China is on track to be the world’s largest lottery gambling market who has more than 300,000 lottery betting shops and more than 750,000 employees. Since its issuance in 1987, it has sold 3443.5 billion yuan (Chinese Academy of Social Sciences, 2022; General Administration of Sport of China, 2022). There are huge hidden worries behind the widespread expansion of lottery gambling in China, especially the growing number of problem lottery gamblers. Problem lottery gambling refers to lottery gambling behaviors that have negative impacts on individuals and related others (such as family or friends) (Ferris & Wynne, 2001). A previous study found that approximately 2% of Chinese lottery players were problem gamblers in 2009 (Li et al., 2011). This proportion had increased to 5% by 2013 (Wang et al., 2013), and it tends to continue to grow (Xu et al., 2018). Studies have verified that problem lottery gambling can induce negative emotions (Rizeanu, 2013), increase risk behaviors (Kuoppamäki et al., 2014), damage family harmony (Dowling et al., 2016a, 2016b), and destroy peer relationships (Rychtarik & McGillicuddy, 2006). Therefore, exploring factors associated with problem lottery gambling is helpful to better understand lottery gambling behaviors and provide useful information for interventions.

Pathways model of problem gambling (Blaszczynski & Nower, 2002) was one of the most popular theories in gambling research that was cited (Clark & Wohl, 2022). It puts forward three subtypes of people with problematic gambling and erroneous beliefs about gambling are a feature found across the subtypes (Blaszczynski & Nower, 2002). Among these erroneous gambling beliefs, the illusion of control (IOC) may play a more important role in the problem lottery gambling. Compared with other gambling activities such as casino gambling, horse racing, and slot machines, lottery gambling is characterized by a low chance of winning a very large jackpot prize (Brochado et al. 2018). People are likely to overestimate their control under heavily chance-determined situations of lottery gambling (Clark & Wohl, 2022). Moreover, compared with Western individualistic cultures, people from Chinese collectivistic cultures have stronger Feng Shui beliefs (Lam, 2014) and beliefs related to fate, luck, risk, and control (Papineau, 2005). Chinese lottery gamblers were more likely to believe that they have number-picking skills that can affect the reward (Zeng & Zhai, 2009), which showed stronger IOC (Un & Lam, 2016). Therefore, a unique contribution of this study is to gain a better understanding of the association between IOC and problem lottery gambling and the mechanisms underlying them within the context of Chinese culture. Additionally, Tabri et al. (2023) stated that erroneous beliefs might not be the only core psychopathology at play in the development of problem gambling. This study tested other factors that help to maintain the problem gambling and extended the pathways model.

IOC refers to beliefs that one believes behavior can increase the chance of winning (Cowley et al., 2015). The pathways model of problem gambling notes that IOC plays a crucial role in the progression and maintenance of problem gambling (Blaszczynski & Nower, 2002). Gamblers with a stronger IOC tend to believe that their behaviors can influence the outcomes of winning (Zeng & Zhai, 2009; Lian et al., 2020). They may pay more attention to what they are doing which leads to “winning” outcomes and ignore the “losing” outcomes (Goodie & Fortune, 2013). While constantly confirming this predictive connection, they gradually become addicted and lose control of their gambling behavior (Un & Lam, 2016). The relationship between IOC and problem lottery gambling has been found among Chinese lottery gamblers (Hu et al., 2018; Xu et al., 2018). Therefore, we proposed the following hypothesis:

  • Hypothesis 1. IOC is positively correlated with problem lottery gambling.

Metacognitions are beliefs about one’s thoughts that involve cognitive-affective experiences and ways of controlling these experiences (Rogier et al., 2021). Metacognitions about lottery gambling refer to the beliefs about lottery gambling-related cognitions and ways of controlling gambling behaviors (Caselli et al., 2018). It can be divided into two categories: positive and negative metacognitions. Positive metacognitions (PM) regard the beliefs about the benefits of lottery gambling. Negative metacognitions (NM) regard the beliefs about the uncontrollability of lottery gambling (Caselli et al., 2018; Zhou et al., 2022). Previous studies have shown that PM can motivate individuals to engage in addictive behaviors, while NM contributes to the maintenance of problem behaviors (Caselli et al., 2018; Spada et al., 2015). Therefore, lottery gamblers with strong metacognitions tend to invest more time and money into gambling, resulting in problem lottery gambling. Studies have found that PM and NM are positively correlated with problem gambling (Mansueto et al., 2016; Spada & Roarty, 2015; Rogier et al., 2021; Zhou et al., 2022).

