Abstract
Aims
To evaluate whether a personalized and normative feedback (PNF) intervention for responsible gambling increases gambling insight as well as intention and self‐efficacy to engage in responsible gambling and behavioral engagement.
Design
Two‐arm randomized controlled trial. Outcome measurements occurred post‐randomization and 3 months later.
Setting
Online, Canada.
Participants
Canadian community members who gambled at a land‐based casino or online in the last 3 months [61.9% men; mean age = 56.52 (standard deviation = 14.80)] recruited via an online panel (n = 4091).
Intervention and comparator
Participants were randomized to receive PNF (n = 1940) or no feedback (n = 2151).
Measurements
Primary outcomes included gambling insight, intentions and self‐efficacy to engage in seven responsible gambling behaviors post‐randomization as well as engagement in these behaviors during the 3‐month follow‐up.
Findings
Post‐intervention, participants receiving PNF (relative to no feedback) had greater gambling insight (d = 0.32, P = 4.59 e‐25) as well as greater intentions and self‐efficacy to learn about how the games they play work (d intention = 0.31, P = 4.92 e‐24; d self‐efficacy = 0.25, P = 4.35 e‐16), learn about the odds of winning at these games (d intention = 0.30, P = 1.43 e‐21; d self‐efficacy = 0.25, P = 2.13 e‐15) and use operator‐provided tools to help limit their gambling (d intention = 0.20, P = 1.36 e‐10; d self‐efficacy = 0.18, P = 3.92 e‐9). However, post‐intervention differences in intention and self‐efficacy to limit time and money spent gambling, openness about gambling with others and balancing gambling with other activities were not observed. Meaningful increases in behavioral engagement 3 months later were observed but were not significant.
Conclusions
PNF for responsible gambling (relative to no feedback) appears to increase gambling insight, intentions and self‐efficacy to engage in gambling literacy and use of limit‐setting tools. Exploratory analyses indicated receiving PNF (relative to no feedback) led to behavioral changes during the 3‐month follow‐up period.
Keywords: gambling, intentions, intervention, personalized and normative feedback, positive play, responsible gambling, self‐efficacy
INTRODUCTION
Gambling addiction presents significant and pervasive challenges, profoundly disrupting, among other things, mental and physical health, relationships and financial stability [1, 2]. Personalized and normative feedback (PNF) interventions offer a promising approach for individuals seeking to overcome addiction. These interventions highlight discrepancies between individuals' self‐reported addictive behaviors and the normative behaviors of their peer group. A recent systematic review and meta‐analysis demonstrated that PNF interventions are effective in reducing the short‐term severity of problem gambling symptoms [3]. Their effectiveness lies in helping individuals living with addiction recognize and correct misperceptions about their behavior, thereby motivating meaningful behavior change.
Despite the intervention utility of PNF for people living with gambling problems [4, 5], there is a notable gap in research exploring whether PNF interventions can promote responsible gambling among vulnerable individuals who have not yet developed a gambling disorder. The extant literature has, however, shown that PNF interventions directed at individuals who do not engage in an addictive behavior at problematic levels does not increase their engagement in the addictive behavior. For instance, a re‐analysis of three PNF intervention studies for alcohol use found no increase in alcohol consumption among light‐drinking university students when informed that they drank below the descriptive drinking norm of university students [6]. What remains unclear is why these light‐drinking students maintained their low levels of alcohol consumption. A potential reason is that the PNF‐based intervention promoted non‐problematic (i.e. responsible) engagement in addictive behavior.
Assessing the potential preventive utility of PNF interventions among people who do not currently have a gambling disorder adds significant value in at least two ways. First, it extends the reach of PNF interventions, potentially benefiting not only those with problematic addictive behaviors (i.e. gambling), but also those currently engaging in these behaviors at non‐problematic levels. Second, focusing on responsible engagement offers a valuable opportunity to investigate the psychological factors that may encourage and sustain responsible behavior patterns of people engaging in addictive behaviors. Two key factors are behavioral intentions to engage in specific responsible actions and perceived self‐efficacy to do so. Both factors are central to various theories of behavioral engagement. For example, the theory of reasoned action [7], and later, the theory of planned behavior [8] both positioned behavioral intention as the primary determinant of whether a person will perform a behavior. In the theory of reasoned action, perceived descriptive norms (i.e. how much others engage in the behavior) predict behavioral intention. In the theory of planned behavior, perceived self‐efficacy is a predictor of behavioral intention. Similarly, in social cognitive theory [9, 10], perceived self‐efficacy directly influences behavior. In social cognitive theory, people are motivated to develop and control their actions through self‐reflection, which helps them make sense of their experiences, explore their beliefs and engage in self‐evaluation. Therefore, providing PNF to people who gamble (but not at problematic levels) and promoting responsible gambling should enhance people's self‐efficacy to act responsibly.
In the current research, we tested the potential preventative utility of PNF using the novel Positive Play Quiz (Quiz), which is grounded in the positive play approach to responsible gambling [11]. The positive play approach defines responsible gambling as a multidimensional construct comprising two fundamental beliefs—personal responsibility and gambling literacy—and two behaviors—honesty and control and pre‐commitment [11]. Personal responsibility refers to the extent to which an individual believes they should take ownership of their gambling behavior. Gambling literacy involves having an accurate understanding of the nature of gambling. Honesty and control pertain to being truthful with others about one's gambling behavior and feeling in control of it. Precommitment involves considering and setting limits on the amount of money and time spent gambling. These four dimensions are measured using the reliable and validated Positive Play Scale [11, 12]. Additionally, there is growing empirical support that positive play is fundamental to the prevention of disordered gambling [11, 12, 13, 14, 15, 16]. The Quiz was developed as a self‐test based on the Positive Play Scale that provides both personalized and normative feedback on each positive play dimension.
We hypothesized that people would benefit from PNF based on how they score on each positive play dimension as well as normative scores for each dimension. Such feedback should result in increased knowledge about their own (as well as most others) gambling, thereby increasing their intentions and self‐efficacy to take action to gamble responsibly in the future.
METHOD
All materials and data from the present research are publicly available via the Open Science Framework (OSF): https://osf.io/6qkyh. The research received ethical approval from Carleton University's Research Ethics Board–B.
Trial design
A parallel randomized controlled design with a 3‐month follow‐up assessment in an online setting was conducted.
Participants
People who gamble completed a short online questionnaire to check their eligibility. They were eligible to participate if they resided in Canada and gambled in the last 3 months either at a land‐based casino, over the Internet, or both.
