language is English or translation to English is accessible	Language is other than English or is not able to be translated
Population is casino players or online gamblers	Population is not in the context of gamblers • Community members who have not gambled online or in person • Undergraduate students
Evaluates the effectiveness of RG messages and campaigns from the players’ perspective • Experimental manipulations or measurements • Correlational designs (i.e., cross-sectional or longitudinal) • Qualitative studies	Does not evaluate the effectiveness of RG messages and campaigns from the players’ perspective • Operators, academics, casino employees

In addition to inclusion and exclusion criteria, we conducted a basic quality assessment of each full-text study to ensure the rigor of our review. Following REA best practices, each study was appraised based on three criteria: (1) clarity of research objectives, (2) appropriateness of methodological design, and (3) transparency in reporting of methods and results (e.g., sample characteristics, data collection, and analysis procedures). Two reviewers independently assessed each study against these criteria, and any discrepancies were resolved through discussion. Studies that failed to meet minimum thresholds across these dimensions were excluded from the final synthesis.

Study Selection

The search strategy yielded 3,337 records related to RG messages. After we removed 1,464 duplicates using Covidence (Veritas Health Innovation, 2022), we were left with 1,873 unique records. Two members of the research team conducted the title and abstract screening and reviewed in full those records that we deemed potentially relevant during the title and abstract screening (k = 109). Following a full-text reading, 14 research articles (containing 14 unique studies) passed the screening process (see Table 2).

Table 2.

Key information about the studies included in the REA

Authors	Objective	Sample Size	Sample	Design	Methods	Results
Armstrong et al. (2018)	Assess the effectiveness of pop-up messages in reducing gambling intensity during a simulated gambling task	N = 172	South Australian men and women who had a gambling problem	Quantitative	Experimental study	Betting speed, betting persistence, and losses reduced when participants were shown negatively framed self-evaluative messages and challenging self-monitoring messages
Auer and Griffiths (2015)	To test the efficacy of an enhanced pop-up message compared to a simple message in stopping play	N = 23110	German online slots players	Quantitative	Secondary data analysis	RG behaviors improved following an enhanced message
Caillon et al. (2021)	Compare the impact of self-appraisal and informative pop-up messages on gambling behavior	N = 58	French online gamblers	Mixed methods	Experimental study	Improvement in RG behaviors and reduction in erroneous beliefs
Davies et al. (2022)	To generate new, more effective RG messages	N = 45 (focus group) N = 987 (pilot test)	British gamblers	Mixed methods	Focus group and longitudinal study	Effectiveness of RG messages depends on players’ age and gambling style
Department of Social Services (2014)	To assess the effectiveness of dynamic warning messages in encouraging RG	N = 667 and N = 17 (longitudinal cohort)	Australian gamblers	Mixed methods	Focus group and longitudinal study	RG behaviors improved following self-appraisal and informative messages
Gainsbury et al. (2018)	To understand potential differences between different populations of gamblers in terms of RG message effectiveness	N = 39	Online gamblers from Manitoba, Canada	Qualitative	Focus group	Effectiveness of RG messages depends on players’ age and gambling style
Gainsbury et al. (2015)	Assess the effect of RG message content on RG thoughts and behaviors	N = 667	Australian casino players	Mixed methods	Focus groups and longitudinal trial	Players recalled self-appraisal messages to a greater extent than informative messages
Houghton and Moss (2024)	To assess the effectiveness of delivering RG messages to sports betters via social media as well as which types of messages are most effective	N = 281	Online sports betters	Mixed methods	Correlational study with qualitative questions and experimental study	RG beliefs and behaviors improved during the intervention
Marko et al. (2023)	To explore how people who have experienced gambling-related harm interpret and apply RG messages in their life	N = 21	Gamblers who had been harmed by their own gambling behavior or people harmed by someone else’s gambling	Qualitative	Focus groups	Frame RG messages akin to smoking or alcohol campaigns
Michalska et al. (2020)	To assess online poker players’ views on harm reduction strategies	N = 311	Online poker players	Quantitative	Secondary data analysis	Players prefer messages about time limits and time wagered and lost
Monaghan and Blaszczynski (2009)	To assess whether the content of pop-up messages are recalled better than static messages	N = 124	EGM gamblers at two venues in Sydney, Australia	Quantitative	Experimental study	Participants were more likely to recall pop-up messages than static messages
Morvannou et al., 2020	To assess poker players’ perceptions of RG messages	N = 12	Poker players from casinos in Quebec, Canada	Qualitative	Semi-structured interviews	Participants thought messages are ineffective as they target players with gambling problems (which they did not feel they belonged to)
Munoz et al. (2010)	To assess if threatening messages change attitudes surrounding gambling behaviors	N = 258	VLT players in Montreal, Canada	Quantitative	Correlational study	Threatening messages (i.e., those that underscore long term consequences of problem gambling) are more effective
Mutti-Packer et al. (2022)	To examine the efficacy of fear based pop-up messages. Specifically, if high-threat warning messages are more effective than low-threat warning messages	N = 62	Online gamblers	Quantitative	Experimental study	Participants rated the high threat message as overall more effective
Newall et al. (2023)	Explore differences between two established RG messages against nine new messages and to assess whether rating differences are due to gender, age, or problem gambling symptomatology	N = 1865	Online gamblers or sports betters from the United Kingdom or United States	Quantitative	Correlational study	Messages such as “What are you prepared to lose today” or “Imagine what you could be buying instead” were rated higher than messages such as “What are you really gambling with?”
Riley-Smith and Binder (2003)	To assess the effectiveness of ten harm minimization messages on gambling behavior	N = 45	Australian problem gamblers or regular gamblers who played poker machines	Qualitative	Focus groups	Messages about lying were effective because players considered the consequences of their gambling on their relationships. Messages that end with a call to a helpline exclude players who do not feel they have a gambling problem
Rockloff et al. (2024)	To explore the effectiveness of positive-emotional and norm-based messages compared to a generic “gamble responsibly” message in improving gambling outcomes	N = 2074	Online sports betters from Australia	Quantitative	Experimental study	RG behaviors and beliefs did not differ depending on message type
Thomas et al. (2015)	To assess how gamblers view campaigns focusing on the harms related to problem gambling	N = 100	Gamblers from Victoria, Australia	Mixed methods	Phone interviews and correlational study	Campaigns should focus on multiple audience segments and also focus on the risks associated with products/industries, not just individual behavior

