An addiction is an ongoing failure to resist an impulse or urge to engage in a certain response, despite experiencing repeated harm by such engagement (American Society of Addiction Medicine, 2019 ; Grant et al., 2010). Addictions can involve engagement with substances, such as alcohol, or with behaviors, such as gambling (Kim & Hodgins, 2018 ). Personality refers to characteristic ways of thinking, feeling and acting across time and place. Researchers have argued that establishing links between psychological disorders and personality dimensions are important as they can enhance our understanding of diatheses, cause, progression, prognosis, and the treatment of psychological disorders (Costa & Widiger, 1994 ; Markon et al., 2005 ; Watson et al., 1994 ). In recent years, the findings from numerous studies have shown that the different types of addictions are associated with different types or groups of personality dimensions (Grant et al., 2010 ). However, we will argue in this introduction that existing findings in this area have serious omissions and limitations. Given this, the aim of the current study was to examine the associations of ten different types of common addictions (alcohol, smoking, drug, sex, social media, shopping, exercise, gambling, internet gaming, and internet use addictions) with the personality dimensions in the Five-Factor Model (FFM; McCrae & Costa, 1985 ; McCrae & Costa, 1999 ), controlling for the omissions and limitations in existing studies.
The Big-5 Model of Personality
The Five-Factor Model is at times referred to as the Big-5 personality model (Big-5; Goldberg, 1992 ). In this model, the major dimensions are extraversion (E; individual differences in reactivity to positive environmental stimuli); neuroticism (N; individual differences in reactivity to negative environmental stimuli) or its opposite pole called emotional stability (ES); agreeableness (A; reflecting how a person is generally with others and interacts with others); conscientiousness (C; reflecting how organized, responsible and task-focused an individual is in pursuing goals); and openness to experience (O; relates to being ‘open-minded’) (McCrae & Costa, 1999 ). The Big-5 personality dimensions have been widely used in addiction research.
Major Findings of the Associations of Different Types of Addictions with Big-5 Personality Dimensions
Because 10 different addiction types are examined in the study, and as there have been many studies in this area, a complete and comprehensive review of the associations for these different addition types with the Big-5 personality dimensions is beyond the scope of this paper. Consequently, to place our study in context, we will summarize the major findings in this area using the findings from meta-analysis, where available. When such findings are not available, the findings from systematic reviews, when available, will be summarized. Where deemed necessary, the findings from both meta-analysis and systematic reviews will be supplemented with more recent findings not included in the original reviews. For addictions without meta-analysis or systematic reviews, we will attempt to synthesize and summarize recent individual studies.
For alcohol addiction, the meta-analysis by Malouff et al. (2007 ) that included 20 studies concluded that alcohol addiction is associated negatively with ES, C, and A. A subsequent meta-analysis by Kotov et al. (2010 ), based on 11 (for C and A) to 26 (for ES) studies concluded that it was associated with only ES and C. A recent study not included in these meta-analyses reported negative associations with ES, C, and A (Dash et al., 2019 ). In relation to drug addiction, the meta-analysis by Kotov et al. (2010 ; 12 studies for E and ES, and 4 studies for C, A, and O) concluded that it was associated negatively with E, ES, C, and A. A study by Zilberman et al. (2018 ) that was not included in the Kotov et al. (2010 ) study reported that drug addiction was associated with A, C, and ES. In relation to smoking addiction, Smith (1970 ) concluded a positive association between smoking and E, and a subsequent review by Gilbert (1995 ) concluded that there was an association with high psychoticism (psychoticism is a personality construct associated negatively with C and A) and low ES. In a review of 25 studies that only focused on the E and ES personality dimensions, Malouff et al. (2007 ) reported that smokers had higher E and lower ES scores than non-smokers. More recently Hakulinen et al. (2015 ), covered nine studies finding that current smoking