Procedure
Participants were recruited from the AskingCanadians panel, which informs panellists of available consumer research studies and asks screening questions. For this study, panellists were asked whether they gamble (among an array of other questions not relevant to the current study). Those who answered ‘yes’ were directed to complete a short questionnaire to check their eligibility. This recruitment method prevents participants from falsifying eligibility, which enhances internal validity and reduces self‐selection bias.
Between June and August 2023, eligible participants provided consent and completed a baseline questionnaire before being randomized to either the intervention or control condition. Post‐randomization, they completed a questionnaire after which they received loyalty points from AskingCanadians. They were also asked for consent to be re‐contacted for the 3‐month follow‐up assessment. Three months later, consenting participants completed an online questionnaire. They again received loyalty points, were debriefed and asked to provide consent to use their data.
Randomization and blinding
Participants were assigned either to the intervention or control condition using simple randomization with a 1:1 allocation ratio in Qualtrics. The researchers and participants were both blind to the condition to which participants were allocated. Only participants were blind to the condition to which they were allocated when they completed the 3‐month follow‐up assessment.
Intervention
In the intervention condition, participants received PNF feedback for each Positive Play subscale via the Quiz. 1 The normative Positive Play subscale scores were obtained from a large Canadian national study of the Positive Play Scale [12]. PNF scores for each Positive Play subscale were juxtaposed. Participants were then provided brief responsible gambling tips for each subscale. In the control condition, participants did not receive any feedback.
Measures
Baseline measures
Participants completed demographic items to indicate their age, gender identity, cultural identity, province or territory of residence, personal income before taxes and transfers and their education level (see Table 1). They completed items that assessed their gambling involvement in the past year, including the games they played and their frequency of play (see Table 2). For each game, participants indicated their frequency of play using a response scale with endpoints never (1) and a few times a week or more (7). They also completed the validated Brief Biosocial Gambling Screen (BBGS) [17, 18], which we used to assess risk for problem gambling. Items were: ‘During the past 12 months, have you become restless, irritable or anxious when trying to stop/cut down on gambling?,’ ‘During the past 12 months, have you tried to keep your family or friends from knowing how much you gambled?,’ and ‘During the past 12 months did you have such financial trouble as a result of your gambling that you had to get help with living expenses from family, friends or welfare?’ Participants responded to each item by indicating ‘yes’ or ‘no.’ The BBGS was scored in terms of the number of participants who indicated ‘yes’ to any single item, which corresponds to risk for gambling problems.
TABLE 1.
Sample demographic characteristics.
| Total sample | Control | Intervention | Difference | |
|---|---|---|---|---|
| Gender identity | n (%) | n (%) | n (%) | χ2(4) = 8.41, P = 0.08 |
| Man | 2533 (61.9) | 1365 (63.5) | 1168 (60.2) | |
| Woman | 1536 (37.5) | 778 (36.2) | 758 (39.1) | |
| Non‐binary | 13 (0.3) | 5 (0.2) | 8 (0.4) | |
| Prefer not to answer | 8 (0.2) | 2 (0.1) | 6 (0.3) | |
| Two‐spirited | 1 (0.02) | 1 (0.04) | – | |
| Cultural identity | n (%) | n (%) | n (%) | χ2(10) = 11.71, P = 0.31 |
| Indigenous | 60 (1.5) | 36 (1.7) | 24 (1.2) | |
| Black | 66 (1.6) | 36 (1.7) | 30 (1.5) | |
| Caucasian/White | 2919 (71.4) | 1535 (71.4) | 1384 (71.3) | |
| East Asian | 524 (12.8) | 274 (12.7) | 250 (12.9) | |
| Latin American | 55 (1.3) | 39 (1.8) | 16 (0.8) | |
| Pacific Islander/Polynesian | 9 (0.2) | 5 (0.2) | 4 (0.2) | |
| South Asian | 205 (5) | 103 (4.8) | 102 (5.3) | |
| Southeast Asian | 80 (2) | 36 (1.7) | 44 (2.3) | |
| West Asian | 35 (0.9) | 19 (0.9) | 16 (0.8) | |
| Multi‐ethnic | 70 (1.7) | 35 (1.6) | 35 (1.8) | |
| Other | 67 (1.6) | 33 (1.5) | 34 (1.8) | |
| Province or Territory of residence | n (%) | n (%) | n (%) | χ2(10) = 11.71, P = 0.47 |
| Alberta | 515 (12.6) | 252 (11.7) | 263 (13.6) | |
| British Columbia | 758 (18.5) | 404 (18.8) | 354 (18.2) | |
| Manitoba | 165 (4) | 90 (4.2) | 75 (3.9) | |
| New Brunswick | 79 (1.9) | 38 (1.8) | 41 (2.1) | |
| Newfoundland and Labrador | 53 (1.3) | 30 (1.4) | 23 (1.2) | |
| Northwest Territories | 1 (0.02) | – | 1 (0.1) | |
| Nova Scotia | 145 (3.5) | 81 (3.8) | 64 (3.3) | |
| Ontario | 2034 (49.7) | 1070 (49.7) | 964 (49.7) | |
| Prince Edward Island | 29 (0.7) | 15 (0.7) | 14 (0.7) | |
| Quebec | 204 (5) | 104 (4.8) | 100 (5.2) | |
| Saskatchewan | 103 (2.5) | 65 (3) | 38 (2) | |
| Yukon | 2 (0.04) | 1 (0.04) | 1 (0.1) | |
| Age, income and education | M (SD) | M (SD) | M (SD) | |
| Age | 56.52 (14.80) | 56.68 (14.68) | 56.35 (14.93) | t (4053) = −0.73, P = 0.47 |
| Self‐reported personal income | 7.72 (3.09) | 7.68 (3.06) | 7.76 (3.12) | t (4046) = 0.80, P = 0.43 |
| Education level | 3.57 (1.21) | 3.56 (1.24) | 3.58 (1.18) | t (4079) = 0.47, P = 0.64 |
Note: One participant did not report their cultural identity and one participant did not report their province or territory of residence. Thirty‐six participants did not report their age or did not do so correctly, 43 participants did not report their personal income and 10 did not report their level of education. Degrees of freedom vary because of missing data. A Cohen's d‐value of |0.15| or greater that has a P‐value <0.0007 was considered statistically significant in the current research. Total n = 4091.
TABLE 2.