At this stage, the research team decided it would be beneficial to include grey literature in the REA. We conducted a search of the GREO Evidence Center using the aforementioned search terms and filtered by reports. This yielded 1,327 records related to RG messages. We reviewed in full those records that we deemed potentially relevant during the title and abstract screening (k = 20). Following a full-text reading, three research articles (containing five unique studies) passed the screening process. Lastly, when cross-checking the reference pages of the studies included in the REA, we found an additional article (containing one study) that was not included in the initial title and abstract screening that we included in the review.

In total, 18 research articles (containing 20 unique studies) were included in the review. For details of the record screening process, reviewed in a PRISMA diagram, see Fig. 1.

Fig. 1.

PRISMA diagram

Results

Study Characteristics

Country

A total of eight samples were drawn from Australia (Armstrong et al., 2018; Department of Social Service, 2014; Gainsbury et al., 2015; Monaghan & Blaszczynski, 2009; Riley-Smith & Binder, 2003; Rockloff et al., 2024; Thomas et al., 2015), three samples were drawn from Canada (Gainsbury et al., 2018; Morvannou et al., 2020; Munoz et al., 2010), one sample was drawn from the United States (Newall et al., 2023), three samples were drawn from the United Kingdom (Davies et al., 2022; Newall et al., 2023), and two samples were drawn from Europe (Auer & Griffiths, 2015; Caillon et al., 2021). Four studies did not specify where their samples were drawn from (Houghton & Moss, 2024; Marko et al., 2023; Michalska et al., 2020; Mutti-Packer et al., 2022).

Method Employed

Of the 20 studies included in our review, nine used quantitative methods (Armstrong et al., 2018; Auer & Griffiths, 2015; Gainsbury et al., 2015; Michalska et al., 2020; Monaghan & Blaszczynski, 2009; Munoz et al., 2010; Mutti-Packer et al., 2022; Newall et al., 2023; Rockloff et al., 2024), four used qualitative methods (Gainsbury et al., 2018; Marko et al., 2023; Morvannou et al., 2020; Riley-Smith & Binder, 2003), and seven used a mixed method approach (i.e. quantitative and qualitative; Caillon et al., 2021; Davies et al., 2022; Department of Social Services, 2014; Houghton & Moss, 2024; Thomas et al., 2015).

