COVID-19 Pandemic as Subjective Repeated Strains and its Effects on Deviant Behavior in a Sample of Italian Youth

Abstract

Although over the past 2 years several studies have been carried out on the psychological effects of the COVID-19 pandemic on young people, few of them investigated the pandemic as psychosocial strain and its effects on deviant behaviors. According to Agnew’s General Strain Theory, a repeated objective psychosocial strain, such as the pandemic, exerts pressure on deviance when individuals associate with deviant peers and have weak attachment to parents. Using a sample of 568 young Italians (ages: 15–20 years; 65.8% females, 34.2% males) from north, central and south Italy, we tested for the possible correlation between COVID-19 as a repeated psychosocial strain, deviant behaviors and the role of some coping strategies not included in the Agnew’s original theoretical formulation. Results back the thesis that, considering the COVID-19 pandemic as a repeated subjective strain, affect deviance results primarily through association with deviant peers and less through weak attachment with family. The mediating role of coping strategies was found to be weak. The predominant role of the peer group in the genesis of deviant responses to strain will be discussed.

Keywords

COVID-19 pandemic subjective repeated strain coping strategies juvenile deviance

Introduction

The COVID-19 pandemic appeared in Italy in February 2020 and quickly represented a health, social, psychological and economic emergency during the following 2 years. Although young people are not the most infected age group, many research and analysis by qualified observers have shown that Covid-19 has had an effect on the psychological and social conditions of Italian young people (see Mazza et al., 2020). Psychological research, especially during lockdown, has since then focused on the stress responses and psychopathological effects of the pandemic, such as Post Traumatic Stress Disorder (PTSD), Depression, and Anxiety Disorders. However, despite some official reports (Commissione Bicamerale, 2021), little is known about adolescents’ psychosocial reactions during the different waves of the pandemic and the possible association between COVID-19 related stress and some deviant behaviors in Italian young people. Using the “integrated approach model” developed by Agnew (2006), “according to which repeated strains would lead to deviance through weakened social control and delinquent peer affiliation,” this study investigates COVID-19 as psychosocial subjective strain and its non-deviant/deviant effects (change over time of strain and coping responses) during the three pandemic waves. It was run on a sample of 568 Italian youth. We also examined the mediating effects of some coping strategies that were not included in Agnew’s original model.

The COVID-19 Pandemic as a Psychosocial Strain

Numerous psychological research deal with “COVID-19 related stress,” yet a more careful psychosocial analysis has lead us to talk about “COVID-19 related strain” instead. According to Zhang et al. (2014), a strain is different from a simple stress-factor, since it consists of at least two stressors either pulling or exerting a pressure on an individual in different manners. A single directional stress is usually less harmful to an individual’s psychological wellbeing than a strain, and a given subject may have experienced numerous stresses but not many strains. Zhang’s thesis is based on Merton’s framing, meaning that a “strain” refers to the gap between cultural goals and legitimate means. Also taking inspiration from Merton’s framework, Agnew (1992) expands this theory to include additional sources of strain in his “General Strain Theory (GST)”; he originally conceived it to explain why individuals might engage in criminal behavior, stating that strain might refer to any event or condition disliked by individuals experiencing them (Agnew, 2006, p. 4). In the “GST,” he affirms that some adverse/negative environmental demands (e.g., failure to achieve a goal, the removal of positive stimuli or the exposure to negative stimuli) lead to strain in the absence of “adequate” coping strategies. It is clear that the COVID-19 pandemic created adverse conditions for young people: “failure to achieve a goal” (autonomy or independence), “absence of positive stimuli” (no social experiences, reduced personal freedom, not being able to celebrate a birthday or Christmas) or “exposure to negative stimuli” (harsher parental control, loss of a relative or a friend, and/or the worsening of family economic conditions).

