The analysis of socio-demographic characteristics shows that males were more vulnerable than females to develop gambling-related problems. This finding is in line with earlier studies both nationally [5–7, 21] and internationally [22–24]. Additionally, young age was strongly associated with all levels of gambling. However, this association needs to be interpreted with caution, since even though the odds ratio was significant in the model, the strength of association was quite low. Divorced or separated individuals were also found to have more severe gambling problems compared to those married. Low level of education was associated with low and moderate levels of gambling problems and unemployment was associated with more severe gambling problems. These results resemble previous findings concerning socio-demographic characteristics of problem and pathological gamblers [7–11]. A vast body of evidence suggests that overall low socio-economic status is a risk factor for gambling, as also stated by Jimenez-Mucia et al. .
The results from this study show that the prevalence of problem gambling is in line with the findings of Williams, Volberg and Stevens . The prevalence rate in Finland falls in the average rates category among other European countries such as Sweden, Switzerland, Estonia and Italy. The findings of this study are also in line with previous studies from Finland indicating that the trend of problem gambling has been more or less unchanged during the past few years. There appears to be a declining trend based on Finnish Gambling Population Surveys [4–6] but it is not statistically significant according to the recent analysis by Raisamo and Salonen . A unique finding of this study was that the moderate level of gambling was 5.5% in this sample. This is important to acknowledge, since moderate level gamblers may be at risk of developing gambling problems and so suffer from an increasing amount of negative consequences caused by gambling. Thus moderate levels of gambling need to be monitored closely in the future.
Types of gambling
According to our results, Lotto was the most popular type of gambling, which has also been the case with previous population surveys in Finland . Males bet on Lotto more frequently than females. On the other hand, females were more attracted to less risky gambling types such as lotto and scratch card gambling as compared to males also found by Hraba and Lee . Frequent gambling in Lotto, daily lotteries, horse race betting and slot machine gambling were found to be associated with a more severe level of gambling.
In this study, slot machine gambling attracted the respondents gambling on a weekly basis. Slot machine gambling is classified as being addictive by nature  and has been reported to be the most trouble causing type of gambling among Finnish treatment-seeking gamblers . The same trend was reported by Turja et al.  in a population sample where those who scored 5 or higher in SOGS reported their preferred game being slot machines. The rather high level of involvement in slot machine gambling, especially in Finland, could be explained by easy access and abundant availability. In Finland, slot machine gambling is easily and conveniently available in restaurants, grocery stores, shopping centres, kiosks and fuel stations around the country. A recent study in Finland by Warpenius et al.  investigated the enforcement of legal age limits on purchases of alcohol, tobacco and gambling slot machines. Their study showed that the enforcement of legal age limits for gambling slot machines was the weakest (4%) compared to purchases of alcohol (49%) and cigarette (43%). One of the reasons for the insufficient enforcement of the law is, at least in kiosks, shopping centres and fuel stations, that the locations of slot machines are often out of reach of the shop keepers’ desks, gambling time can be rather short and the gambler does not need to confront the shop keeper directly.
Availability, proximity and convenience of gambling venues have been found as being clear risk factors for gambling problems [30–32]. As a whole, males seem to choose riskier and faster game types, such as games like internet poker, which could partially explain why males tend to have more severe levels of gambling problems. Our results show that frequent involvement in internet gambling is associated with more severe levels of gambling problems. Our results also show that female moderate level of gambling was 2.6%, which is nearly three times lower compared to males. Females have been found to gamble often because of boredom, loneliness and isolation and thus maximize their gambling time with slower games [33, 34]. Nevertheless, females have a growing risk in developing gambling problems, because progression in gambling problems is reported to be more rapid with females  and is known as telescoping phenomenon .
Comorbid alcohol and nicotine dependency
In this sample, males consumed remarkably more alcohol than females. Risk-level alcohol consumption and male gender were strongly associated with low and moderate levels of gambling problems. Comorbidity with substance use, especially alcohol, has been frequently reported in earlier gambling studies. The rate of pathological gambling among substance abusers has been reported to be four to ten times greater when compared to the general population [37, 38]. Similarly, the study by Hakkarainen et al.  from Finland stated that especially heavy episodic drinking increased the risk of problem gambling. Petry has reported that the odds ratio for any alcohol use disorder and alcohol dependence as well as drug use and smoking are significantly related to pathological gambling .
In fact, nicotine dependence “accounts for some elevated risks for psychopathology with subsyndromal and problem/pathological levels of gambling” and subsyndromal level of gambling are associated with more severe psychopathology, as stated by Grant, Desai and Potenza . In our study, nicotine dependence was significantly associated with all levels of gambling severity. This is consistent with previous findings [37–39, 41]. Nicotine dependence is the second most frequent addiction after alcohol use disorder. Moreover, Petry, Stinson and Grant  reported that nicotine dependent individuals had a seven times higher odds ratio to be pathological gamblers when compared to non-smokers. Additionally, women with nicotine dependence were 14 times more likely to be pathological gamblers when compared to non-smoking women. In contrast, the likelihood for nicotine dependent males being pathological gamblers was five times higher when compared to non-smokers. Petry and Oncken  found that daily smokers were less able to control their gambling and had more severe gambling problems when compared to non-smokers. The association between gambling problems and nicotine dependence is evident, and one dependence may serve as a prime for another, as suggested by McGrath and Barrett .
Associations of socio-demographic characteristics, alcohol consumption, smoking and gambling severity
The results from this study revealed that younger age, male gender and daily smoking were associated with all levels of gambling problems. The notable likeness between low level and moderate level of gambling problems were less than twelve years of education and risk-level alcohol consumption. In contrast, unemployment was associated with problem gambling. These findings are soundly in line with previous findings and confirms that there appears to be certain groups of individuals that are more vulnerable to developing gambling problems.
Despite the large sample size and good representation of the Finnish population, there are a number of limitations with the data, possibly the self-administered survey being the most important. A self-administered survey is not as candid and honest as a face-to-face interview, and future studies would benefit from face-to-face interviews or correction weights (1.00) as suggested by Williams et al. . Two previous studies have found a post-questionnaire to produce higher rates of prevalence compared to telephone interviews . Another aspect to mention is the response rate of 57%, which is low compared to face-to-face interviews or telephone interviews, but still adequate .
Another limitation is that this study reflects its results with previous prevalence studies from Finland. Previous prevalence studies from Finland [4–6] have used the South Oaks Gambling Screen (SOGS) as a measure of gambling. This study uses Problem Gambling Severity Index (PGSI) which is more conservative than SOGS, because it uses 8 as a cut-off point as compared to SOGS that uses 5. This study’s gambling-related questions were part of a larger health survey. This means that this study’s prevalence rate ought to be more realistic compared to specific gambling surveys, because specific gambling surveys may attract gamblers more than health surveys and produce higher prevalence rates .
This study used PGSI which has not been validated in a Finnish cultural context. So far, there are no validated gambling scales available in Finland.