Sample Size

We captured 31,087 participants across the eligible research studies. The sample sizes used in the 20 identified studies varied considerably from a focus group of N = 12 (Morvannou et al., 2020) to N = 23,110 (Auer & Griffiths, 2015).

Target Population

A total of 27,800 participants gambled online (e.g., sports betting or poker; Auer & Griffiths, 2015; Caillon et al., 2021; Gainsbury et al., 2018; Houghton & Moss, 2024; Michalska et al., 2020; Mutti-Packer et al., 2022; Newall et al., 2023; Rockloff et al., 2024), 3,115 participants gambled at land-based casinos (e.g., EGMs, VLT, or poker machines; Davies et al., 2022; Department of Social Services, 2014; Monaghan & Blaszczynski, 2009; Morvannou et al., 2020; Munoz et al., 2010; Riley-Smith & Binder, 2003; Thomas et al., 2015), and 172 were community members (Armstrong et al., 2018) or people who had been impacted by gambling-related harms (e.g., having a significant other, friend, or family member who has a gambling disorder; Marko et al., 2023).

Emerging Themes

The 18 records (comprising 20 unique studies) were analyzed using narrative synthesis, guided by thematic analysis principles and consistent with best practices for REAs (Thomas et al., 2013) and narrative synthesis frameworks (Popay et al., 2006). We employed an inductive coding approach, allowing themes to emerge organically from the data rather than imposing a pre-existing framework. Two members of the research team independently reviewed each full-text article and generated initial codes that captured key findings, patterns, and author interpretations related to RG message content, preferences, and effectiveness.

Following this, the coders met to compare and consolidate their initial codes. Through iterative discussion, these codes were refined into broader, higher-order themes. Disagreements or discrepancies were resolved through consensus-building conversations. While we did not calculate formal intercoder reliability coefficients—given the interpretive, rather than statistical, focus of our synthesis—we ensured rigor through multiple coding rounds and collaborative theme development.

To be included in the final thematic framework, themes had to be (a) represented across multiple studies, (b) supported by data or author interpretation, and (c) directly relevant to the aims of the REA. Within each identified theme, findings were synthesized narratively to explore points of convergence and divergence across studies, as well as contextual factors (e.g., message framing, delivery mode, population characteristics) that might explain variations in outcomes. This process resulted in two major themes: (1) RG message preferences and (2) RG message effectiveness, each with several subthemes described in the Results section..

Theme I: RG Message Preferences

Theme I comprised 12 studies that examined the type of messages that resonate with players (Caillon et al., 2021; Davies et al., 2022; Gainsbury et al., 2018; Houghton & Moss, 2024; Marko et al., 2023; Michalska et al., 2020; Morvannou et al., 2020; Mutti-Packer et al., 2022; Newall et al., 2023; Riley-Smith & Binder, 2003; Thomas et al., 2015). This theme was further subdivided into two sub-themes: 1) preference for self-appraisal messages; 2) need for segmented RG messages.

Preference for Self-Appraisal Messages

Players expressed dislike towards generic messages (i.e., a universal or standard RG message aimed at all players, such as “Gamble responsibly”; Marko et al., 2023). Instead, they preferred self-appraisal messages and RG messages that underscored the consequences of problem gambling (Houghton & Moss, 2024; Marko et al., 2023). For example, Newall et al. (2023) found that players were responsive to messages such as “Imagine what you could be buying instead” or “What are you prepared to lose today” because they were perceived as being more believable and helpful than generic RG messages. Moreover, players tended to rate messages about the social consequences of gambling that were highly threatening (as opposed to low-threat messages) as more believable (Caillon et al., 2021; Mutti-Packer et al., 2022). For instance, Riley-Smith and Binder (2003) found that messages that highlighted how gambling may have harmed their interpersonal relationships encouraged respondents to consider changing their gambling behaviors and call a problem gambling helpline. However, it is unknown whether intentions translated into action. Thus, it is still to be determined whether such messages promote positive play, reduce problem gambling behavior, or facilitate treatment seeking. Overall, although these findings suggest that self-appraisal messages may be more effective in helping players to reflect on their problem gambling behavior compared to generic messages, more research is needed to assess whether this translates to behavior change.