Additionally, lottery gamblers with strong IOC tend to believe that they hold more gambling-related skills and knowledge (Myrseth et al., 2010). They may then experience a strong sense of control and positive emotions after winning. This makes them overestimate the benefits of lottery gambling (Hu et al., 2018), thereby promoting PM. They are also more likely to analyze the rules of winning patterns and spend more time seeking trends in number picking after losing (Hu et al., 2018). Then, they may find it difficult to control their gambling behavior in the process of continuously verifying their prediction, which leads to NM. To sum up, we proposed the following hypothesis:

  • Hypothesis 2. Metacognitions mediate the link between IOC and problem lottery gambling.

Urges are a state of being motivated to search for specific objects or feelings (Rock & Kambouropoulos, 2008). Lottery gambling urges can be defined as current temptations to gamble felt by lottery gamblers (Raylu & Oei, 2004a; Rockloff et al. 2024). According to the conceptual model of gambling behavior (King & Delfabbro, 2016), IOC can promote hazardous patterns of gambling towards automation and then make it difficult for gamblers to control their urge for gambling. Moreover, urges play a pivotal role in maintaining lottery gambling problems. Lottery gamblers with a strong desire to play lotteries are generally more likely to be attracted to lottery-related information (Ciccarelli et al., 2022). They may have poor emotional regulation ability and less effective coping strategies (Michalczuk et al., 2011; Williams et al., 2012; Moccia et al, 2017). They are also less likely to believe in their ability to refuse playing lotteries (Iliceto et al., 2021). Previous research has shown that gambling urge is positively correlated with problem gambling (Ashrafioun et al., 2012; Quilty et al., 2017; Iliceto et al., 2021). To sum up, we proposed the following hypothesis:

  • Hypothesis 3. Urges mediate the link between IOC and problem lottery gambling.

The self-regulatory executive model suggests that metacognition about lottery gambling is associated with maladaptive coping strategies that lead to the perseveration of negative emotions (Wells, 2002). Previous studies have shown that metacognitions are positively correlated with anxiety and depression (Hjemdal et al., 2013; Jauregui et al., 2016), which can increase the urge to gamble (Quilty et al., 2017). Additionally, maladaptive coping strategies play a key role in developing negative interpretations of thoughts and increase the perception of being unable to control them (Hamonniere & Varescon, 2018). This perceived inability to control gambling is positively associated with gambling urges (Casey et al., 2008). Taken together, we proposed the following hypothesis:

  • Hypothesis 4. Metacognition and urges sequentially mediate the effect of IOC on problem lottery gambling.

To address some of the gaps in the literature, the present study aimed to shed light on the psychological mechanisms between IOC and problem lottery gambling. Specifically, we need to answer the following questions: (a) whether IOC was related to problem lottery gambling; (b) whether metacognitions mediated the association between IOC and problem lottery gambling; (c) whether urges mediated the association between IOC and problem lottery gambling; (d) whether IOC can influence problem lottery gambling through the serial mediating roles of metacognitions and urges.

Method

Participants and Procedure

A total of 402 lottery gamblers participated in this study. The age range of the participants was 18–83 years (Mage = 44.50 years, SD = 14.15). Seventy-seven percent of the participants were males, and most had a high school diploma (30%). On average, they have bought lottery tickets for 11.56 years (SD = 8.30). Participants were recruited from the lottery stores in Tianjin and Wuhan, China. They were invited to complete questionnaires in nearly 30 min in the lottery stores under the supervision of our research assistant. The authenticity of all answers was emphasized to all participants by well-trained research assistants. This study was approved by the Research Ethics Committee of Tianjin University of Sport (2023-014). Informed consent was obtained from all individual participants included in the study. All participants informed consent for study participation and publication of results.

Measures

Illusion of Control

We used the illusion of control subscale in the Gambling Cognition Questionnaire developed by Raylu and Oei (2004b) to measure IOC. The questionnaire contained four questions, such as “I always collected some items that can increase my chances of winning.” The questionnaire was rated on a five-point Likert-type scale, with higher scores indicating a stronger IOC. The Cronbach’s α coefficient was 0.89 in this study.

Problem Lottery Gambling

Problem lottery gambling was measured by the Problem Gambling Questionnaire developed by Ferris and Wynne (Ferris & Wynne, 2001). It included 9 items. Each item is scored from 1 (strongly disagree) to 7 (strongly agree). The Cronbach’s α coefficient was 0.90 in this study.

Lottery Gambling Urge

The gambling urge questionnaire developed by Raylu and Oei (2004a) was used to measure the degree of lottery gambling urge for lottery gamblers. This scale contained 5 items and adopted the five-point scoring method, from 1 (strongly disagree) to 7 (strongly agree). The Cronbach’s α coefficient was 0.92 in this study.