was associated positively with E, and negatively with ES and C. Additionally a study by Dash et al. (2019 ), which was not included in these reviews, concluded that smoking addiction was associated negatively with A, ES, and C. For internet addiction, a meta-analysis involving 12 studies, by Kayiş et al. (2016 ) reported that it had negative associations with all Big-5 personality dimensions. As for gambling, the meta-analysis by MacLaren et al. (2011 ), involving 44 studies, found that it was associated negatively with ES, C, and A. This study did not examine the relationship with O. For internet gaming, a systematic review that included a wider range of personality dimensions than the Big-5 dimensions concluded that for internet gaming there was a negative association or no relationship with E, ES, and O; and mixed results with C and A (Şalvarlı & Griffiths, 2019 ). Another systematic review by Gervasi et al. (2017 ) covering 27 studies concluded that internet gaming is associated negatively with ES, A and C. In relation to exercise addiction, studies have shown that exercise is associated positively with E and negatively with A (Costa & Oliva, 2012 ; Hausenblas & Giacobbi, 2004 ; Mathers & Walker, 1999 ). Though findings for ES shown some inconsistency, with some studies reporting a negative association (Courneya & Hellsten, 1998 ; Yeung & Hemsley, 1997 ), and others reporting positive associations (Costa & Oliva, 2012 ; Hausenblas & Giacobbi, 2004 ). However, in a systematic review that included 5 studies, Bircher et al. (2017 ) concluded that there was no clear association between exercise addiction with any Big-5 personality dimensions. As for any associations involving sex, social media use, and shopping addictions with the Big-5 dimensions, we did not find reviews to cover. For sex addiction, at least two studies have reported negative associations with C and ES (Amamou et al., 2020 ; Zilberman et al., 2018 ). Additionally, Amamou et al. (2020 ) also reported a negative association with E, and Zilberman et al. (2018 ) reported a negative association with A. In relation to shopping addiction, findings have generally reported a positive association with E (Andreassen et al., 2013 ; Uzarska et al., 2021 ; Verplanken & Herabadi, 2001 ), and a negative association with ES (Andreassen et al., 2013 ; Mowen & Spears, 1999 ), C (e.g., Andreassen et al., 2013 ; Mowen & Spears, 1999 ; Verplanken & Herabadi, 2001 ; Wang & Yang, 2008 ), A (e.g., Andreassen et al., 2013 ; Uzarska et al., 2021 ) and O (Andreassen et al., 2013 ). In relation to (general) social network use addiction, the findings have generally been inconsistent. Studies have reported positive associations with E and negative associations with ES (Peris et al., 2020 ; Wang et al., 2015 ). Sumaryanti et al. (2020 ) found a positive association with A and E, and a negative association with ES, C, and O. Kavčič et al. (2019 ) reported negative associations with ES. Wilson, Fornasier and White (2010 ) have reported a positive association with E, and a negative association with C. Negative associations with C has also been reported by others (e.g., Andreassen et al., 2012 ).
In summary, the findings suggest that alcohol addiction is likely to be associated with low ES and C, and possibly low A. For drug addiction, it is highly probable that it is associated with low A, C, and ES, and possibly low E. Although there is notable inconsistency in the findings, it can be speculated that smoking addiction will be associated with low A, C, and ES. All Big-5 personality dimensions may be associated negatively with internet addiction; and gambling addiction is likely to be associated with low ES, C, and A. Furthermore, internet gaming has been associated with low ES, A, and C. Whilst sex addiction has been associated with low C and ES, and possibly low A and E. Shopping addiction can be expected to be associated with low C, A, ES, and O, and high E. Despite these inconsistencies, the findings for social media use addiction appear to indict associations with high E and low ES and C. Although it is difficult to speculate clearly the associations between exercise addiction and the Big-5 dimensions, exercise addiction is most likely to be associated positively with E and negatively with A. Overall therefore, past findings across most of the different types of addictions appear to indicate fairly consistent associations with low ES, C, and to a lesser extend A (especially for biological addictions), with less consistent associations with high E and low O.