Sample gambling characteristics in terms of gambling involvement, positive play and problem gambling risk.
| Total sample | Control | Intervention | Difference | ||||
|---|---|---|---|---|---|---|---|
| M (SD) | n | M (SD) | n | M (SD) | n | ||
| Gambling involvement | |||||||
| Lottery draw games (e.g. Powerball, Mega Millions; not on the internet) | 4.00 (2.05) | 4087 | 4.02 (2.07) | 2147 | 3.99 (2.03) | 1940 | t (4085) = −0.47, P = 0.64 |
| Scratch tickets (not on the internet) | 3.14 (1.77) | 4081 | 3.17 (1.79) | 2143 | 3.12 (1.76) | 1938 | t (4079) = −0.73, P = 0.47 |
| Sports betting (not on the Internet) | 1.88 (1.56) | 4080 | 1.88 (1.55) | 2143 | 1.88 (1.57) | 1937 | t (4078) = −0.07, P = 0.95 |
| Horse racing (not on the internet) | 1.53 (1.18) | 4079 | 1.55 (1.21) | 2143 | 1.51 (1.15) | 1936 | t (4077) = −0.91, P = 0.36 |
| Bingo (not on the internet) | 1.69 (1.33) | 4079 | 1.70 (1.35) | 2143 | 1.67 (1.31) | 1936 | t (4077) = −0.74, P = 0.46 |
| Pull‐tabs (not on the internet) | 1.61 (1.25) | 4077 | 1.63 (1.28) | 2142 | 1.58 (1.21) | 1935 | t (4075) = −1.47, P = 0.14 |
| Slot machines (not on the internet) | 2.82 (1.52) | 4080 | 2.81 (1.52) | 2145 | 2.82 (1.53) | 1935 | t (4078) = 0.24, P = 0.81 |
| Casino style card or table games (e.g. poker, blackjack, roulette; not on the internet) | 2.02 (1.42) | 4081 | 2.02 (1.43) | 2145 | 2.03 (1.42) | 1936 | t (4079) = 0.20, P = 0.84 |
| Video lottery terminals (not on the internet) | 1.77 (1.41) | 4080 | 1.78 (1.44) | 2145 | 1.76 (1.37) | 1935 | t (4078) = −0.45, P = 0.66 |
| Lottery draw games (e.g. Powerball, Mega Millions; on the internet) | 2.73 (2.12) | 4087 | 2.79 (2.15) | 2149 | 2.67 (2.08) | 1938 | t (4085) = −1.79, P = 0.07 |
| Scratch tickets (on the internet) | 1.83 (1.53) | 4084 | 1.86 (1.58) | 2150 | 1.78 (1.47) | 1934 | t (4082) = −1.66, P = 0.10 |
| Sports betting (on the internet) | 1.86 (1.64) | 4085 | 1.87 (1.65) | 2149 | 1.85 (1.63) | 1936 | t (4083) = −0.36, P = 0.72 |
| Horse racing (on the internet) | 1.45 (1.23) | 4081 | 1.48 (1.28) | 2145 | 1.41 (1.17) | 1936 | t (4079) = −1.89, P = 0.06 |
| Bingo (on the internet) | 1.52 (1.29) | 4083 | 1.54 (1.32) | 2148 | 1.50 (1.26) | 1935 | t (4081) = −0.89, P = 0.38 |
| Pull‐tabs (on the internet) | 1.38 (1.12) | 4083 | 1.41 (1.17) | 2147 | 1.36 (1.07) | 1936 | t (4081) = −1.49, P = 0.14 |
| Slot machines (on the internet) | 1.99 (1.67) | 4083 | 2.01 (1.70) | 2148 | 1.96 (1.64) | 1935 | t (4081) = −1.01, Pp = 0.31 |
| Casino style card or table games (e.g. poker, blackjack, roulette; on the internet) | 1.82 (1.53) | 4084 | 1.82 (1.54) | 2149 | 1.81 (1.52) | 1935 | t (4082) = −0.32, P = 0.75 |
| Positive Play and problem gambling risk | |||||||
| Gambling literacy | 5.86 (1.24) | 4086 | 5.85 (1.31) | 2151 | 5.87 (1.16) | 1935 | t (4086) = 0.59, P = 0.55 |
| Personal responsibility | 6.41 (0.99) | 4086 | 6.29 (1.10) | 2151 | 6.54 (0.84) | 1935 | t (4084) = 7.88, P = 4.32 e‐15 |
| Precommitment | 6.22 (1.12) | 4086 | 6.19 (1.15) | 2151 | 6.26 (1.08) | 1935 | t (4082) = 1.84, P = 0.07 |
| Honesty and control | 6.14 (1.28) | 4086 | 6.13 (1.27) | 2151 | 6.14 (1.29) | 1935 | t (4084) = 0.22, P = 0.82 |
| Percentage of participants who endorsed one or more of the three BBGS items | 15.7% | 4088 | 15.9% | 2149 | 15.4% | 1939 | t (4086) = −0.52, P = 0.60 |
Note: A Cohen's d‐value of |0.15| or greater that has a P‐value <0.0007 was considered statistically significant in the current research. A bolded value in the intervention condition is significantly larger than its corresponding value in the control condition.
Abbreviation: BBGS, Brief Biosocial Gambling Screen.
Participants also completed the Positive Play Scale [11, 12]. The scale includes four subscales: gambling literacy (three items; e.g. ‘Gambling is not a good way to make money’), personal responsibility (four items; e.g. ‘It's my responsibility to spend only money that I can afford to lose’), honesty and control (three items; e.g., ‘I felt in control of my gambling behavior’) and precommitment (four items; e.g. ‘I only gambled with MONEY that I could afford to lose’). Participants responded to the gambling literacy and personal responsibility items using a response scale with endpoints strongly disagree (1) and strongly agree (7). Participants responded to the honesty and control and precommitment items using a response scale with endpoints never (1) and very much (7). The Positive Play subscales had good internal consistency reliability (αs ranged from 0.71–0.91) and items for each subscale were averaged with higher scores indicating higher levels of the corresponding construct.
Primary outcomes
The primary outcomes were assessed post‐intervention and 3 months later. Participants completed items we developed to assess their intentions to engage in seven responsible gambling behaviors over the next 3 months. The seven responsible gambling behaviors are listed in Table 3. For each behavior, participants indicated their plans to engage in the behavior using a response scale with endpoints not at all (1) and extremely likely (7). Similarly, participants indicated their self‐efficacy to engage in each behavior over the next 3 months. They did so by indicating their degree of confidence in their ability to engage in each behavior using a response scale with endpoints not confident (1) and extremely confident (11).
TABLE 3.
