key: cord-0018176-pqdbsxym authors: Akbari, Mehdi; Bahadori, Mohammad Hossein; Mohammadkhani, Shahram; Kolubinski, Daniel C.; Nikčević, Ana V.; Spada, Marcantonio M. title: A discriminant analysis model of psychosocial predictors of problematic Internet use and cannabis use disorder in university students date: 2021-05-27 journal: Addict Behav Rep DOI: 10.1016/j.abrep.2021.100354 sha: f956725d2d36d97d778f70adc4f1d6d1d0a7f4cc doc_id: 18176 cord_uid: pqdbsxym Researchers have found similarities and differences between behavioral and drug addictions. The present study was designed to explore which of a series of psychosocial predictors of addictive behaviors could differentiate problematic Internet use (PIU) and Cannabis Use Disorder (CUD) in a sample of University students. A total of 144 participants (76 males, mean age = 23.03 years ± 2.83) were separated into three groups: those presenting with PIU (18 females, Mean age = 22.27 years), those presenting with CUD (22 female, Mean age = 22.73 years), and a control group (28 female, Mean age = 24.04 years). Participants completed the Internet Abusive Use Questionnaire (IAUQ), the Severity of Dependence Scale (SDS), the Multidimensional Scale of Perceived Social Support (MSPSS), the Barratt Impulsiveness Scale-11 (BIS-11), the Multidimensional Distress Tolerance Scale (MDTS), the Emotion Regulation Questionnaire (ERQ), the Metacognitions Questionnaire-30 (MCQ-30), and the Repetitive Thinking Questionnaire-10 (RTQ-10). The classification analysis results showed that 68.8% of the control group, 70.8% of the PIU group, and 81.3% of the CUD group were correctly classified in their respective groups. In addition, the results of the discriminant function analysis showed that there was a significant difference between members of the PIU and CUD groups in the degree of family support (0.45), significant other (0.33), tolerance of physical discomfort (0.30), reappraisal (0.42), and cognitive confidence (0.35). The findings provide evidence that specific psychosocial predictors can discriminate PIU from CUD. The phenomenon of non-chemical addictions, or 'behavioral addictions', was first introduced over thirty years ago (Marks, 1990) and the subsequent criteria to define and diagnose these different types of addictive behaviors have been highly controversial ever since (Kardefelt-Winther et al., 2017) . Shaffer and Stimmel (2014) defined three core characteristics of an addictive behavior. The first is the presence of psycho-physiological responses when exposed to the environmental stimulus for that behavior. The second is the engagement in repetitive behavior despite its harmful consequences. The third is wanting to stop the behavior but being unable to do so. Common features of behavioral addictions suggest that these share similar features to drug addictions, but there are also noticeable differences. A key similarity is the emergence of a short-term reward that continues to intensify despite the negative consequences of engaging in the addictive behavior (Thege, Woodin, Hodgins, & Williams, 2015) . Key dissimilarities include the dependence on a substance (Alavi et al., 2012) and the severity of symptoms (Thege et al., 2015) not attenuating as fast in drug addictions. Several studies have also examined the differences in personality or psychopathology between behavioral addictions and drug addictions (Ko et al., 2006; Maremmani et al., 2018; Zilberman, Yadid, Efrati, Neumark, & Rassovsky, 2018) . Ko and colleagues (2006) , for example, showed that novelty seeking, harm avoidance, and reward dependence predicted addictive behaviors generally. Maremmani and colleagues (2018) found higher general psychopathology in people with heroin use disorder than people with gambling disorder. Zilberman and colleagues (2018) reviewed the importance of personality in discriminating the type of addictive behavior. The personality profile of individuals with alcohol use disorder showed lower extraversion and openness to experience, and the personality profile of individuals with gambling disorder showed higher impulsivity and neuroticism (Zilberman et al., 2018) . The similarities and differences between behavioral and drug addictions remain a contentious matter. In the current study, and using a discriminant analysis model, we aimed to determine whether certain psychosocial factors could differentiate problematic Internet use (PIU) and Cannabis Use Disorder (CUD) in University students. We chose PIU and CUD as 'representatives' of behavioral and drug addictions, respectively. The prevalence of PIU is between 9% and 11% (Moreno, Eickhoff, Zhao, Young, & Cox, 2019) and it can cause significant problems in the lives of University students (Costa, Patrão, & Machado, 2019; Meda et al., 2017; Pearson, Hustad, Neighbors, Conner, & Bravo, 2018; Troup, Andrzejewski, Braunwalder, & Torrence, 2016) . The prevalence of CUD is 4.4% among those aged 18-29 years, according to the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5; American Psychiatric Association, 2013) , making it the most prevalent illicit/non-medical drug disorder in University students. In a study conducted in Iran on 4261 university students, the prevalence of PIU was estimated at 27.3% (Poorolajal et al., 2019) . Furthermore, many studies have shown the relationship between insomnia (Shadzi, Salehi, & Vardanjani, 2020) , impulse control disorders (Mazhari, 2012) , poor general health (Poorolajal et al., 2019) , anxiety (Hashemi et al., 2020) , and PIU among the Iranian population. In a meta-analysis, the prevalence of CUD among Iranian college students was 2%-3% (Nazarzadeh et al., 2015) . Also, a negative correlation has been observed between Iranian college students' level of quality of life, sleep quality (Kushkestani, Parvani, Ebrahimpour Nosrani, & Bathaeezadeh, 2020) , and CUD, and a positive correlation with depression, anxiety (Noori et al., 2015) , and CUD. What follows is a brief review of research data on both PIU and CUD and several key psychosocial factors (social support, impulsiveness, distress tolerance, emotion regulation, metacognitions, and repetitive negative thinking) which have been found to predict these addictive behaviors. Internet users have grown by 1,266% to 5 billion people over the last 20 years Internet World Stats (2020). However, with all its undeniable uses and benefits, the Internet can have many destructive psychological and social effects for its users (Elhai, Dvorak, Levine, & Hall, 2017) . such as PIU (Davis, Flett, & Besser, 2002; Spada, 2014) , Internet dependency (Tewildt, 2011 ), pathological Internet use (Morahan-Martin & Schumacher, 2000 , and compulsive computer use (Black, Belsare, & Schlosser, 1999) effectively describe the same phenomenon (Mihajlov & Vejmelka, 2017) : excessive and possibly