Need for Segmented RG Messaging

Six studies reported results that suggest a need to segment the content of the RG message by player demographic characteristics and their play style. For instance, three studies found that message preferences differed depending on a player’s age. Young adults (i.e., 18–24 years old) preferred RG messages that that supported informed and strategic play, such as those providing guidance on managing behaviors like chasing losses (Gainsbury et al., 2018). Additionally, although both younger and older players valued messages that were respectful and avoided inducing feelings of guilt or blame, young adults were highly attuned to the tone of RG messages (Davies et al., 2022). Specifically, young adults were more likely to disregard messages that conveyed social judgment (e.g., “You may be the last person to realize you have a gambling problem; Riley-Smith & Binder, 2003). Younger adults also appeared more responsive to emotionally engaging messages than older adults. For instance, Riley-Smith and Binder (2003) found that a self-appraisal message (i.e., “Do you lie to hide the extent of your gambling?”) resonated with younger adults, particularly among young women, because it encouraged reflection on how gambling behaviors could affect personal relationships and signaled potential problematic play. In contrast, the self-appraisal message was less effective with older adults. Instead, older adults preferred messages that were lighthearted, optimistic, and provided clear information on accessing support for gambling-related challenges.

Additionally, findings from our REA suggest there is a need to tailor messages based on gambling styles. For instance, Gainsbury et al. (2018) examined whether RG messaging preferences vary by the type of game played. Skill-based gamblers and frequent gamblers both preferred simple and direct messages, but with notable differences: skill-based gamblers favored blunt, personalized messages tailored to their gambling behavior, while frequent gamblers responded better to positive messages, such as “Keep it a game.” An additional four studies assessed whether RG message preference differed between players at low-risk of a gambling problem and those at a higher risk of a gambling problem (Houghton & Moss, 2024; Morvannou et al., 2020; Riley-Smith & Binder, 2003; Thomas et al., 2015). Players at low-risk of problem gambling often expressed that RG messages were not aligned with their play experience (Houghton & Moss, 2024; Morvannou et al., 2020; Thomas et al., 2015). Instead, they felt RG messages are directed at players who have symptoms of problem gambling—a demographic they did not feel they aligned with. As such, messages targeted at players at high-risk of developing a gambling problem (e.g., messages exclusively containing helpline information) may inadvertently exclude players who do not feel they have a gambling problem (Riley-Smith & Binder, 2003).

Theme II: RG Message Effectiveness

The second major theme that emerged from the REA was the effectiveness of RG messages in moderating gambling behavior. Two sub-themes were identified: 1) the type of RG message (e.g., self-appraisal vs. informative) influences its effectiveness and 2) how message delivery formats (e.g., static vs. dynamic) and perceived threat levels influence players’ responses.

Type of RG Message

Auer and Griffiths (2015) compared the effectiveness of an “enhanced” pop-up message incorporating self-appraisal, normative feedback, and behavioral recommendations to a simple RG message. The enhanced message stated: “We would like to inform you that you have just played 1,000 slot games. Only a few people play more than 1,000 slot games. The chance of winning does not increase with the duration of the session. Taking a break often helps, and you can choose the duration of the break” (pp. 3–4). In contrast, the simple message merely informed players they had played 1,000 games and asked if they wanted to continue. Players who received the enhanced message were more likely to stop gambling, though it remains unclear which component of the message (e.g., social norm feedback, self-appraisal) was most effective. It is possible that the social norm feedback helped challenge cognitive biases, or that the self-appraisal component facilitated reflection. Alternatively, a combination of these elements might be more impactful than any single component, though further research is needed to test this hypothesis.