Descriptive and inferential statistics for intentions, self‐efficacy and insight as a function of receiving PNF or no feedback.
| Control | Intervention | Difference | |||||
|---|---|---|---|---|---|---|---|
| M (SD) | n | M (SD) | n | M and 95% CI | P | d | |
| Over the next 3 months, I plan to… | |||||||
| …learn more about how specific games that I play work | 2.03 (1.19) | 2150 | 2.41 (1.21) | 1938 | 0.38 [0.30, 0.46] | 4.92 e‐24 | 0.31 |
| …learn more about the odds of winning for the games that I play | 2.06 (1.23) | 2150 | 2.43 (1.25) | 1937 | 0.37 [0.29, 0.45] | 1.43 e‐21 | 0.30 |
| …ensure I limit how much time I spend gambling | 3.37 (1.39) | 2150 | 3.53 (1.25) | 1937 | 0.16 [0.08, 0.26] | 6.98 e‐5 | 0.18 |
| …ensure I limit how much money I spend gambling | 3.69 (1.34) | 2149 | 3.79 (1.20) | 1935 | 0.10 [0.02, 0.18] | 0.009 | 0.14 |
| …be open with others (e.g. friends, family) about my gambling | 3.73 (1.34) | 2149 | 3.73 (1.22) | 1935 | <0.01 [−0.08, 0.08] | 0.98 | <0.01 |
| …use responsible gambling tools (e.g. set a time or spend limit) supplied by the gaming companies to help me control my gambling | 2.85 (1.53) | 2149 | 3.15 (1.40) | 1935 | 0.29 [0.20, 0.38] | 1.36 e‐10 | 0.20 |
| …balance my gambling with other recreational activities | 3.78 (1.31) | 2150 | 3.81 (1.19) | 1935 | 0.02 [−0.04, 0.10] | 0.49 | 0.02 |
| How confident are you right now in your ability to do the following activities over the next 3 months? | |||||||
| Learn more about how specific games that I play work | 6.30 (3.69) | 2150 | 7.19 (3.25) | 1936 | 0.89 [0.67, 1.11] | 4.35 e‐16 | 0.25 |
| Learn more about the odds of winning for the games that I play | 6.24 (3.68) | 2147 | 7.11 (3.25) | 1935 | 0.87 [0.65, 1.08] | 2.13 e‐15 | 0.25 |
| Ensure I limit how much time I spend gambling | 8.77 (2.98) | 2149 | 8.90 (2.67) | 1931 | 0.13 [−0.04, 0.30] | 0.14 | 0.04 |
| Ensure I limit how much money I spend gambling | 9.10 (2.80) | 2149 | 9.14 (2.55) | 1929 | 0.04 [−0.12, 0.20] | 0.63 | 0.01 |
| Be open with others (e.g. friends, family) about my gambling | 8.80 (3.06) | 2148 | 8.83 (2.78) | 1928 | 0.03 [−0.14, 0.20] | 0.74 | 0.01 |
| Use responsible gambling tools (e.g. set a time or spend limit) supplied by the gaming companies to help me control my gambling | 6.07 (3.91) | 2148 | 6.77 (3.56) | 1928 | 0.69 [0.46, 0.92] | 3.92 e‐9 | 0.18 |
| Balance my gambling with other recreational activities | 8.84 (2.98) | 2149 | 8.95 (2.61) | 1931 | 0.10 [−0.06, 0.28] | 0.22 | 0.04 |
| Insight | |||||||
| Insight into one's gambling | 4.24 (1.53) | 2150 | 4.72 (1.43) | 1937 | 0.48 [0.39, 0.57] | 4.59 e‐25 | 0.32 |
Note: A Cohen's d‐value of |0.15| or greater that has a P‐value <0.0007 was considered statistically significant in the current research. A bolded value in the intervention condition is significantly larger than its corresponding value in the control condition.
Furthermore, participants completed a three‐item measure we developed to assess the amount of insight participants thought they gained from completing the questions about their gambling beliefs and behaviors. The opening stem was ‘Completing the questionnaires about my gambling beliefs and behaviors…’ and the items were ‘…made me more informed about different aspects of my gambling,’ ‘…gave me more insight into my gambling,’ and ‘…helped me better understand my gambling.’ Responses were averaged with higher scores indicating greater insight (α = 0.95).
Three months following the intervention, participants completed a survey that assessed whether they engaged in the seven responsible gambling behaviors during the past 3 months (see Table 4). For each behavior, participants indicated whether they engaged in the behavior using a response scale with endpoints not at all (1) and extremely likely (7).
TABLE 4.
Descriptive and inferential statistics for the 3‐month follow‐up analyses.
| Control | Intervention | Difference | |||||
|---|---|---|---|---|---|---|---|
| M (SD) | n | M (SD) | n | M and 95% CI | Pp | d | |
| Did you engage in the following activities in the past 3 months? | |||||||
| I learned more about how specific games that I play work | 2.41 (1.53) | 973 | 2.55 (1.56) | 901 | 0.14 [0.01, 0.28] | 0.04 | 0.09 |
| I learned more about the odds of winning for the games that I play | 2.48 (1.55) | 973 | 2.63 (1.59) | 901 | 0.14 [0.01, 0.29] | 0.04 | 0.09 |
| I limited how much time I spend gambling | 3.96 (1.75) | 973 | 4.02 (1.71) | 901 | 0.05 [−0.09, 0.21] | 0.47 | 0.03 |
| I limited how much money I spend gambling | 4.29 (1.69) | 974 | 4.37 (1.66) | 901 | 0.07 [−0.07, 0.23] | 0.32 | 0.04 |
| I was open with others (e.g. friends, family) about my gambling | 4.60 (1.63) | 974 | 4.52 (1.66) | 901 | −0.07 [−0.21, 0.07] | 0.34 | −0.04 |
| I used operator supplied responsible gambling tools/programs to help control my gambling | 2.20 (1.69) | 974 | 2.26 (1.71) | 901 | 0.05 [−0.09, 0.20] | 0.47 | 0.03 |
| I balanced my gambling with other recreational activities | 4.45 (1.63) | 974 | 4.49 (1.60) | 901 | 0.04 [−0.10, 0.18] | 0.57 | 0.02 |
Note: A Cohen's d‐value of |0.15| or greater that has a P‐value <0.0007 was considered statistically significant in the current research. A bolded value in the intervention condition is significantly larger than its corresponding value in the control condition.
Data quality check
At the end of the post‐intervention assessment, participants responded to two yes/no data quality check items: ‘Did you provide good, high quality responses?’ and ‘Did you provide honest responses to all items.’ Participants were told that they would be compensated regardless of how they responded. These questions were used to filter out low‐quality data.