compulsive use of the Internet along with adverse consequences. Yau, Crowley, Mayes, and Potenza (2012) reviewed the clinical and biological characteristics of PIU and suggested that that it could be considered a behavioral addiction. Cannabis is the most abused illicit drug (Centre, 2020) and about one out of every ten people who have used cannabis are likely to develop CUD (Copeland & Swift, 2009 ). Long-term marijuana use can cause memory impairment (Hall & Lynskey, 2016) and can significantly negatively impact on physical health (Feeney & Kampman, 2016) . Furthermore, regular cannabis use is associated with structural and functional changes in the brain (Nader & Sanchez, 2018) . A systematic review by Volkow et al. (2016) reported that cannabis use can affect people's behavior, causes learning and memory impairments, decreases motivation, and increases the risk for psychosis. In general, it can be said that frequent cannabis consumption can harm mental and physical health in various ways (Karila et al., 2014) . Moreover, CUD has associations with anxiety and mood disorders, antisocial personality disorder, and conduct disorder (Agosti, Nunes, & Levin, 2002; Schoeler et al., 2018) . Social support refers to the resources provided to people by others (Cohen & Syme, 1985) and protects individuals from stress and negative emotions (Wood, Maltby, Gillett, Linley, & Joseph, 2008) . Studies showed that when social resources are relatively scarce, some people choose the internet or cannabis use as a medium through which to meet their needs and gain social support, especially when they are required to reestablish social networks because of environmental changes (Carter et al., 2016; Yao & Zhong, 2014; Zhang et al., 2018) . Perceived lack of social support has been found to be indirectly associated with PIU (Prievara, Piko, & Luszczynska, 2019) and, additionally, PIU has been found to be less common in those who perceive more social support (Gunuc & Dogan, 2013) . Lack of social support has been found to increase the likelihood of CUD (2013b; van der Pol et al., 2013a) . Cougle, McDermott, Hakes, and Joyner (2020) , for example, highlight that cannabis dependence is associated with higher rates of personality disorders and lower social support. Patients who used cannabis more often also showed an absence of social support (Dorard, Bungener, Corcos, & Berthoz, 2013) . Evenden (1999) defined impulsiveness as the behavior that is performed with little or insufficient tact which often leads to adverse consequences. Impulsiveness is a predictor of PIU (Mottram & Fleming, 2009 ) and individuals with PIU, exhibited more trait impulsivity than control group (Choi et al., 2014) . People with higher impulsiveness have higher levels of severity of Internet addiction symptoms (Evren, Dalbudak, Evren, & Ozen, 2019) . A recent systematic review demonstrated a positive association between impulsiveness and Internet Gaming Disorder, which is one of the facets of PIU (Ş alvarlı & Griffiths, 2019). Impulsiveness is also an important risk factor for CUD (Brewer & Potenza, 2008; Day, Metrik, Spillane, & Kahler, 2013; DeVito et al., 2020; VanderVeen, Hershberger, & Cyders, 2016; Verdejo-García, Lawrence, & Clark, 2008) . Higher impulsiveness, with greater access to marijuana, for example, is related to CUD (Haas, Zamboanga, Bersamin, & Hyke, 2018) . Distress tolerance refers to a person's ability to tolerate negative emotions and unpleasant situations (Zvolensky, Vujanovic, Bernstein, & Leyro, 2010) . Distress tolerance plays an important role in PIU. Akbari (2017b) showed that distress tolerance is a significant mediator of the relationship between emotional dysregulation and PIU. Distress tolerance is also associated with different aspects of cannabis use, such as withdrawal severity and coping-oriented use (Manning et al., 2018) . There is evidence to suggest that cannabis use can cause more problems in people with low distress tolerance (Bujarski, Norberg, & Copeland, 2012) . Buckner, Walukevich Dienst, and Zvolensky (2019) showed that low distress tolerance is associated with an increase in cannabis cravings. Emotion regulation allows for the ability to moderate emotional experiences and to create appropriate responses to stressful and negative events (Gross & Muñoz, 1995) . The development of emotion regulation skills has been found to be key in preventing Internet addiction (Karaer & Akdemir, 2019) . Difficulties in emotion regulation have also been observed to be associated with PIU (Estévez, Jáuregui, Sánchez-Marcos, López-González, & Griffiths, 2017; Evren, Evren, Dalbudak, Topçu, & Kutlu, 2018; Spada & Marino, 2017) . Difficulties in emotion regulation have also been connected to cannabis use (Allen, 2019) . In several studies, emotion regulation plays a mediating role between marijuana use and other psychological phenomena. For example, it appears to mediate the relationship between post-traumatic stress and coping approaches related to marijuana use (Bonn-Miller, Vujanovic, Boden, & Gross, 2011). Metacognitions refer to higher order thinking that involves active control over one's cognitive processes and establishes the pattern of responses to one's thoughts (Wells & Matthews, 1996) . Generally, metacognitions play a key role in choosing adaptive or maladaptive coping strategies when faced with unpleasant emotions (Spada, Nikčević, Moneta, & Wells, 2008) . A recent systematic review has showed the central role that metacognitions play in the development and maintenance of both PIU (and other behavioral addictions), as well as drug addictions (e.g., alcohol use and nicotine use; Spada, Caselli, Nikčević, & Wells, 2015; Hamonniere & Varescon, 2018) including cannabis use (Brosnan, Kolubinski, & Spada, 2020) . Repetitive negative thinking (RNT) refers to the style of thinking about one's problems or negative experiences that is repetitive, intrusive, and difficult from which to shift attention (Ehring et al., 2011) . Research has shown that RNT levels are elevated across behavioral and drug addictions (Bravo, Sotelo, Pilatti, Mezquita, & Read, 2019; Li & Wang, 2013) . Rumination, a perseverative style of thinking that focuses on the problem without any attempt to problem-solve, is also one of the instances of RNT that is associated with both PIU (Bagatarhan & Siyez, 2020) and CUD (Memedovic et al., 2019) . A discriminant analysis model based on the above psychosocial factors was devised to differentiate between PIU and CUD predictors in a sample of university students. By separating the factors that predict PIU and CUD, not only will it be easier to identify vulnerable people in PIU and CUD presentations, but also more specialized prevention programs could be developed. We predicted that there would be a difference on all measures between