Metacognitions About Lottery Gambling

The Gambling Metacognition Questionnaire developed by Caselli et al. (2018) was adopted to measure metacognitions about lottery gambling. The questionnaire included two dimensions, PM and NM. Each dimension contains 5 items. A seven-point scale for each item was used, ranging from 1 (strongly disagree) to 7 (strongly agree). The Cronbach’s α coefficient was 0.91 in this study.

Statistical Analyses

The present study conducted statistical analysis using SPSS 22.0 and MPLUS 7.0. Descriptive statistics and Pearson correlation were performed. Structural equation modeling was performed to examine the mediational role of metacognitions and urges in the relationship between IOC and problem lottery gambling. Tucker Lewis index (TLI) and Comparative Fit Index (CFI) served as model fit indices with a reasonable fit for values above 0.90 (Hu & Bentler, 1999). Statistical stability of the model was also performed based on the bootstrap procedure (1000 bootstrapped samples). Additionally, Harman’s single factor was applied to assess the common method bias (Aguirre-Urreta & Hu, 2019). The results showed that there were seven factors with an eigenvalue greater than one, and the variation explained by the first factor was 13.45%, less than 40%. It indicated that there was no serious common method deviation in this study.

Results

Descriptive Statistics and Correlations

Descriptive statistics and correlation between the variables are shown in Table 1. Problem lottery gambling was positively associated with IOC, PM, NM, and urges. Furthermore, IOC was positively associated with PM, NM, and urges.

Table 1 Descriptive statistics and correlations among study variables
Full size table

Serial Mediation Analysis

The model fit indices were within the acceptable range: χ2 = 1030.71; χ2/df = 3.03; RMSEA = 0.07; TLI = 0.90; CFI = 0.91. Figure 1 and Table 2 summarize the standardized path coefficients of the serial-mediated structural model. It was shown that IOC was positively associated with problem lottery gambling (β = 0.193; 95% CI 0.046, 0.243). Results also reported the significantly indirect effect of NM (β = 0.092; 95% CI 0.05, 0.154) and urge (β = 0.030; 95% CI 0.009, 0.070) in the relationship between IOC and problem lottery gambling. The indirect effect of PM was not significant. The indirect role of PM and urge (β = 0.014; 95% CI 0.005, 0.033) and the indirect role of NM and urge (β = 0.035; 95% CI 0.015, 0.066) was also significant. Moreover, the difference between these two coefficients was not significant (Δβ = −0.013; 95% CI −0.032, −0.003). Additionally, the total model explained 31.4% of the total variance.

Fig. 1
figure 1

Serial mediating effect model of metacognitions and urge between the illusion of control and problem lottery gambling

Full size image
Table 2 Direct, indirect, and total effects for the final model
Full size table

Discussion

As the only legalized gambling activity in the Chinese mainland, lottery gambling is rapidly gaining popularity and has become one of the largest gambling industries in the world. Previous researchers have shown the association between IOC and problem gambling, but the mediating mechanism remains unclear, particularly for Chinese lottery gamblers. The current study questioned how IOC, PM, NM, and urge affected problem lottery gambling among Chinese lottery gamblers. First, we reported a positive relationship among problem lottery gambling, IOC, PM, NM, and urge. Second, structural equation modeling analysis revealed that urge and NM could positively mediate the association between IOC and problem lottery gambling. Finally, the association between IOC and problem lottery gambling was also partially mediated by the serial mediating role of both NM and urge and the serial mediating role of both PM and urge.

Firstly, results reported that IOC played a significant and positive role in the increase of problem lottery gambling, and this role remained when PM, NM, and urge were controlled. This finding agrees with our hypothesis 1. This also supports the pathways model of problem gambling, which notes that cognitive factors such as IOC could promote gamblers constantly chasing “losing” and “winning,” and their losses gradually exceeding their expectations which induces gambling problems (Blaszczynski and Nower, 2002). Studies in China have found a positive association between IOC and problem lottery gambling (Hu et al., 2018; Xu et al., 2018). Similar patterns are visible in Social Affairs (Orlowski et al., 2020), France (Moreau et al., 2020) and Italy (Nigro et al., 2022). For example, a study in Social Affairs reported that higher IOC was associated with problematic gambling (Orlowski et al., 2020).

Partly agree with hypothesis 2, the relationship between IOC and problem lottery gambling was only partially mediated by NM. This finding implies that lottery gamblers with strong IOC could reduce their confidence in controlling their lottery gambling behaviors and impair their self-regulation ability. Then, lottery gamblers were more likely to be addicted to gambling. Moreover, this result indicated that metacognitions might be a more proximal cognitive risk factor that influenced the occurrence and development of problem lottery gambling, compared to the illusion of control. Similar patterns have been found between NM and problematic networks (Zhou et al., 2021), game addiction (Dang et al., 2022), and excessive drinking (Janssen et al., 2020). Additionally, the mediating role of PM was not significant, suggesting that NM may have a stronger effect on problem behavior than PM. The results do align with previous studies. They found that PM reduced the perception of gambling’s adverse consequences and affected the choice of gambling activities (Caselli & Spada, 2015; Caselli et al., 2018). NM was reported to influence the maintenance of gambling activities and self-control of gambling behaviors (Caselli & Spada, 2015; Caselli et al., 2018).