Eight studies assessed the effect of different types of RG messages (i.e., self-appraisal and informative) on problematic behaviors. Although participants indicated they found self-appraisal messages and informative messages to be similarly useful (see Department of Social Services, 2014), players tended to recall self-appraisal messages to a greater extent than informative messages (Gainsbury et al., 2015). Moreover, self-appraisal messages had a greater influence on facilitating engagement with RG behaviors. That is, players who saw self-appraisal messages were more likely to decrease their betting speed and betting amounts and cashed out and left the gaming floor more often than players who saw informative messages. Similarly, Armstrong et al. (2018) found that negative self-appraisal messages (e.g., “You’re playing faster than most people. You’re playing at similar speeds to most problem gamblers”) were associated with decreased betting speed compared to negative informative or positive informative messages. This could be because this message involves normative influence by informing players that they are gambling more than the average player.

Caillon et al. (2021) also demonstrated that self-appraisal RG messages led to reduced gambling time and fewer illusions of control compared to a control group. In contrast, players exposed to informative messages showed no meaningful changes in gambling behaviors or cognitions. Some challenging informative messages, such as “Betting quickly equals losing quickly,” even increased betting persistence and losses. Similarly, Rockloff et al. (2024) reported that there were no differences in betting amounts, gambling urges, time spent gambling, risk-related cognitions, or gambling-related harms whether players were shown positive-emotional, norm-based, or generic RG messages. Thus, it appears that self-appraisal messages may be more useful in facilitating RG behaviors than other types of messages.

RG Message Delivery Format

The effectiveness of RG messages also appears to be influenced by their delivery format (e.g., static vs. dynamic) and perceived threat level. Monaghan and Blaszczynski (2009) found that players retained RG message content better when it was delivered dynamically (e.g., as pop-up messages) rather than in static formats. Additionally, Munoz et al. (2010) found that players who viewed highly threatening messages (e.g., highlighting the long-term consequences of problem gambling) were more likely to change their behavior compared to those exposed to low-threat messages.

Exploratory Analyses

To explore the literature further, we coded for whether: 1) the messages used in the reported research was provided; 2) the message was grounded in theory; 3) the message was a compound message (i.e., there was more than one type of message in a given RG message); 4) researchers had considered demographic information (age, gender) in their analyses; and 5) the tone of the messages was assessed (e.g., whether the message was supportive or condescending).

Only eight of the 20 studies indicated what kind of RG message they used (i.e., informative messages, self-appraisal messages, and normative feedback; Armstrong et al., 2018; Auer & Griffiths, 2015; Caillon et al., 2021; Department of Social Services, 2014; Gainsbury et al., 2015; Houghton & Moss, 2024; Monaghan & Blaszczynski, 2009; Rockloff et al., 2024). Moreover, two of the 20 studies used compound messages (Auer & Griffiths, 2015; Davies et al., 2022) and two studies did not show the messages that they used. Only five of the 20 studies generated their messages based on any kind of relevant theory (Armstrong et al., 2018; Davies et al., 2022; Department of Social Services, 2014; Munoz et al., 2010; Mutti-Packer et al., 2022). Approximately half of the identified studies (N = 11) considered demographic characteristics. Specifically, six studies considered age of the participant (Armstrong et al., 2018; Gainsbury et al., 2015, 2018; Newall et al., 2023; Riley-Smith & Binder, 2003; Rockloff et al., 2024), five examined the gender of the participant(Armstrong et al., 2018; Gainsbury et al., 2015; Newall et al., 2023; Riley-Smith & Binder, 2003; Rockloff et al., 2024), four evaluated disordered gambling symptomatology (Armstrong et al., 2018; Houghton & Moss, 2024; Michalska et al., 2020; Thomas et al., 2015), and two studies looked at gambling game preferences (Gainsbury et al., 2018; Morvannou et al., 2020). Lastly, only six studies regarded the tone of the message (Davies et al., 2022; Gainsbury et al., 2018; Marko et al., 2023; Munoz et al., 2010; Riley-Smith & Binder, 2003; Rockloff et al., 2024).

Discussion

In this REA, we synthesized and evaluated the research literature on RG messages to identify best practices for effective RG messages and promoting positive play. Two overarching themes emerged: 1) types of RG messages players preferred, and 2) the RG message effectiveness. In terms of the type of RG message, most players preferred messages that were not generic (e.g., “Gamble responsibly” messages). Instead, they preferred self-appraisal messages, which encouraged them to reflect on their gambling behaviors. In particular, players reported that messages that highlighted the risks associated with gambling were more effective in encouraging them to change their gambling behavior (Caillon et al., 2021). Taken together, the literature suggests that self-appraisal messages are more effective than informative or normative feedback messages in promoting safer gambling and supporting positive play.