Limitations of Existing Findings on the Associations of Different Types of Addictions with Big-5 Personality Dimensions
Despite the existence of a large number of studies that have examined the associations of different types of addictions with the Big-5 personality dimensions, it is argued here that existing findings from these studies are limited. Firstly, to date, most of the studies in this area have examined only one or at most a small handful of addictions in the same study. Although the study by Villella et al. (2011 ) examined five types of addictions, they were all behavioral addictions (pathological gambling, compulsive buying, internet addiction, workaholism, and exercise addiction). Likewise, although Andreassen et al. (2013 ) examined seven different addictions, they were also all behavioral addictions (Facebook addiction, video game addiction, internet addiction, exercise addiction, mobile phone addiction, compulsive buying, and study addiction). Additionally, both Andreassen et al. (2013 ) and Villella et al. (2011 ) did not include many behavioral addictions that are of current interest, such as gaming, sex, and social media. Given these limitations and omissions, it can be argued that further research is needed, examining concurrently, for the same group of participants, the associations of the Big-5 personality dimensions with a wider range of both biological and behavioral addictions (especially those of the current interest). Such studies will reveal shared and unique associations across different addictions and will not be confounded by differences in the sample and study characteristics (Andreassen et al., 2013 ). As noted by others, although these cross-study comparisons of personality correlates of the different addictions are informative, the most incisive approach is to make these comparisons within the same study (Krueger et al., 1996 ).
Second, the vast majority of the past studies in this area, including the meta-analysis, systematic reviews, and individual studies cited above have examined bivariate correlations between the different pairs of addictions and the Big-5 personality dimensions. Given this, and as there are many shared variances across the Big-5 personality dimensions (Costa and McCrae, 1992 ), the findings from the correlation analyses, therefore, do not indicate the unique associations between the personality and addiction variables. There is however some limited data on unique associations. Although our review of past studies suggested that shopping addiction can be expected to be associated with low C, A, ES, and O, and high E, and that exercise addiction will be associated positively with E and negatively with A, and internet addictions will be associated negatively with all five personality dimensions, Andreassen et al. (2013 ) have reported that shopping, exercise, and internet addictions were all uniquely associated negatively with ES, A, and C, and positively with E. While our review suggested that gambling addiction is likely to be associated with low ES, C, and A, Andreassen et al. (2013 ) have reported that it is uniquely associated with low C. While our review suggested that gambling will be associated with low ES, C and A, Miller et al. (2013 ) found that only low ES and O predicted gambling problems uniquely. As will be noticed, when examined for unique associations, the findings are different than when examined for bivariate associations. As unique associations provide a more accurate understanding of the personality-addiction links, we need such data. Statistically, they can be obtained by using multiple regression analysis in which each of the addiction types is regressed on the Big-5 personality dimensions simultaneously.
Third, two important considerations when examining the associations for the different addictions with personality dimensions are gender and age. This is because both age and gender are known to influence different addiction types (e.g., Andreassen et al., 2013 ; Becker et al., 2017 ; Cotto et al., 2010 ; Thege et al., 2015 ). This means that existing correlation and unique association data on how the different addiction types are associated with personality dimensions that have not controlled for gender and age effects (which has more often been the case) are confounded and misleading. The importance of controlling for age and gender effects is emphasized in the study by Jaradat et al. (2017 ). They found that while high E and O, and low ES predicted social media addiction in a multiple regression analysis model, only low C predicted social media addiction when gender was included in the regression model, and none of the personality dimensions predicted social media addiction when age was included in the regression model. The unique associations, controlling for age and gender, can be obtained using multiple regressions, in which each of the addiction types is regressed on the Big-5 personality dimensions, and also age and gender simultaneously.
Aims of the Current Study
Given the omissions and limitations in the existing literature, the current study used multiple regression analysis to examine the associations of 10 different types of addictions with the Big-5 personality dimensions (E, ES, A, C, and O) in the same study. The addictions examined were alcohol, smoking, drug, sex, social media, shopping, exercise, gambling, internet gaming, and internet use. Age and gender were controlled in the multiple regression analyses. As there has been limited data on the unique associations between the different types of addictions with the Big-5 personality dimensions, controlling for age and gender effects, we made no clear predictions. However, based on the literature reviewed for bivariate correlations, it is speculated that many (but not all) of the following associations would hold. Alcohol, drug, and smoking addictions will be associated negatively with A, C, and ES; internet gaming and internet addictions could be associated negatively with all five personality dimensions; pathological gambling will be associated negatively with ES, C, and A; sex addiction will be associated negatively with C and ES; social media use addiction will be associated positively with E, and negatively with ES; shopping addiction will be associated positively with E and negatively with the remaining four personality dimensions. Exercise addiction is most likely to be associated positively with E and negatively with A.