the PIU and CUD groups on the one hand and the control group on other, with higher scores in the PIU and CUD groups. The study was conducted on a convenience sample of 144 students, aged 18-29, from three Universities in Tehran, Iran (Amirkabir, Kharazmi, and Tehran). There were 18 women and 30 men (M = 22.27 years, SD = 2.83) in the PIU group, 22 women and 26 men (M = 22.73 years, SD = 2.54) in the CUD group, and 28 women and 20 men (M = 24.04 years, SD = 2.40) in the control group. In terms of education in the PIU group, 29 were undergraduate students and 19 were graduate students, in the CUD group 34 were undergraduate students and 14 were graduate students, and in the control group, 26 were undergraduate students and 22 were graduate students. A face-to-face clinical interview i.e., the Structured Clinical Interview for DSM-5 (SCID-5, First, 2014) was conducted with students who formed the PIU, CUD and control groups. All questionnaires administered in the study were the Persian forms of the original scales. DSM-5 (SCID-5; First, 2014) The SCID-5, a structured clinical interview, was used to diagnose CUD. The validity and reliability of the original version have been proven in several studies and the calculated kappa value showed excellent agreement for substance use disorders (Osório et al., 2019; Shankman et al., 2018) . Furthermore, the validity and reliability study of Persian version of the SCID-5 was established in an Iranian sample and good internal consistency (0.95-0.99), test-retest reliability (0.60-0.79), and Kappa reliability (0.57-0.72) have been reported (Mohammadkhani, Forouzan, Hooshyari, & Abasi, 2020) . Gossop et al., 1995) The SDS is a 5-item self-report measure that assesses the severity of substance dependence that has been used in this study to measure to measure the severity level of CUD. Responses are based on the following 4-point Likert-style scoring: 0 (never/almost never) to 3 (always/nearly always) for items 1-4 and 0 (not difficult) to 3 (impossible) for item 5. Higher scores indicate higher levels of cannabis dependence. There was acceptable reliability of the SDS total score in cannabis users (Cronbach's alpha = 0.70) and at the optimal differentiating cut-off (cutoff point = 3) sensitivity (61.3%) was acceptable (2013b; van der Pol et al., 2013a) . Additionally, the validity and reliability study of Persian version of the SDS was established in an Iranian sample and acceptable internal consistency (Cronbach's alpha = 0.64) was demonstrated (Habibi & Alahdadi, 2018) . The IAUQ is a 12-item self-report measure that evaluates level of abusive use of internet. Responses are based on the following 5-point Likert-style scoring: 0 (totally disagree) to 4 (totally agree). Higher scores indicate higher levels of abusive internet use. The IAUQ shows good structural validity and discriminative capacity (Calvo-Francés, 2016) . The cut-off point for the IAUQ was 24 (Calvo-Francés, 2016). The validity and reliability study of IAUQ was established in an Iranian sample and strong internal consistency (Cronbach's alpha = 0.91) and adequate convergent reliability (r = 0.82) and test-retest reliability were demonstrated (Mottaghi & Safaie, 2017) . Zimet, Dahlem, Zimet, & Farley, 1988) The MSPSS is a 12-item self-report measure that assesses preceived social support. The MSPSS has three factors that include family, friends, and significant others. Responses are based on the following 7-point Likert-style scoring: 1 (very strongly disagree) to 7 (very strongly agree). Higher scores indicate higher levels of perceived social support. The MSPSS possesses good psychometric properties in various studies (e. g., Pedersen, Spinder, Erdman, & Denollet, 2009; Stewart, Umar, Tomenson, & Creed, 2014) . Also, Salami and colleagues (2009) reported Cronbach's alpha of over 0.80 in the Persian version in all three subscales (friends, family, and significant other). Patton, Stanford, & Barratt, 1995) The BIS-11 is a 30-item self-report measure that assesses impulsiveness. It comprises three factors: attentional impulsiveness, motor impulsiveness, and non-planning impulsiveness. Responses are based on the following 4-point Likert-style scoring: 1 (never) to 4 (always). Higher scores indicate higher levels of impulsiveness. The BIS-11 has good psychometric properties to measure impulsiveness among patients and other populations (Patton et al., 1995) . The validity and reliability study of Persian version of the BIS-11 was established in an Iranian sample. Cronbach's alpha and test-retest were 0.81 and 0.77, respectively (Javid, Mohammadi, & Rahimi, 2012) . Also, in another study Cronbach's alpha for the Persian version was 0.84 (Ekhtiari, Rezvanfard, & Mokri, 2008) . Thomas, 2018) The MDTS is a 20-item self-report measure that assesses one's perceived ability to tolerate physical discomfort, frustration, ambiguity, and emotional pain. Responses are based on the following 5-point Likert-style scoring: 1 (not at all) to 5 (extremely). Higher scores indicate lower levels of distress tolerance. Thomas (2018) study provided evidence of adequate psychometric properties for the MDTS as indicated by the level of internal consistency (The Cronbach's alpha for total score, tolerance of physical discomfort, frustration, ambiguity, and emotional pain was.0.94, 0.92, 0.91, 0.89, and 0.89, respectively). Because the MDTS has not been validated in Iran, for the present study, the English version of the MDTS was translated into Persian by the Iranian authors and a bilingual psychologist. Following this, the equivalence of the two versions was checked and confirmed. The first version for rating the understandability and fluency of each item was completed by 86 participants from the general population of Tehran (F = 28.53%; mean age = 21.03 ± 1.44 years) in a pilot study. This process did not result in any changes in items, and the final version of the Persian MDTS was confirmed. In this pilot study the Cronbach's alpha for total score, tolerance of physical discomfort, frustration, ambiguity, and emotional pain was 0.73, 0.81, 0.78, 0.78, and 0.86, respectively. (Gross & John, 2003) The ERQ is a 10-item self-report measure that assesses emotion regulation. The ERQ has two opposing factors that include cognitive reappraisal, which involves changing the way one thinks about a potentially emotion-eliciting event, and suppression, which involves changing the way one responds behaviorally to an emotion-eliciting event. Responses are based on the following 7-point Likert-style scoring: 1 (strongly disagree) to 7 (strongly agree). Higher scores indicate higher levels of emotion regulation strategies use. The original version of the ERQ had a Cronbach's alpha of 0.79 for reappraisal and 0.73 for suppression. Test-retest reliability across three