Statistical analysis
Descriptive and confirmatory analyses were conducted using SPSS version 29. Confirmatory analyses were preregistered on OSF in August 2023 before completing the post‐randomization data collection: https://osf.io/qcu6j.
Sample size
Cohen's d value of 0.15 was used as the smallest effect size of interest in a priori power analyses described in the preregistration. We also used α = 0.0007 in the a priori power analyses to control for the Type I error‐rate based on a Bonferroni correction to the P‐value. Therefore, a given result must have a Cohen's d of 0.15 or larger to be considered practically meaningful and a corresponding P‐value <0.0007 to be considered statistically significant. With these parameters, the results of the power analysis indicated that a minimum total sample size of 784 would be needed for 80% power.
Descriptive analyses
Descriptive analyses were conducted to describe the demographic and gambling characteristics of the total sample as well as the subsamples that were randomized to the control or intervention conditions. Demographic characteristics included participants' age, gender identity, cultural identity, province or territory of residence, education level and personal income. Participants' gambling characteristics included their gambling involvement in various games and frequency of play for these games as well as their problem gambling risk and positive play subscale scores. We also examined differences in demographic and gambling characteristics as a function of condition.
Confirmatory analyses
The post‐randomization analyses involved independent samples t tests to examine differences in participants' insight into their gambling, intentions to engage in the seven responsible gambling behaviors in the next 3 months, and self‐efficacy to engage in the seven responsible gambling behaviors in the next 3 months as a function of condition (control vs. intervention). For the 3‐month follow‐up analyses, independent sample t tests were used to examine differences in participants' engagement in the seven responsible gambling behaviors during the 3‐month follow‐up period as a function of condition (control vs. intervention).
In line with the preregistered data analytic plan, bootstrapping with 5000 resamples was used to determine statistical significance of the t test results. The reason is that the dependent variables were non‐normally distributed in that they each had a statistically significant Shapiro–Wilk test (see OSF for these analyses). Therefore, the 95% bias‐corrected bootstrapped CI and P‐value of the t test were used to assess statistical significance.
Missing data handling
There was minimal missing data post‐randomization (0.1%–0.4%), so analyses were based on available data. The 3‐month follow‐up retention rate was 45.80%, exceeding the minimum sample size (n = 784) for 80% power per the preregistration. Following the preregistered plan, we found that participants' ethnicity and province were linked to the propensity of missing data, suggesting a missing at random mechanism. Sensitivity analyses using full information maximum likelihood, available on OSF, produced results virtually identical to those reported herein based on the available data.
RESULTS
Recruitment, losses, exclusions and participant flow
As shown in Figure 1, 15 763 panel members accessed our survey, with 7260 eligible and 4599 providing consent. After excluding 143 participants who dropped out before randomization, 81 who failed data quality checks and 284 who did not respond to the data quality check questions, there were 4091 participants post‐randomization. These participants were re‐contacted to complete the 3‐month follow‐up, with 2027 accessing the survey. After excluding 64 who did not consent to the follow‐up, 63 who did not respond to the request to provide consent to use their data, and 24 declined consent to use their data, 1876 participants were included in the 3‐month follow‐up analyses.
FIGURE 1.
CONSORT diagram depicting participant exclusions and allocation to condition.
Descriptive analyses
Demographic characteristics
Demographic characteristics for the total sample and by condition are reported in Table 1. Overall, 61.9% identified as men, with an average age of 56.5 years (SD = 17.4; range: 17–97) and 71.4% identified as Caucasian/White. Nearly half (49.7%) resided in Ontario, Canada. The average income was between $60 000 and $79 000, and the average education level was between ‘Trade or technical certificate’ and ‘Bachelor's degree.’ No statistically significant demographic differences were found between conditions (see Table 1).
Gambling characteristics
Gambling characteristics for the total sample and by condition are in Table 2. In the total sample, the most frequent game played was non‐internet lottery draws (M = 4.00, SD = 2.05, or ‘once a month’). The next most common were non‐internet scratch tickets (M = 3.14, SD = 1.77), non‐Internet slot machines (M = 2.82, SD = 1.52) and internet lottery draws (M = 2.72, SD = 2.12), all corresponding to ‘a few times a year.’ Casino card/table games (M = 2.02, SD = 1.42, ‘about once a year’) were fifth. Other games had average scores below 2. There were no statistically significant differences in gambling involvement between conditions (see Table 2).
In terms of problem gambling risk, 15.7% of the total sample endorsed one or more of the three BBGS items, with no statistically significant difference between conditions. For Positive Play, participants scored high on personal responsibility (M = 6.41, SD = 0.99) and relatively lower on gambling literacy (M = 5.86, SD = 1.24). Scores were similarly high on both precommitment (M = 6.22, SD = 1.22) and honesty and control (M = 6.14, SD = 1.28). The only statistically significant difference between conditions was a small effect for personal responsibility (d = 0.25) (Table 2). 2
Confirmatory analyses
Post‐randomization
Descriptive and inferential statistics for the post‐randomization analyses are reported in Table 3.
Insight into gambling. Participants in the intervention condition reported having more insight into their gambling because of completing the questionnaires about their gambling beliefs and behaviors compared to participants in the control condition.
Intentions to engage in responsible gambling behaviors. Participants in the intervention condition reported greater intentions to learn more about how the games they play work and learn about the odds of winning at the games they play over the next 3 months compared to those in the control condition. Similarly, participants in the intervention condition reported greater intentions to limit how much time they spend gambling in the next 3 months compared to those in the control condition. Likewise, participants in the intervention condition reported greater intentions in the next 3 months to use operator‐provided tools to help limit their gambling compared to those in the control condition. No other differences had a d‐value of 0.15 or larger and a P‐value <0.0007 (see Table 3).
Self‐efficacy to engage in responsible gambling behaviors. Participants in the intervention condition reported greater confidence in their ability to learn more about how the games they play work and their odds of winning at the games they play over the next 3 months compared to the control condition. Likewise, participants in the intervention condition reported greater confidence in their ability over the next 3 months to use operator‐provided tools to help limit their gambling compared to the control condition. No other differences had a d‐value of 0.15 or larger and a P‐value <0.0007 (see Table 3).