Agnew (2001) distinguishes between “objective” and “subjective” strains. An objective strain refers to negative events or conditions that are disliked by most people in a group, such as: economic breakdowns, wars, and pandemic. Subjective strains, on the other hand, refer to events that are rejected by the individuals experiencing them. Individuals often differ in their subjective evaluation of the same objective strain. For example, while for many young people the lockdown had resulted in frustration and loneliness, others have had the opportunity to enjoy greater intimacy within their own family. The kind of emotional response and its intensity to a negative event or condition are closely related to the subjective evaluation to strain. Two people may evaluate the lockdown to be equally harmful and this means they have the same level of subjective strain. On the other hand, one of them may feel (more or less) anger in response to strain, while the other may become anxious. To predict deviance, strains should trigger particularly negative emotional states such as anger, fear, or depression, where anger is thought to be the most criminogenic emotion. However, research has found that strains can trigger different emotional states which, in turn, can be linked to different forms of deviance. For example, Kaufman (2009) expands the GST’s original formulations and states that serious and relevant strains (for both males and females), include externalizing (such as anger, aggression) and internalizing forms of negative emotions (such as depression, suicidal thoughts, and anxiety). The first are dependent variables of strain and mediators for delinquency, victimization, antisocial behaviors; while the latter are dependent variables and mediators for eating disorders, self-harm, suicide, or drug abuse.

A recurring strain, such as a phased epidemic, can require adaptation over prolonged periods. The GST talks about “chronic strains” (Agnew, 2001, p. 334), suggesting that they have a greater negative impact on individuals’ wellbeing. Research on COVID-19 seems to support this statement, backing the idea that COVID-19 is a long-term chronic stress having consequences on people's mental health (Gabrielli & Lund, 2020). Repeated strains foster deviance by creating a general predisposition for crime through reducing social control and creating the conditions for social learning of crime. Direct association and interaction with others who engage in deviant behaviors are the most important contexts where criminal behavior is learned, including peer groups and family units. By using cross-sectional design, Bao et al. (2014) confirmed the relation between attachment to parents, social learning, association with delinquent peers, repeated strain and their effect on youth deviant behavior.

As we have seen, GST suggests that deviance is a coping strategy to adapt to negative emotions related to strain, especially anger. Recently, Agnew (2013, p. 659) stated that a great number of stress, coping, and resilience papers have indicated several key conditioning variables not listed by GST, and that researchers should devote more attention to different coping strategies. According to Broidy (2001), GST should examine the coping strategies that individuals first employ in response to strains. It is said that some individuals tend to employ maladaptive strategies that eventually result in deviant coping, including the maladaptive strategies of which Lazarus and Folkman (1984) talk : opposition, rumination, panic, self-blame, blaming, and denial. This assertion was anticipated and supported by Leeper Piquero and Sealock (2000) when they tested the GST . By using Coping Resources Inventory (Hammer & Hammer, 1988), they found some coping strategies not included on Agnew’s list to have an inhibitory effect on emotions.

Studies on COVID-19 and Juvenile Deviance

International research has investigated two different forms of juvenile deviance (Hay & Meldrum, 2010): “externalizing” acts committed against other individuals (i.e., violent and property crimes, victimization) and “internalizing” deviance (e.g., eating disorders, drug addiction) caused by COVID-19 related stress and negative emotions like guilt, depression, and anxiety.

Research on juvenile delinquency (as an externalizing deviance) and on COVID-19 pandemic has so far produced some results. Estévez-Soto (2021) observed significant variations in the crime typology during pandemic times, due to the changes in people’s routine activities. Campedelli et al. (2021) found that in the first weeks after the implementation of anti-COVID-19 measures, social distancing had a more direct impact on less serious crimes, such as online victimization, involvement in cyberbulling, and revenge porn. Other research confirmed this relationship between COVID-19 and cyberbulling (Agus et al., 2021; Barlett et al., 2021) and other forms of online deviance such as revenge porn (Irwin, 2021).

Regarding “internalizing” deviance, results found by Sun et al. (2020) indicate a substantial increase in COVID-19 related alcohol and drug abuse in young people who used substances before the pandemic. Gili et al. (2021) argue that the lockdown measures produced significant changes in these abuse patterns, with a shift toward the use of substances that were more easily accessible, which were used to alleviate negative emotions and abstinence from drugs that were no longer readily available. According to Taub and McLorg (2011) eating disorders can also be considered as forms of internalizing deviance. Some correlation between COVID-19 effects and the latter was found in recent researches. Specifically, the epidemic is considered to have had a pejorative impact on symptoms among women suffering from Bulimia Nervosa (Schlegl et al., 2020) and Anorexia Nervosa (Dumitraşcu et al., 2021). The impact of COVID has also been verified in other forms of internalizing deviance, such as non-suicidal self-harm (self-cutting, burning, branding, and bone breaking; Adler & Adler, 2007; Hasking et al., 2021).