months was 0.69 for both scales (Gross & John, 2003 ). In the current study emotion regulation was measured using the Persian version of the ERQ. Hasani (2017) reported good reliability for the Persian version in total score (Cronbach's alpha = 0.91), reappraisal (Cronbach's alpha = 0.87), and suppression (Cronbach's alpha = 0.85). (Wells & Cartwright-Hatton, 2004) The MCQ-30 is a 30-item measure that assesses metacognitions. The MCQ-30 has five factors: positive beliefs about worry; (ii) negative beliefs about thoughts concerning uncontrollability and danger; (iii) cognitive confidence; (iv) beliefs about the need to control thoughts; and (v) cognitive self-consciousness. Responses are based on the following 4point Likert-style scoring: 1 (do not agree) to 4 (agree very much). Higher scores indicate higher levels of maladaptive metacognitions. The MCQ-30 has demonstrated good internal consistency and convergent validity and has acceptable test-retest reliability (Spada, Mohiyeddini, & Wells, 2008; Wells & Cartwright-Hatton, 2004) . The Persian MCQ-30 was used in the present study. This measure has been shown to have acceptable internal consistency (Cronbach's alpha = 0.73), test-retest reliability (ICC = 0.73) and acceptable validity (Shirinzadeh, Goudarzi, RAHIMI, & Naziri, 2009) . Also, Cronbach's alpha coefficients of all subscales range between 0.59 and 0.87 (Shirinzadeh et al., 2009 ). Repetitive thinking Questionnaire-10 (RTQ-10; McEvoy, Thibodeau, & Asmundson, 2014) The RTQ-10 is a 10-item self-report measure that assesses RNT. Responses are based on the following 5-point Likert-style scoring: 1 (not true at all) to 5 (very true). Higher scores indicate higher levels of RNT. The RTQ-10 exhibits excellent internal consistency (Cronbach's alpha = 0.94) and is significantly correlated with worry, depression, and anxiety (McEvoy et al., 2014) . The validity and reliability study of Persian version of the RTQ-10 was established in an Iranian sample (Akbari, 2017b) and Cronbach's alpha is reported to be 0.91 (Akbari, 2017b) . It also showed good convergent and divergent validity with measurements of anxiety, depression, rumination, worry, and affect in Iranian samples (Akbari, 2017b) . The total sample of 144 participants were selected via convenience sampling from three universities in Tehran, Iran. The inclusion criterion for the PIU group was scoring higher than the cutoff point on the IAUQ; the exclusion criterion was receiving a CUD diagnosis based on the SCID-5. This was done to prevent the inclusion of participants who had both PIU and CUD. The inclusion criterion for the CUD group was receiving a CUD diagnosis based on the SCID-5; the exclusion criterion was a score greater than the cutoff point on the IAUQ. This was done to prevent the inclusion of participants who have both CUD and PIU. The inclusion criteria for the control group were a score lower than the cutoff point on the IAUQ and the SDS. The exclusion criterion was receiving a CUD diagnosis based on the SCID-5. Participants were recruited using advertisements on social network groups at from three universities in Tehran, Iran. There were three stages to the recruitment of participants. To begin with students who used cannabis daily were invited to fill the IAUQ and the SDS and provide their contact information. Three hundred eighty-three students completed this phase of the recruitment process. Following a review of the responses obtained, 203 students who had recorded the highest score on the SDS (score of 3 or more) and not scored more than the cutoff point (cutoff = 24) in the IAUQ were invited for a face-to-face interview. There were 83 students who agreed to attend the face-to-face interview i.e., the SCID-5 interview, which was conducted by two authors of the present study; out of these, 48 students who met the CUD criteria formed the CUD group. In the second phase of the recruitment process, students who used the internet excessively (more than 3 h per day; Berchtold, Akre, Barrense-Dias, Zimmermann, & Surís, 2018). Were invited to complete the IAUQ and the SDS and provide their contact information. Out of the 443 students who completed the questionnaires, 294 who had the lowest score on the SDS and a score higher than the cutoff point on the IAUQ, were invited to attend the SCID-5 interview. Out of 294 students who were contacted, 102 students agreed to attend the face-to-face interview. Two of the authors of the present study interviewed the participants based on the SCID-5. Finally, to match this group with the CUD group in terms of number, 48 people who were not diagnosed with CUD and scored more than the cutoff point on the IAUQ formed the PIU group. The PIU group was evaluated with the IAUQ. We used a simple random sampling based on an exact number of cases using SPSS 22. In the third phase of the recruitment process, students were invited to complete the SDS and provide their contact information. Seven hundred and thirty people completed this task and 105 of these who had the lowest score on the SDS, and did not have a score more than the cutoff point on the IAUQ, attended the face-to-face interview. Two of the authors of the present study interviewed the participants based on the SCID-5. Finally, to match this group with the CUD group in terms of number, 48 people who were not diagnosed with CUD and scored less than the cutoff point on the IAUQ formed the control group. We used a simple random sampling based on an exact number of cases using SPSS 22. All groups were asked to complete a battery of measures assessing social support, impulsivity, distress tolerance, emotion regulation, metacognitions, and RNT. All participants were informed of the study aims and gave written informed consent before completing the study pack. All procedures performed in studies involving human participants were in accordance with the 1989 Helsinki Declaration and its later amendments or comparable ethical standards. A discriminant function analysis was employed to evaluate predictors of PIU, CUD and control group membership. In this analysis, after examining the differences between the three groups, we sought to find the canonical functions to determine the relationship between predictor variables and criterion variable. Then we identified the best predictor variables across the three groups. Finally, we examined how many members of each group identified with these predictor variables and obtained the accuracy of classifying members. There were no significant differences in gender (χ 2 = 4.235, p = 0.120), age (F = 6.001, p = 0.130) and education (χ 2 = 2.883, p = 0.237) between the PIU and CUD groups and the control group. Before performing the analyses related to the statistical tests, the assumptions of normality, homogeneity of variances and multiple sequence