The third hypothesis, regarding the mediation role of urges between IOC and problem lottery gambling, was also accepted. Previous studies have found that gambling urges are positively related to IOC (Raylu & Oei, 2004a) and problem gambling (Quilty et al., 2017; Iliceto et al., 2021). In addition, past researchers applied psychological interventions to reduce gambling urges. They conclude that cognitive reappraisal focused on promoting gamblers to think about the adverse consequences of problem gambling and mindfulness training can improve risk perception, self-control, and coping ability (Dowling et al., 2016a, 2016b; Mallorquí-Bagué et al., 2019; Wu et al., 2020). Practically, policymakers and researchers can also develop some easily popularizing, simple, and convenient self-intervention tools or internet-based psychological interventions for lottery gamblers, for example, audio and video from mindfulness interventions, warning messages, and educational animation videos of how lottery work.

Finally, our study found that the IOC could affect problem lottery gambling via the serial mediation role of metacognitions and urges, which supports H4. This suggests that IOC may promote the development of PM and NM and then trigger strong urges, which eventually lead to lottery gambling problems. Moreover, the single mediating role of PM was not significant, which indicates that the mediating effect of PM between IOC and problem lottery gambling can only be achieved through urges. More importantly, the serial mediation effect of NM and urges was significantly stronger than PM and urges. These results indicate that NM may play a more prominent role in maintaining problem lottery gambling for gamblers with strong IOC. Indeed, previous researchers suggested that PM and NM might work at different stages of gambling engagement (Caselli et al., 2018; Zhou et al., 2022). Specially, PM mainly works in the early stage as a coping strategy to motivate individuals to engage in gambling activities, while NM works in the late stage and tends to strengthen gamblers’ negative emotions and reduce their control over gambling (Caselli et al., 2018; Zhou et al., 2022). It is reasonable to assume that lottery gamblers with strong IOC tend to develop PM and continue to gamble without considering the potential negative consequences. With the accumulation of gambling losses, these gamblers gradually experience negative emotions and thoughts and develop NM. This in turn leads to strong gambling urges and a lack of control to cease gambling, thus increasing the vulnerability to lottery gambling problems. Additionally, although the influence of IOC on problem lottery gambling was partly explained by metacognitions and urges, the magnitude of the indirect effect was modest. That is to say, it appears that the IOC may not solely exert its influence on problem lottery gambling via metacognition and urges.

There are several limitations in this study. First, a cross-sectional research design conducted in this study could only obtain simultaneous relationships, inviting future work to focus on longitudinal studies to infer causal conclusions. Second, participants selected in this study were mainly from lottery stores, and the data collection was in the period of non-sport events. Thus, there were fewer female participants included in our study. Future research should use other methods such as WeChat and QQ to collect more data from female lottery players. Third, the data were self-reported which may be affected by social expectations and individuals’ recall bias (Lopez-Gonzalez et al., 2020). Ecological momentary assessment technology (Hawker et al., 2021) can be used to overcome this issue in the future. Finally, the age range (18–83) of the present study contained different age generations. Although this aligns with previous studies in the gambling field (Nigro et al., 2022; Lian et al., 2020), the results may differ in various age groups. Previous studies have shown that young people were more vulnerable to gambling problems because of cognitive immaturity (e.g., illusions of control; Chambers & Potenza, 2003). In contrast, most elderly individuals focus on maintaining the enjoyment of gambling (Fontaine et al., 2023a) and show less illusion of control (Fontaine et al., 2023b). These findings invite future work to explore the age differences in the mediating model of the present study.

Conclusion

Problem lottery gambling is a growing public health problem in China (Bin et al., 2021). IOC plays a key role in the problem lottery gambling. However, little is known about the underlying mechanisms between IOC and problem lottery gambling. The present study provides empirical evidence for the structure relationships among IOC, PM, NM, urges, and problem lottery gambling. Specifically, we found the direct effect of a lower illusion of control on higher problem lottery gambling. Additionally, there was a marginally significant indirect effect from the illusion of control to problem lottery gambling via higher negative metacognitions, negative metacognitions and urges, and positive metacognitions and urges. This study extended and deepened the gambling path model with metacognitions and urges to enhance the understanding of Chinese lottery gambling behavior.