Current Study

The previous discussion allows us to pin down some hypotheses which will be tested thanks to the data we have collected from a sample of young Italians.

H1. The consequences of the COVID-19 pandemic represent a strain that has had a psychosocial impact on Italian young people (strain reactions and correlated functioning indicators: well-being, psychosocial impairment, emotions).

H1.1. The impact is associated to the subjective evaluation of the event. We expect the strain to vary between individuals who conceive the pandemic as more “adverse/negative,” as opposed to those who consider it presents “advantageous/positive” conditions and aspects.

H1.2. A repeated strain can maintain its strength provided that the evaluation, and emotions do not change significantly over the course of subsequent presentations and coping strategies extinguish their effectiveness. We expect respondents to affirm that they have retained the same evaluations and experienced the same emotions during the three pandemic waves.

H2. Some researchers suppose that the GST should also examine other coping strategies discussed in stress literature, which play a role as conditioning factors that are able to exacerbate strain. We expect maladaptive coping strategies to be conditioning factors for the reported negative emotions during the three waves, thus on their effects on deviant behavior.

H3. Agnew’s GST theorizes that repeated strains increase deviant behavior when the subject associates with deviant peers and when he/she has low attachment to parents. We expect the repeated strains experienced during the pandemic to be coped with deviance: firstly, when young people associate with peers involved in deviant behavior, and secondly, when they have a weak attachment to parents (as per GST).

Methods

The following study was conducted to test the previously mentioned hypotheses.

Participants and Procedure

A total of 568 Italian young people completed an anonymous open-ended online survey, (1.331 participants, 43% response rate. In the case of total non-response, we assume that missing units were completely at random. In the case of missing values, we have imputed an extra category). Our final sample is composed of respondents aged 15 to 20 years, of which: 65.8% are females, 34.2% males; 73.6% are high school students; 4.0% are middle school students; 16.5% are university students; 5.8% are workers. Most of the respondents (76.8%) were enrolled in schools in central Italy (Tuscany, Latium, Marche, Umbria); 9.4% were enrolled in northern Italy (Valle d’Aosta, Liguria, Lombardy, Piedmont, Trentino-South Tyrol, Veneto, Friuli-Venezia Giulia, Emilia-Romagna); 13.9% in southern Italy (Abruzzo, Molise, Campania, Basilicata, Apulia, Calabria) and the islands (Sicily and Sardinia). 7.9% of the respondents reported living with one parent; 70.3%, reported living in a family composed of three to four people (likely two parents and a sibling); 21.8% reported living in a family composed of five or more people. Finally, 7.4% of respondents affirm to have contracted the COVID-19 disease, where the national data for aged 10 to 20 years is 9% (ISS, 2021).

Participant recruitment and survey completion occurred between April 10 and July 15, 2021, when schools were closed and distance-learning was in place. Recruitment was carried out through “convenience sampling,” considered a useful sampling technique for online surveys (Ritter & Sue, 2007). With the purpose of obtaining the sample, the recruitment process was divided into two smaller steps: (1) contacting the headmasters of some high schools and middle schools, and the deans of two universities, and the director of some youth sports and recreational centers who were contacted by email, where we presented our research project and asked for their cooperation and permission to involve the students, (2) A “formal invitation email” with information about the research for participants and a link to the survey questionnaire was sent to headmasters, deans and directors who had agreed to collaborate, who would then forward it to students and young people aged 15 to 20. Participants had to provide informed consent before completing the 30-min online survey (a long and articulated questionnaire of 360 variables), which was completely anonymous. They were informed that the research respected all the ethical standards on the protection of personal data, that participation was voluntary and did not involve any risk for one’s health. Participants were asked to answer questions as spontaneously and sincerely as possible. Finally, they were informed that they would often find the reference to the “three waves” of the epidemic in the questionnaire. “First wave” referred to the moment of the declaration of the first lockdown (March 2020, with the first schools closing in some Italian regions), the second wave (beginning of November 2020) and the third wave (beginning February 2021 to date).