alignment between the study variables were checked. Kolmogorov-Smirnov tests for all variables (except for the cognitive impulsiveness and unplanned variables in CUD group) were not significant, so the assumption of normality for the final analysis was accepted, since skewness and kurtosis for these two variables was very low. Furthermore, the M-box test (as an index of multivariate normality) was not significant (F = 1.128, p = 0.099). To evaluate the alignment the tolerance index and the variance inflation factor (VIF) were calculated. Since none of the values related to the tolerance value was less than 0.01 and none of the values related to the VIF exceeded the cutoff of 10, these assumptions were confirmed (Tabachnick & Fidell, 2013) . Lastly, Levene's test was used to evaluate homogeneity of variance and this was found not to be significant in most instances, indicating that the variances were approximately equal (see Table 1 for Levene's statistics in all variables). However, Levene's test was significant in two of the study variables (reappraisal and suppression), but according to the volume of all three groups, Tabachnick and Fidell (2013) suggest that the overall results are not impacted if the larger variance can be divided into the smaller variance where the result is less than 10. This was the case with the current data, which indicated that the assumption of homogeneity of variance was met. Table 2 presents mean and standard division for all subscales of variables among the PIU group, CUD group, and control group that be compared with ANOVA. Three canonical functions were produced in three separate discriminant function analyses to determine the relationship between predictor variables and the group criterion variable (see Table 3 ). The first discriminant function (F1) was between the PIU group and the control group, the second discriminant function (F2) was between the CUD group and the control group, and the third discriminant function (F3) was between the PIU group and the CUD group. The first discriminant function (F 1 ) had a significant relationship (r = 0.66, χ 2 = 48.96, p < 0.001) with the dependent variables of group membership (PIU and control). The second discriminant function (F 2 ) had a significant relationship with the dependent variable of group membership (r = 0.74, χ 2 = 66.72, p < 0.001). The third discriminant function (F 3 ) also had a significant relationship with the dependent variable of group membership (r = 0.74, χ 2 = 66.82, p < 0.001). Once the relationship between the canonical function, the independent variables consisting of social support, impulsiveness, distress tolerance, emotion regulation, metacognitions, and repetitive negative thinking, and the dependent variable were determined, we progressed to examine the contribution of the independent variables in the formation of the canonical function. Table 4 shows that the first function has a significant relationship with the following variables: non-planning (0.45), motor (0.64), and attentional (0.66), which belong to impulsiveness; frustration (0.40), ambiguity (0.43) and emotional pain (0.51), which belong to distress tolerance; cognitive confidence (0.43), positive beliefs about worry (0.39), negative beliefs about thoughts (0.62), need to control thoughts (0.38) which belong to metacognitions; and RNT (0.66). the second function demonstrated a significant relationship with the following variables: significant other (0.37), family support (0.59), which belongs to social support; motor impulsiveness (0.30), and non-planning impulsiveness (0.41), which belong to impulsiveness; suppression (0.30), which belongs to emotion regulation; need to control thoughts (0.31), which belongs to metacognitions; and RNT (0.66). the third function demonstrated a significant relationship with the following variables: significant other (0.34), family support (0.43), which belong to social support; motor impulsiveness (0.32), which belongs to impulsiveness; physical discomfort (0.34), and emotional pain (0.35) which belong to distress tolerance; reappraisal (0.37), which belongs to emotion regulation; cognitive confidence (0.37), which belongs to metacognitions. As shown above, the first function distinguishes individuals in the PIU group from those in the control group, whilst the second function distinguishes individuals in the CUD from those in the control group. The third function distinguishes individuals in the PIU from those in the CUD group. Finally, the results of classification analysis showed that based on the predictor variables in general, 73.6% of the study participants were correctly placed in their respective groups. Since the degree of separation is more than 50%, it can be said that the discriminant functions worked well (Meyers, Gamst, & Guarino, 2016) . Based on the classification analysis, it was also found that 68.8% of the control group, 70.8% of the PIU group, and 81.3% of the CUD group were correctly classified in their respective group. In the present study, participants in the PIU group had higher scores, compared to the control group, on motor, attentional and non-planning impulsiveness (Lee, Hoppenbrouwers, & Franken, 2019; Mottram & Fleming, 2009 ), tolerance of frustration, emotional pain, and ambiguity (Akbari, 2017a; Ko, Yen, Yen, Chen, & Wang, 2008) , negative beliefs about thoughts, positive beliefs about worry, need to control thoughts, and cognitive confidence (Akbari, Bahadori, Bouruki Milan, Caselli, & Spada, 2021; Hamonniere & Varescon, 2018) , and RNT (Bagatarhan & Siyez, 2020; Dempsey, O'Brien, Tiamiyu, & Elhai, 2019) . Also participants in the CUD group had lower scores, compared to the control group, on family, and significant other support (Ates, Unubol, Bestepe, & Bilici, 2019; Leos-Toro, Shiplo, & Hammond, 2018) , and had higher scores compared to the control group on motor, and non-planning impulsiveness (Lee et al., 2019) , tolerance of frustration, emotional pain, and ambiguity (Hasan, Babson, Banducci, & Bonn-Miller, 2015; Ko et al., 2008) , and suppression (Blanchard, Stevens, Cann, & Littlefield, 2019; Buckner, Walukevich, Zvolensky, & Gallagher, 2017) , and need to control thoughts (Hamonniere & Varescon, 2018; Wasmuth et al., 2015) , and RNT (Hamonniere et al., 2020; Hill, 2020) . These findings are fully aligned with previous research. Individuals in the CUD group reported receiving less social support from family and significant others than those in the PIU group. One reason for this may be that individuals in the PIU group perceive greater support from online social networks and that this online support itself predicts PIU (Hardie & Tee, 2007; Tudorel & Vintilȃ, 2018) . The reason for preferring online social support to offline could be a lack of face-toface social skills (Caplan, 