The effectiveness of self-appraisal messages may be explained through several well-established psychological theories. According to dual-process models of decision-making, such as the Elaboration Likelihood Model (Petty & Cacioppo, 1986), self-appraisal messages engage the central route to persuasion, prompting more deliberate cognitive processing and increasing the likelihood of lasting behavior change. Rather than passively receiving information, players are encouraged to actively reflect on their current gambling behaviors. This aligns with self-regulation theory (Baumeister & Vohs, 2007), which emphasizes the importance of self-monitoring for impulse control and goal alignment. By prompting real-time reflection—for example, through questions like “Have you been playing longer than planned?”—self-appraisal messages may disrupt dissociative gambling and foster behavioral regulation. These introspective prompts are also more likely to be perceived as personally relevant, enhancing message salience, a key predictor of behavior change in health communication (Noar et al., 2007). Collectively, these theoretical frameworks help explain why self-appraisal messages often outperform more generic or didactic RG content.

Players’ preference for self-appraisal messages may also reflect a broader perception that RG tools are primarily intended for individuals with gambling problems. Indeed, research suggests that many players view RG content as relevant only for those experiencing harm (Bagot et al., 2021). This perception may shape how messages are interpreted. For regular or higher-risk players—who made up the majority of participants in the studies reviewed—self-appraisal messages may feel more legitimate and appropriately targeted. In contrast, low-risk players often reported that they saw little relevance in RG content, particularly messages emphasizing personal responsibility (Houghton & Moss, 2024; Morvannou et al., 2020; Thomas et al., 2015). These players perceived themselves as already engaging in healthy play and thus frequently ignored messages that seemed irrelevant to their gambling experience.

To address players’ misperceptions that RG messages are only intended for those living with a gambling disorder, it may be beneficial to clearly differentiate between messaging for disordered gambling and messages aimed at promoting positive, safer play. Although it is crucial to target players at risk of developing a gambling problem—given their heightened vulnerability to gambling-related harms—focusing RG messages solely on those with gambling disorders may inadvertently alienate players who do not experience such harms. This approach may also reinforce stigma surrounding RG (Miller & Thomas, 2018). We propose that separating messages about disordered gambling from messages that promote positive, responsible play could reduce stigma and enhance the appeal of RG tools across a wider audience. This could involve developing targeted messages for individuals seeking help for gambling disorders while also promoting messages that highlight the array of RG tools available, such as limit-setting features or player feedback quizzes, which may resonate with players at the lower end of the risk spectrum.

Results also point to the need to consider how demographic characteristics, such as age, and gambling styles influence players’ preferences for RG messages. Younger players tend to respond more positively to non-accusatory language (Davies et al., 2022) and messages that provide concrete actionable tips to improve their gambling practices, such as setting a spending limit to avoid spending more than intended (Gainsbury et al., 2018). Consequently, RG messages for younger players could emphasize how RG programs help them save money and manage their play effectively, fostering greater engagement with these tools.

Older players, although being more likely to endorse messages that highlight the consequences of problem gambling, may respond less favorably to messages with a negative tone. For this demographic, using more optimistic and upbeat language could be more effective. Additionally, skill-based and frequent players generally prefer simple and generic RG messages (Gainsbury et al., 2018). Despite generic RG messages appealing to frequent players due to their non-intrusive nature, relying solely on these messages limits the reach of RG initiatives. Frequent and skill-based players, who may be more susceptible to gambling fallacies or cognitive distortions, may benefit more from self-appraisal messages that promote self-reflection and explicitly address risks like debt and addiction (Caillon et al., 2021). To engage a broader audience, gambling operators should consider using a combination of generic RG messages and targeted self-appraisal messages on signage and EGMs. This approach balances simplicity and impact, ensuring that RG messaging resonates with diverse player demographics and gambling behaviors.