Three‐month follow‐up
Descriptive and inferential statistics for the 3‐month follow‐up analyses are reported in Table 4. Unexpectedly, none of the differences had a d‐value of 0.15 or larger and a P‐value <0.0007. Therefore, in exploratory analyses that were not pre‐registered, we used the two one‐sided t tests (TOST) procedure [19] to test for statistical equivalence between the intervention and control conditions, with the equivalence bound using d‐values of −0.14 and 0.14. As shown in Table 5, for all behaviors except for openness with others about one's gambling, the lower bound was statistically significant, whereas the upper bound was not. This indicates that equivalence between conditions was not supported, and that effect sizes for the difference may exceed a d‐value of 0.14. This suggests practical significance, which aligns with our expectations. In contrast, for openness, the upper bound was statistically significant, whereas the lower bound was not. This also indicates a lack of equivalence. However, in this case, the effect size may exceed a d‐value of −0.14, which suggests practical significance that is inconsistent with our expectations (less openness in the intervention relative to control condition).
TABLE 5.
Inferential statistics from the TOST procedure.
| Upper bound | Lower bound | |
|---|---|---|
| Did you engage in the following activities in the past 3 months? | ||
| I learned more about how specific games that I play work | t (1872) = 0.98, P = 0.16 | t (1872) = −5.07, P = 2.20 e‐7 |
| I learned more about the odds of winning for the games that I play | t (1873) = 0.98, P = 0.16 | t (1873) = −5.06, P = 2.30 e‐7 |
| I limited how much time I spend gambling | t (1872) = 2.31, P = 0.01 | t (1872) = −3.75, P = 9.11 e‐5 |
| I limited how much money I spend gambling | t (1873) = 2.04, P = 0.02 | t (1873) = −4.02, P = 3.03 e‐5 |
| I was open with others (e.g., friends, family) about my gambling | t (1873) = 3.99, P = 3.43 e‐5 | t (1873) = −2.08, P = 0.02 |
| I used operator supplied responsible gambling tools/programs to help control my gambling | t (1873) = 2.31, P = 0.01 | t (1873) = −3.75, P = 9.11 e‐5 |
| I balanced my gambling with other recreational activities | t (1873) = 2.47, P = 0.007 | t (1873) = −3.59, P = 1.69 e‐4 |
Note: Because we determined that a Cohen's d of |0.15| was the smallest effect of interest, we set the equivalence region to range from −0.14 to 0.14. A P‐value <0.0007 was considered statistically significant in the current research.
Abbreviation: TOST, two one‐sided t tests.
DISCUSSION
Participants who received PNF via the Quiz reported a significant increase in insight into their gambling beliefs and behaviors compared to those who received no feedback. Through the normative feedback, participants became more aware of how their gambling patterns compared to community norms. This heightened self‐awareness and critical reflection likely helped participants identify issues in their gambling practices that they might not have previously recognized. These findings underscore the internal validity of the PNF intervention and highlight the Quiz as an effective tool for enhancing self‐knowledge about responsible gambling.
Additionally, participants who received PNF reported greater intentions and self‐efficacy to engage in actions that promote gambling literacy compared to those who received no feedback. More specifically, those in the intervention (compared to control) condition expressed greater intentions and self‐efficacy to learn about how the games they play work and the odds of winning at these games. Gambling literacy is important because misconceptions about how games work and the odds of winning are key factors that help maintain excessive gambling involvement despite mounting financial losses [20]. Critically, understanding the mechanics and odds of the games played empowers people to make informed decisions about their gambling. When people recognize the low probabilities of winning money on the long run through gambling, they may be less likely to engage in high‐risk gambling behaviors [21].
Unexpectedly, however, differences in gambling literacy behaviors 3 months later did not reach significance based on our preregistered stringent thresholds. In exploratory analyses, evidence indicated that receiving PNF may have meaningfully increased gambling literacy behaviors, which warrants further investigation. The time gap between the PNF intervention and the behavior measurement may help explain these findings. A shorter timeframe might better capture the link between participants' intentions and their behaviors, providing a clearer picture of the PNF's impact. Additionally, connecting participants immediately to relevant gambling literacy resources following the PNF intervention may help strengthen the intention‐behavior link. Future research can examine these possibilities.
Moreover, participants who received PNF reported greater intentions and self‐efficacy to use operator‐provided limit‐setting tools compared to participants who received no feedback. This finding is noteworthy given the low usage of operator‐provided limit‐setting tools among people who gamble. Although there is evidence that limit‐setting tools are effective [22, 23], a systematic review highlighted that few people who gamble use these tools [24]. For example, only 1% of people who gambled on bwin—an internet sports betting site—used the limit‐setting feature during the 18‐month study period [25]. Similarly, in a more recent study of 40 000 Australians who gamble online, 83% indicated that they did not use responsible gambling tools [26]. These findings underscore the potential impact of PNF via the Quiz in promoting the adoption of operator‐provided limit‐setting tools.
Unexpectedly, however, the difference in use of operator‐provided limit‐setting tools 3‐months later did not reach significance based on our preregistered stringent thresholds. In exploratory analyses, evidence indicated that receiving PNF may have meaningfully increased the use of operator‐provided limit‐setting tools, which warrants further investigation. There are likely many reasons for the intention‐behavior gap for using operator‐provided limit‐setting tools. Participants may have faced barriers to the behavior that prevents them from acting on their intentions, including lack of time, competing priorities and insufficient information needed to access the operator‐provided tools. It would behoove future research to examine whether overcoming some of these barriers would facilitate the transformation of intentions and self‐efficacy into action.
Additionally, there were inconclusive results for responsible gambling limit‐setting behaviors. Although participants who received PNF reported greater intentions to limit their gambling time compared to participants who received no feedback, this finding was attenuated when baseline differences in personal responsibility were statistically controlled in the analysis (see footnote 2 and OSF for these results). Likewise, the difference in self‐efficacy to control gambling time was inconclusive. Similarly, the difference between those who received PNF and those who received no feedback in terms of intention and self‐efficacy to limit gambling expenditure were inconclusive. Similarly, differences in intentions and self‐efficacy for being open with others about gambling habits and balancing gambling with other recreational activities between participants who received PNF and those who received no feedback were inconclusive. These differences were inconclusive because they did not reach significance based on our preregistered stringent thresholds.
Furthermore, differences in limit‐setting behaviors, openness with others about gambling habits and balancing gambling with recreational activities between those who received PNF and those who did not receive feedback 3 months later did not reach significance based on our preregistered stringent thresholds. However, exploratory analyses for limit‐setting and balancing behaviors suggested that receiving PNF may have meaningfully increased these behaviors. In contrast, exploratory analyses suggested that receiving PNF may have meaningfully reduced openness, which is inconsistent with expectations (less openness in the intervention relative to control condition). These findings warrant further investigation. A possible direction for future research is to examine whether incorporating safer gambling messages that have been shown to have responsible gambling utility [27, 28] into the tips provided to people who complete the Quiz improves their responsible gambling intentions, self‐efficacy and behaviors.