2005) with offline social support having been shown to have a negative relationship with the PIU (Lin et al., 2018) . Furthermore, some personality traits might explain this difference. For example, cannabis users are less agreeable and conscientious than the general population (Terracciano, Löckenhoff, Crum, Bienvenu, & Costa, 2008) . Additionally, personality disorders, such Borderline Personality Disorder (Vest & Tragesser, 2019) and Schizotypal Personality Disorder (Spriggens & Hides, 2015) are more likely to be present in CUD. It could be that these characteristics give-rise to more problems in social communication. In line with this finding, several studies have shown that family conflicts increase the likelihood of substance use disorder (Agha, Zia, & Irfan, 2008; Schafer, 2011) . Individuals in the PIU group reported lower levels of tolerance of physical discomfort (Schmidt, Richey, & Fitzpatrick, 2006) than those in the CUD group. Given that individuals who use cannabis commonly experience physical symptoms such as tachycardia and hypotension (Caplan, 2013; Patel & Marwaha, 2019) it could be that, over time and with use, their ability to tolerate physical discomfort grows. This type of discomfort is limited in individuals who use internet problematically, with users typically reporting postural pain or blurred vision (Panova & Carbonell, 2018) . Therefore, greater discomfort tolerance may be a byproduct of the addictive behavior. On the other hand, it could also be that individuals who have low physical discomfort tolerance are more likely to engage in internet, rather than cannabis use, as internet is a readily available mean of escape, avoidance, and mood regulation (Skues, Williams, Oldmeadow, & Wise, 2016) . In individuals who use cannabis, it has been suggested (Buckner, Keough, & Schmidt, 2007) that the greatest risk of cannabis problems is in those individuals with high capacity to tolerate physical discomfort. Thus, low physical discomfort tolerance could be a protective factor in problematic cannabis use (Buckner et al., 2007) . Our findings are aligned with this suggestion by Buckner et al. (2007) . Further research is needed to disentangle fully these relationships. A systematic literature review showed, for example, associations between distress tolerance factors and psychopathological symptoms and disorders, which included behavioral and drug addictions . Individuals in the CUD group reported lower levels of reappraisal than those in the PIU group. Previous research has suggested that a deficit in reappraisal is associated with craving tobacco (Garland et al., 2018; Szasz, Szentagotai, & Hofmann, 2012) . One explanation for the current findings could be that the CUD group experience more cravings than the PIU group, since nicotine craving is more strongly linked to 'internal' symptoms such as withdrawal (Vollstädt-Klein et al., 2011) . As previously mentioned, physical symptoms are more common in drug addiction. Generally, due to more craving in the CUD group, reappraisal strategies may become more difficult. Another reason could be that drug use can interrupt cognitive and executive function (Cyrus et al., 2021; Dolcos, Iordan, & Dolcos, 2011; Núñez et al., 2016) . Therefore, due to these cognitive changes, the use of reappraisal strategies occurs less in the CUD group. Individuals in the PIU reported lower levels of cognitive confidence than those in the CUD group. Low cognitive confidence leads to greater difficulty in problem-solving, which brings to dysfunctional coping . Also, lower cognitive confidence may be associated with a more consistent way of responding to a range of situations, leading to persistence of responses (Hezel & McNally, 2016; Nedeljkovic & Kyrios, 2007) . This may explain excessive use of the Internet as a means of gaining greater confidence in decision-making . This said, research has also shown that cognitive confidence is a predictor of cannabis use (Brosnan et al., 2020) so further research to tease out the degree of importance of this construct across PIU and CUD is warranted. Several limitations of this study should be noted. First, a crosssectional design cannot be used to infer causality. Second, data were collected from self-report measures, which may have some disadvantages regarding social desirability. Third, recruiting only university students as participants has limitations with respect to generalizability. Fourth, other mental disorders such as anxiety and depression were not assessed when using SCID-5 during the interviews. Fifth, PIU has a wide scope and can include a variety of areas such as Internet Gaming Disorder and problematic social networking sites use. Future studies could examine the differences between the two addictive behaviors at various levels of severity across the general population. Additionally, to aid further generalization, future studies should be conducted on other behavioral addictions, such as gambling or compulsive shopping, and other drug addictions, such as stimulants or opioids lastly, future studies should also examine other psychological variables to differentiate behavioral from drug addictions. This could include, but is not limited to, delay-discounting or experiential avoidance. Despite these limitations, the present study introduces a valuable model based on a variety of psycho-social variables that can distinguish PIU from CUD. It has been shown that social support, tolerance of physical discomfort, reappraisal and cognitive confidence play a significant role in discriminating between these two addictive behaviors. These findings can be of interest to psychological and mental healthcare practitioners working with behavioral (i.e., PIU) and drug addiction (i. e., CUD) clinical groups, during the process of assessment as well as treatment for example, in designing behavioral addictions interventions, emphasis could be placed on increasing cognitive confidence, whereas in designing drug addictions interventions, more attention could be paid to improving family relationships. The authors declare that they have no known competing financial interests or personal relationships that could influence the work reported in this paper. Psychological problems and family functioning as risk factors in addiction Rates of psychiatric comorbidity among US residents with lifetime cannabis dependence Metacognitions or distress intolerance: The mediating role in the relationship between emotional dysregulation and problematic internet use Psychometric Properties of Repetitive Thinking Questionnaire in Nonclinical Sample: Transdiagnostic Tool Metacognitions as a predictor of online gaming in adolescents: Psychometric properties of the metacognitions about online gaming scale among Iranian adolescents Behavioral addiction versus substance addiction: Correspondence of psychiatric and psychological views Emotion Regulation, Motives and Personality Risk for Cannabis Use Problems in Emerging Adulthood: A Person-centered Approach. (Doctoral dissertation) Diagnostic and statistics manual of mental disorders The effect of perceived social support on quality of life in Turkish men with alcohol, opiate and cannabis use disorder Rumination and internet addiction among adolescents: The mediating role of depression Daily internet time: Towards an evidence-based recommendation? The European Journal of Public Health Clinical features, psychiatric comorbidity, and health-related quality of life in persons reporting compulsive computer use behavior Regulate yourself: Emotion regulation and protective behavioral strategies in substance use behaviors Posttraumatic stress, difficulties in emotion regulation, and coping-oriented marijuana use Depressive symptoms, ruminative thinking, marijuana use motives, and marijuana outcomes: A multiple mediation model among college students in five countries The neurobiology and genetics of impulse control disorders: Relationships to drug addictions Parental styles and metacognitions as predictors of cannabis use Problematic alcohol and cannabis use among young adults: The roles of depression and discomfort and distress tolerance Distress tolerance and cannabis craving: The impact of laboratory-induced distress Emotion regulation and coping motives serially affect cannabis cessation problems among dually diagnosed outpatients The association between distress tolerance and cannabis use-related problems: The mediating and moderating roles of coping motives and gender Internet abusive use questionnaire: Psychometric properties A social skill account of problematic Internet use Longitudinal associations of friend-based social support and PTSD symptomatology during a cannabis cessation attempt Dysfunctional inhibitory control and impulsivity in Internet addiction Issues in the study and application of social support Cannabis use disorder: Epidemiology and management Problematic internet use and feelings of loneliness Personality disorders and social support in cannabis dependence: A comparison with alcohol dependence A review investigating the relationship between cannabis use and adolescent cognitive functioning. Current Opinion in Psychology Validation of a new scale for measuring problematic internet use: implications for pre-employment screening Working memory and impulsivity predict marijuana-related problems among frequent users. Drug and Alcohol Dependence Fear of missing out (FoMO) and rumination mediate relations between social anxiety and problematic Facebook use Impulsivity across Substance Use Categories: Consideration of Sex/Gender Neural correlates of emotion-cognition interactions: A review of evidence from brain imaging investigations Self-esteem, coping, perceived social support and substance use in young adults with a cannabis dependence disorder The Perseverative Thinking Questionnaire (PTQ): Validation of a content-independent measure of repetitive negative thinking Impulsivity and its Different Assessment Tools: A Review of View Points and Conducted Researches Problematic smartphone use: A conceptual overview and systematic review of relations with anxiety and depression psychopathology Attachment and emotion regulation in substance addictions and behavioral addictions Varieties of impulsivity The relationship of internet addiction symptom severity with posttraumatic stress disorder symptoms and impulsivity among Turkish university students Relationship of internet addiction severity with probable ADHD and difficulties in emotion regulation among young adults Adverse effects of marijuana use Structured clinical interview for the DSM (SCID). The Encyclopedia of Clinical Psychology Reappraisal deficits promote craving and emotional distress among chronic pain patients at risk for prescription opioid misuse The Severity of Dependence Scale (SDS): Psychometric properties of the SDS in English and Australian samples of heroin, cocaine and amphetamine users Individual differences in two emotion regulation processes: Implications for affect,relationships, and well-being Emotion regulation and mental health The relationships between Turkish adolescents' Internet addiction, their perceived social support and family activities Perceived access and parental monitoring as moderators of impulsivity and marijuana use among adolescents Psychometric Properties of Severity of Dependence Scale (SDS) in People with Drug and Alcohol Dependency Long-term marijuana use and cognitive impairment in middle age Toward a better understanding of the influence of repetitive negative thinking in alcohol use disorder: An examination of moderation effect of metacognitive beliefs and gender Metacognitive beliefs in addictive behaviours: A systematic review Excessive Internet use: The role of personality, loneliness and social support networks in Internet Addiction The prospective effects of perceived and laboratory indices of distress tolerance on cannabis use following a self-guided quit attempt Persian version of the emotion regulation questionnaire: Factor structure, reliability and validity The mediating effect of the cyberchondria and anxiety sensitivity in the association between problematic internet use, metacognition beliefs, and fear of COVID-19 among Iranian online population. Heliyon, 6, Article e05135 A theoretical review of cognitive biases and deficits in obsessive-compulsive disorder Low Regulatory Flexibility as a Mechanism of the Link Between Rumination and Internalizing Symptoms and Substance Misuse in College Freshmen (Doctoral dissertation Psychometric properties of an Iranian version of the Barratt Impulsiveness Scale-11 (BIS-11) Parenting styles, perceived social support and emotion regulation in adolescents with internet addiction How can we conceptualize behavioural addiction without pathologizing common behaviours? Acute and long-term effects of cannabis use: A review Tridimensional personality of adolescents with internet addiction and substance use experience The association between Internet addiction and belief of frustration intolerance: The gender difference The relationship between drug use, sleep quality and quality of life in dormitory students at Allameh Tabataba'i University A systematic meta-review of impulsivity and compulsivity in addictive behaviors Perceived support for medical cannabis use among approved medical cannabis users in Canada Distress tolerance and psychopathological symptoms and disorders: A review of the empirical literature among adults The role of cognitive distortion in online game addiction among Chinese adolescents Association between online and offline social support and internet addiction in a