A Framework for the Development of RG Messages

Building on the findings of the rapid evidence assessment (REA) and established theories of health communication and behavior change, we propose a Responsible Gambling Message Development Framework (RGMDF; see Fig. 2). This framework is designed as a practical and theoretically informed tool to help gambling operators and allied organizations develop more effective RG messages. It draws on dual-process models of persuasion (e.g., Elaboration Likelihood Model; Petty & Cacioppo, 1986) and behavior change theories such as the Health Belief Model (Rosenstock, 1974), which highlight the importance of tailoring messages to an individual’s capability, opportunity, and motivation to change behavior. The RGMDF consists of six core elements: three focused on player profile—risk level, game choice, and demographic characteristics—and three on message design—tone, touchpoint, and frequency.

Fig. 2.

RG message development framework

At its core, the RGMDF emphasizes the need to segment audiences based on their risk profile and tailor messages accordingly. For instance, for lower-risk players, messages may focus on enhancing gambling literacy and promoting positive play strategies such as pre-commitment planning. In contrast, messages targeting high-risk or harmed players should prioritize encouraging help-seeking, promoting treatment resources, or prompting cessation. Game choice also informs message content: skill-based gamblers may be more responsive to messages that correct cognitive distortions, while frequent players may benefit from self-appraisal prompts. Demographic characteristics further nuance message delivery—for example, younger players may prefer direct, action-oriented messaging, while older players may respond better to supportive and reflective tones.

The framework is also grounded in strategic communication principles. “Tone” refers to the emotional and linguistic framing of a message—messages should avoid being moralizing or stigmatizing and instead promote autonomy and reflection. “Touchpoint” refers to where a message is delivered (e.g., on-device pop-ups, ATM screens, loyalty app notifications), and "frequency" refers to how often a message is delivered to maximize recall and minimize fatigue. For example, in-play messages may promote real-time reflection during a gambling session, while consistent messaging on ATMs and cash redemption kiosks ensures visibility to those potentially in distress.

To illustrate the practical application of the RGMDF, consider a casino operator developing a campaign to promote limit-setting among slot players. Based on the framework, the operator would identify the most relevant target group (e.g., casual or newer players), then determine the tone (e.g., positive and empowering), touchpoint (e.g., digital kiosks and loyalty app notifications), and frequency (e.g., a consistent message at login and one during play). They would also consider community demographics (e.g., tailoring content to reflect cultural values or financial norms), thereby sharpening message resonance. This structured approach increases the likelihood that messages are perceived as personally relevant and actionable.

Finally, we stress that the RGMDF is not a static prescription but a flexible, evaluative tool. Its success hinges on continuous monitoring and testing to determine whether messages achieve intended outcomes, such as increased use of RG tools or calls to helplines. Evaluation should be multimethod, drawing on surveys, behavioral data, and player feedback. Moreover, message development and review should include perspectives from researchers, people with lived experience, player-facing casino employees, and behavioral science experts. Through this collaborative, evidence-informed approach, the RGMDF can help foster more personalized, effective, and impactful RG strategies.

Future Directions

Although this REA provides important insight into the landscape of RG message research, it is not intended to serve as a comprehensive systematic review. Rather, it offers a time-sensitive, practice-oriented synthesis designed to inform current implementation strategies and regulatory efforts. We acknowledge that future systematic reviews or meta-analyses may be warranted as the literature grows and converges on more standardized outcomes and methodologies. Nevertheless, we believe the current review offers a valuable stepping stone by identifying evidence-based themes and highlighting areas for further research and message development.

Other limitations warrant consideration. First, most participants across the 19 studies played online gambling games. Consequently, the findings may not be fully generalizable to land-based gambling environments, where RG messages are typically delivered via physical signage or pop-up messages that are restricted to EGMs in land-based venues. The effectiveness of RG messages in online contexts may be partly attributable to their delivery format rather than their content, underscoring the need for research that examines RG messaging across diverse gambling environments.

Additionally, many of the studies included in Theme I relied on qualitative research methods. These findings primarily reflect participants’ recommendations on the types of RG messages they believe would be effective, rather than empirical evaluations of whether the suggested content successfully reduces gambling-related harms. This gap highlights the need for more experimental and longitudinal research to rigorously assess the utility of participant-recommended RG messages in promoting safer gambling behaviors and mitigating harm.