The findings for limit‐setting, openness with others and balancing behaviors may be attributed, in part, to participants' high initial precommitment, honesty and control and personal responsibility Positive Play subscales scores. The high scores likely minimized discrepancies between participants' self‐assessments and community norms presented in the PNF intervention, reducing the impact of the feedback on behavioral engagement. Therefore, it would be helpful for future PNF research using the Quiz to consider the high scoring Positive Play subscales by designing tailored approaches to help increase responsible gambling engagement among people with varying degrees of precommitment, honesty and control and personal responsibility.
Regarding unintended harm, we cannot draw firm conclusions from the follow‐up analyses. However, there is suggestive evidence that some participants who received PNF reported being less open with others about their gambling 3 months later compared participants who received no feedback. In addition, it is possible that receiving PNF may have led some participants to place undue personal responsibility on their ability to control their gambling, leading to feelings of guilt or shame if they struggled to change their behavior. Although this concern is important, it does not invalidate our findings. Participants who received PNF reported a stronger intention to adopt harm‐reduction strategies, such as using a limit‐setting tool. Another potential issue is that some participants may have misinterpreted the normative feedback, either seeing it as encouragement to gamble more or as reassurance that their gambling was not problematic, even when it was harmful. Future research should investigate these potential unintended harms to better understand their impact and develop strategies to mitigate them.
The current research has several strengths, including preregistration, a recruitment strategy that minimized self‐selection bias, a large and demographically diverse sample and the use of random assignment. A limitation is the potential of recall bias in the retrospective accounts of participants' engagement in responsible gambling behaviors during the 3‐month follow‐up period. Another limitation is that participants largely played the lottery monthly as well as scratch tickets and slots a few times a year. The remaining games were of lower frequency. Therefore, the external validity of the results to people who gamble more frequently and who play other games is unknown. A third limitation is that participants, on average, were in the middle‐to‐older age category, and so external validity of the results to emerging and young adults who gamble is unknown.
In sum, receiving PNF via the Quiz increased participants' intentions and self‐efficacy to engage in gambling literacy behaviors and use operator‐provided limit‐setting tools that help control gambling. Although the heightened intentions and self‐efficacy did not translate into actual behavioral changes during the 3‐month follow‐up period based on our stringent criteria for significance, there was some evidence that receiving PNF may have meaningfully increased these behaviors. Taken together, these findings contribute valuable insight into how to foster intentions and self‐efficacy to engage in responsible gambling behaviors. Moreover, the findings highlight the critical need to bridge the intention‐behavior gap, paving the way for future research to investigate the factors that would help strengthen the connection between intentions and engagement in responsible gambling behavior.
AUTHOR CONTRIBUTIONS
Conceptualization: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl. Methodology: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl. Validation: Nassim Tabri. Formal analysis: Nassim Tabri. Investigation: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl. Resources: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl. Data curation: Nassim Tabri. Writing – Original Draft: Nassim Tabri. Writing – Review and Editing: Nassim Tabri, Richard T. A Wood and Michael J. A. Wohl. Visualization: Nassim Tabri. Project administration: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl. Funding acquisition: Nassim Tabri, Richard T. A. Wood and Michael J. A. Wohl.
PREREGISTRATION
The hypotheses and data analytic plan were preregistered on the Open Science Framework: https://osf.io/qcu6j.
DECLARATION OF INTERESTS
N.T. has received consulting fees from Gamres—a research and consultancy service that designs, implements and evaluates responsible gambling strategies for governments and the gambling industry. N.T. has also received funding for gambling research from Gambling Research Exchange Ontario (Canada), Canadian Society of Addiction Medicine (Canada) and International Center for Responsible Gaming (United States [US]). As well, N.T. has received research contracts from the Ontario Lottery Corporation (Canada), Massachusetts Gaming Commission (US) and Gambling Research Exchange Ontario (Canada). Additionally, N.T. has received a speaker/travel honorarium the International Center for Responsible Gaming (US). R.T.A.W. is the President of Gamres—a business that provides research and consultancy services for the design, implementation and evaluation of responsible gambling strategies for governments and the gambling industry. Gamres developed the Positive Play Quiz (Quiz) and so has a commercial interest. Initial funding for the development of the Quiz was provided by Lotto New Zealand and IGT Italy. Neither organization has placed any constraints on development, testing or publication of matters related to the Quiz. M.J.A.W. has received speaker/travel honorarium from Alberta Liquor Gaming Commission (Canada), Indiana Council on Problem Gambling (US); International Center for Responsible Gaming (US), National Association for Gambling Studies (Australia), Massachusetts Council on Compulsive Gambling (US), New York Council of Problem Gambling (US); Problem Gambling Ohio Network (US); Rhode Island Council on Problem Gambling (US); Safe Foundation (US); The Star Entertainment Group (Australia). He has also received consulting fees from Aristocrat Gaming (US); Atlantic Lottery and Gaming Corporation (Canada), GamRes (Canada), Massachusetts Gaming Commission (US), National Council on Problem Gambling (Singapore), New South Wales Government (Australia), Nova Scotia Gaming Corporation (Canada) and Ontario Lottery and Gaming (Canada). Last, he has worked with Massachusetts Gaming Commission on the transfer of data from gaming operators to the Massachusetts Gaming Commission.
ACKNOWLEDGEMENTS
The research was supported by a Seed Grant from the International Center for Responsible Gaming (ICRG) to Nassim Tabri. ICRG has placed no constraints on the development and implementation of the research or on publication.
Tabri N, Wood RTA, Wohl MJA. Effects of personalized and normative feedback via the Positive Play Quiz on responsible gambling intention, self‐efficacy and behavior: A randomized controlled trial. Addiction. 2025;120(5):1016–1027. 10.1111/add.16722
Funding information Seed Grant from the International Center for Responsible Gaming to Nassim Tabri.
Footnotes
For a demo, see: https://demo.positiveplay.org/quiz/.
The confirmatory analyses were repeated with personal responsibility as a control variable using regression to correct for the imbalance, and the results remained virtually identical. The statistical outputs for these analyses are available on the Open Science Framework: https://osf.io/qcu6j
DATA AVAILABILITY STATEMENT
All materials and data from the present research are publicly available via the Open Science Framework: https://osf.io/6qkyh.