representative sample of senior high school students in Taiwan: The mediating role of self-esteem The association between perceived distress tolerance and cannabis use problems, cannabis withdrawal symptoms, and self-efficacy for quitting cannabis: The explanatory role of pain-related affective distress Psychopathology of addiction: Can the SCL90-based fivedimensional structure differentiate Heroin Use Disorder from a non-substancerelated addictive disorder such as Gambling Disorder? Behavioural (non-chemical) addictions Association between problematic Internet use and impulse control disorders among Iranian university students Trait repetitive negative thinking: A brief transdiagnostic assessment Longitudinal influence of alcohol and marijuana use on academic performance in college students Rumination and problematic substance use among individuals with a long-term history of illicit drug use. Drug and Alcohol Dependence Applied multivariate research: Design and interpretation Internet addiction: A review of the first twenty years Psychometric Properties of Persian Version of Structured Clinical Interview for DSM-5-Research Version (SCID-5-RV): A Diagnostic Accuracy Study Incidence and correlates of pathological Internet use among college students Problematic Internet Use: A Longitudinal Study Evaluating Prevalence and Predictors Psychometric properties of Internet abusive use questionnaire Extraversion, impulsivity, and online group membership as predictors of problematic Internet use Effects of regular cannabis use on neurocognition, brain structure, and function: A systematic review of findings in adults Prevalence of cannabis lifetime use in Iranian high school and college students: A systematic review, meta-analyses, and meta-regression Confidence in memory and other cognitive processes in obsessive-compulsive disorder Evaluation of anxiety and depression among female spouses of Iranian male drug dependents Cannabis use and cognitive function in first episode psychosis: Differential effect of heavy use Clinical validity and intrarater and test-retest reliability of the Structured Clinical Interview for DSM-5-Clinician Version (SCID-5-CV) Is smartphone addiction really an addiction Cannabis use disorder Factor structure of the Barratt impulsiveness scale Personality, marijuana norms, and marijuana outcomes among college students Poor perceived social support in implantable cardioverter defibrillator (ICD) patients and their partners: Cross-validation of the multidimensional scale of perceived social support Prevalence of problematic internet use disorder and associated risk factors and complications among Iranian university students: A national survey Problematic internet use, social needs, and social support among youth The association between internet gaming disorder and impulsivity: A systematic review of literature Family functioning in families with alcohol and other drug addiction Discomfort intolerance: Development of a construct and measure relevant to panic disorder Developmental sensitivity to cannabis use patterns and risk for major depressive disorder in mid-life: Findings from 40 years of follow-up Problematic internet use, mental health, and sleep quality among medical students: A path-analytic model The Addictive Behaviors. Routledge Reliability and validity of severity dimensions of psychopathology assessed using the Structured Clinical Interview Study of factor structure, validity and reliability of metacognition questionnaire-30 The effects of boredom, loneliness, and distress tolerance on problem internet use among university students An overview of problematic Internet use Metacognition in addictive behaviors Metacognitions and emotion regulation as predictors of problematic internet use in adolescents Measuring metacognitions associated with emotional distress: Factor structure and predictive validity of the metacognitions questionnaire 30 Patterns of cannabis use, psychotic-like experiences and personality styles in young cannabis users Validation of the multidimensional scale of perceived social support (MSPSS) and the relationship between social support, intimate partner violence and antenatal depression in Malawi Effects of emotion regulation strategies on smoking craving, attentional bias, and task persistence Using Multivariate Statistics Five-Factor Model personality profiles of drug users Internet dependency: Symptoms, diagnosis and therapy Natural course of behavioral addictions: A 5-year longitudinal study Measuring Distress Tolerance: Development and Validation of the Multidimensional Distress Tolerance Scale, Doctoral dissertation The relationship between cannabis use and measures of anxiety and depression in a sample of college campus cannabis users and non-users post state legalization in Colorado The role of social support on Internet addiction Reliability and validity of the Severity of Dependence Scale for detecting cannabis dependence in frequent cannabis users UPPS-P model impulsivity and marijuana use behaviors in adolescents: A meta-analysis. Drug and Alcohol Dependence Impulsivity as a vulnerability marker for substance-use disorders: Review of findings from high-risk research, problem gamblers and genetic association studies Coping motives mediate the relationship between borderline personality features and alcohol, cannabis, and prescription opioid use disorder symptomatology in a substance use disorder treatment sample Effects of cannabis use on human behavior, including cognition, motivation, and psychosis: A review Severity of dependence modulates smokers' neuronal cue reactivity and cigarette craving elicited by tobacco advertisement Metacognition in persons with substance abuse: Findings and implications for occupational therapists: La métacognition chez les personnes toxicomances: Résultats et conséquences pour les ergothérapeutes A short form of the metacognitions questionnaire: properties of the MCQ-30 Modelling cognition in emotional disorder: The S-REF model The role of gratitude in the development of social support, stress, and depression: Two longitudinal studies Loneliness, social contacts and internet addiction: A cross-lagged panel study. Computor in Human Behavior Are Internet use and video-game-playing addictive behaviors? Biological, clinical and public health implications for youths and adults Relationships between social support, loneliness, and internet addiction in Chinese postsecondary students: A longitudinal cross-lagged analysis Personality profiles of substance and behavioral addictions The multidimensional scale of perceived social support Distress tolerance: Theory, measurement, and relations to psychopathology This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The authors confirm that the data supporting the findings of this study are available.