Through the development of the RGMDF, we demonstrated the importance of considering factors such as tone, player demographics, risk of a gambling problem, and play style when designing RG messages. However, none of the studies reviewed incorporated all these elements simultaneously. Neglecting factors like age, gender, play style, and risk category likely limits the effectiveness of RG messages, reducing their ability to engage a broad range of players. A “one-size-fits-all” approach to RG messaging is insufficient, and future research should explore how a variety of tailored messages can effectively reach diverse player demographics.

Moreover, many of the analyzed RG messages were compound in nature, combining multiple strategies or containing elements that could be classified as various types of messages. This lack of clarity complicates efforts to isolate which specific components are most effective. While combining different types of RG messages may enhance their overall impact, this possibility remains unexplored. Future studies should aim to disentangle the effects of message content and delivery to determine the optimal balance for promoting positive play.

This REA could be used as a springboard for collaborative initiatives between researchers and the gambling industry. Such efforts should aim to develop RG messages that are grounded in theory, systematically tested, and tailored to diverse player needs. By addressing these gaps, future research can contribute to creating consistent, evidence-based messaging strategies that foster safer gambling practices and reduce gambling-related harms. Our proposed RGMDF may serve as a guide to facilitate the development of systematically tested RG messages. However, it should first be systematically evaluated.

Conclusion

This REA provides valuable insights into the design and delivery of RG messages, emphasizing the importance of tailoring messages to player demographics, risk levels, and gambling styles to promote safer gambling behaviors and reduce harm. The findings highlight what is currently known about the effectiveness of self-appraisal messages and the role of dynamic delivery formats, such as pop-ups, in enhancing message impact. However, gaps in the literature, including an over-reliance on online gambling contexts and a lack of systematic evaluation of RG message components, underscore the need for further research and collaboration.

The proposed RGMDF offers a practical guide for creating effective RG messages by integrating considerations such as tone, content, and touchpoints. By implementing player-centric, evidence-based strategies, this framework has the potential to transform harm reduction efforts and foster positive play across diverse gambling environments. Through bridging research and practice, RG messages can become an integral component of a comprehensive approach to responsible gambling.

Funding

Gray Gaudett has received funding from the Massachusetts Gambling Commission (US), GamRes (Canada), Carleton University (Canada), and Mitacs (Canada). Paul Pellizzari has not received any funding. Dr. Richard Wood, over the last 25 years, has been working as an applied research psychologist focusing on responsible gambling strategy development and evaluation. Since 2004 he has been the President of Gamres Limited, a Canadian research and consultancy company that develops and evaluates responsible gambling initiatives. He has undertaken responsible gambling consultations and research projects for more than 90 gaming companies, regulators and government organizations world-wide. Dr. Michael Wohl, over the past five years, has received research funding from Alberta Gambling Research Institute (Canada), Carleton University (Canada), Canadian Responsible Gambling Association (Canada), Canadian Centre on Substance Abuse and Addiction (Canada), Canadian Society of Addiction Medicine (Canada), Gambling Research Exchange Ontario (Canada), International Center for Responsible Gaming (US), Massachusetts Gambling Commission (US), Ontario Lottery and Gaming (Canada), and Ontario Ministry of Health and Long-Term Care (Canada). Dr. Wohl has received speaker/travel honorarium from National Association for Gambling Studies (Australia), FunStage (Europe), Indiana Council on Problem Gambling (US), International Center for Responsible Gaming (US), Massachusetts Council on Compulsive Gambling (US), McGill University (Canada), New York Council on Problem Gambling (US), and Safe Foundation (US). He has received fees for academic services from, Atlantic Lottery and Gaming Corporation (Canada), International Center for Responsible Gaming (US), New South Wales Government (Australia), Massachusetts Gambling Commission (US), Ontario Lottery and Gaming (Canada), and Responsible Gambling Council (Canada). Dr. Wohl has also received consulting fees from Aristocrat Leisure Limited (Australia), Atlantic Lottery and Gaming Corporation (Canada), FanDuel (US), GamRes (Canada), Massachusetts Gaming Commission (US), and Ontario Lottery and Gaming (Canada).

Declarations

Competing interests

Financial support for this study came from a GamRes Limited and a Mitacs Business Strategy Internship (IT42668).