REFERENCES
- 1. Langham E, Thorne H, Browne M, Donaldson P, Rose J, Rockloff M. Understanding gambling related harm: a proposed definition, conceptual framework, and taxonomy of harms. BMC Public Health. 2015;16(1):1–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Shannon K, Anjoul F, Blaszczynski A. Mapping the proportional distribution of gambling‐related harms in a clinical and community sample. Int Gambling Stud. 2017;17(3):366–385. 10.1080/14459795.2017.1333131 [DOI] [Google Scholar]
- 3. Saxton J, Rodda SN, Booth N, Merkouris SS, Dowling NA. The efficacy of personalized normative feedback interventions across addictions: a systematic review and meta‐analysis. Cardoso Cruz F, editor. PLoS ONE. 2021;16(4):e0248262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Luquiens A, Tanguy ML, Lagadec M, Benyamina A, Aubin HJ, Reynaud M. The efficacy of three modalities of internet‐based psychotherapy for non–treatment‐seeking online problem gamblers: a randomized controlled trial. J Med Internet Res. 2016;18(2):e36. 10.2196/jmir.4752 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Neighbors C, Rodriguez LM, Rinker DV, Gonzales RG, Agana M, Tackett JL, et al. Efficacy of personalized normative feedback as a brief intervention for college student gambling: a randomized controlled trial. J Consult Clin Psychol. 2015;83(3):500–511. 10.1037/a0039125 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Prince MA, Reid A, Carey KB, Neighbors C. Effects of normative feedback for drinkers who consume less than the norm: dodging the boomerang. Psychol Addict Behav. 2014;28(2):538–544. 10.1037/a0036402 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Ajzen I, Fishbein M. Understanding attitudes and predicting social behavior Englewood Cliffs, NJ: Prentice‐Hall; 1980. [Google Scholar]
- 8. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50(2):179–211. 10.1016/0749-5978(91)90020-T [DOI] [Google Scholar]
- 9. Bandura A. Self‐efficacy: the exercise of control W H Freeman/Times Books/Henry Holt & Co.; 1997. [Google Scholar]
- 10. Bandura A. Social foundations of thought and action: a social cognitive theory Prentice‐Hall, Inc.; 1986. [Google Scholar]
- 11. Wood RTA, Wohl MJA, Tabri N, Philander K. Measuring responsible gambling amongst players: development of the positive play scale. Front Psychol. 2017;8:227. 10.3389/fpsyg.2017.00227 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Tabri N, Wood RTA, Philander K, Wohl MJA. An examination of the validity and reliability of the positive play scale: findings from a Canadian national study. Int Gambling Stud. 2020;20(2):282–295. 10.1080/14459795.2020.1732442 [DOI] [Google Scholar]
- 13. Connolly AJ, Stall LM, Floyd CG, Kraus SW, Grubbs JB. Exploring the differences in positive play among various sports wagering behaviors. J Gambl Stud. 2024:1–24. 10.1007/s10899-024-10304-8 [DOI] [PubMed] [Google Scholar]
- 14. Delfabbro P, King DL, Georgiou N. Positive play and its relationship with gambling harms and benefits. J Behav Addict. 2020;9(2):363–370. 10.1556/2006.2020.00041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Tong KK, Chen JH, Wu AMS. Validation and application of the positive play scale adapted for Chinese gamblers: its relation to disordered gambling and gambling attitudes. Front Psychol. 2020;11:263. 10.3389/fpsyg.2020.00263 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. He M, Yang HM, Chen JH, Tong KK, Dang L, Wu S. Applying the protective gambling beliefs scale (PGBS) among Chinese gamblers and testing its association with responsible gambling. Int Gambling Stud. 2023;1–18. [Google Scholar]
- 17. Brett EI, Weinstock J, Burton S, Wenzel KR, Weber S, Moran S. Do theDSM‐5diagnostic revisions affect the psychometric properties of the brief biosocial gambling screen? Int Gambling Stud. 2014;14(3):447–456. 10.1080/14459795.2014.931449 [DOI] [Google Scholar]
- 18. Gebauer L, LaBrie R, Shaffer HJ. Optimizing DSM‐IV‐TR classification accuracy: a brief biosocial screen for detecting current gambling disorders among gamblers in the general household population. Can J Psychiatry. 2010;55(2):82–90. 10.1177/070674371005500204 [DOI] [PubMed] [Google Scholar]
- 19. Lakens D. Equivalence tests. Soc Psychol Personality Sci. 2017;8(4):355–362. 10.1177/1948550617697177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Blaszczynski A, Nower L. A pathways model of problem and pathological gambling. Addiction. 2002;97(5):487–499. 10.1046/j.1360-0443.2002.00015.x [DOI] [PubMed] [Google Scholar]
- 21. Keen B, Anjoul F, Blaszczynski A. How learning misconceptions can improve outcomes and youth engagement with gambling education programs. J Behav Addict. 2019;8(3):372–383. 10.1556/2006.8.2019.56 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Auer M, Griffiths MD. Voluntary limit setting and player choice in Most intense online gamblers: an empirical study of gambling behaviour. J Gambl Stud. 2012;29(4):647–660. 10.1007/s10899-012-9332-y [DOI] [PubMed] [Google Scholar]
- 23. Wohl MJA, Davis CG, Tabri N. Setting a hard (versus soft) monetary limit decreases expenditure: an assessment using player account data. Int Gambling Stud. 2023;1–17. [Google Scholar]
- 24. Tabri N, Wohl MJA, Xuereb S. Population‐based safer gambling/responsible gambling efforts. In: Hilbrecht M, editorPrevention and education evidence review: gambling‐related harm Guelph, Canada: Greo; 2021. p. 66–108. [Google Scholar]
- 25. Broda A, LaPlante DA, Nelson SE, LaBrie RA, Bosworth LB, Shaffer HJ. Virtual harm reduction efforts for internet gambling: effects of deposit limits on actual internet sports gambling behavior. Harm Reduct J. 2008;5(1):27. 10.1186/1477-7517-5-27 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Heirene RM, Vanichkina DP, Gainsbury SM. Patterns and correlates of consumer protection tool use by Australian online gambling customers. Psychol Addict Behav. 2021;35(8):974–984. 10.1037/adb0000761 [DOI] [PubMed] [Google Scholar]
- 27. Hing N, Browne M, Tulloch C, Russell AMT, Rockloff M. Self‐regulatory strategies reduce gambling spend and harm in a randomised controlled trial of electronic gaming machine players. Int J Mental Health Addiction. 2024:1–25. 10.1007/s11469-024-01308-x [DOI] [Google Scholar]
- 28. Roickloff M, Browne M, Russell AMT, Newall P, Hing N, Armstrong T. Testing the effectiveness of different safer gambling messages for sports and race betting: a five‐week experiment. Addict Behav. 149:107893. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All materials and data from the present research are publicly available via the Open Science Framework: https://osf.io/6qkyh.