Measures

These are the measures we used to test our hypothesis.

Strain Measures

A 19-items scale was adopted to measure strains during the three waves of pandemic. Respondents were asked to what extent (response options: 1 = not at all, 2 = slightly, 3 = moderately, 4 = very much, 5 = very much) they felt that the listed aspects had been adverse or negative for them during the pandemic waves. Most of these events/conditions are of the type suggested by Magson et al. (2021) in COVID-19 Related to Stress Scales (e.g., risk of contagion, social isolation, relationship with relatives, relationship with professors, academic study, relationship with partner). In order to validate the different scales Explorative Factor Analysis (EFA) was implemented in SPSS IBM Statistics (extraction based on eigenvalue, eigenvalues over 1, Varimax rotation method). In the first wave (total variance 38.48%, α = .84) three items were excluded from analysis because they saturated more than a few factors: first factor included five variables, factor loading from .38 to .70; second factor included three variables, factor loadings from .51 to .76; third factor included four variables, factor loadings from .41 to .60; and fourth factor, comprised four variables, factor loadings from .36 to 63. In the second wave (total variance .59,87%, α = .82) three items were excluded from analysis because they saturated more than a few factors. The first factor included five variables, factor loadings from .58 to .78; second factor included three variables, factor loadings from .76 to .79; third factor comprised three variables, factor loadings from .65 to .75; fourth factor encompassed two variables, factor loadings from .72 to.76; fifth factor included three variables, factor loadings from .54 to 79. Finally, in the third wave (total variance 59.87, α = .82) three items were excluded from analysis because they saturated several factors. The first factor encompassed three variables, factor loadings from .51 to .77; second factor included six variables, factor loadings from .37 to .69; third factor included three variables, factor loadings from .51 to .70; fourth factor included four variables, factor loadings from .40 to .65.

An index was created using Categorical Principal Component Analysis (CatPCA) (Gifi, 1990) technique implemented in IBM SPSS Statistics. The component loadings for the first wave from .21 to .73, total variance .8% α = .77; for the second wave from .31 to .61, total variance .7%, α = .90; for the third wave from .31 to .61, total variance 6.7%, α = .90).

In order to maximize the phenomenological aspect of strains (strain reaction depends on the subjective evaluation of the event) in the context of a second scale, respondents were asked to what extent they had experienced each of sixteen valued advantageous/positive conditions (e.g., listen to lots of music; talk more with my parents than before). Response options: 1 = not at all, 2 = slightly, 3 = moderately, 4 = very, 5 = very much). EFA results are as follows: first wave, total variance .33,06%, α = .75: first factor included five variables from .34 to .60; second, three variables, .42 to .80; third, two variables, .48 to .79; fourth factor, three variables, .33 to .53; fifth factor, three variables, 35 to 48. For the second wave, a variable was excluded (saturation in several factors), variance 33.81%, α = .75, first factor, 13 variables, .30 to .52; second factor, two variables, .40 to .44. For the third wave, variance 35.12%, α.73, first factor, two variables, .63 to .72; second, six variables, .32 to .56; third factor, three variables, .38 to .76; fourth, three variables, .39 to .57; fifth factor, two variables, .43 to .59.

CatPCA component loadings for first wave from .30 to .67, total variance 3.6%, α = 77; for second wave from .30 to .61, total variance 2.1%, α = .55; and third wave from .20 to .62, total variance 1.9%, α = .53.

Emotion Measures

As mentioned before, some studies have shown that strains produce different emotions associated to deviant behaviors. We used a scale of 25 items asking respondents how often they had experienced the negative emotions and feelings (e.g., sadness, anger, anxiety) and three positive emotions (e.g., peace of mind, serenity, mastery) hereby reported during the three waves and before the pandemic: 1 = never, 2 = rarely, 3 = sometime, 4 = often, 5 = always. Several items were derived from different validated scales. Two items derived from Nocentini et al. (2021) (“I had trouble concentrating,” “I was nervous”). Nocentini et al., in turn, derived these items from “The Impact of Event Scale–6 (IES-6)” a validated, shortened version (Thoresen et al., 2010), of the full IES-revised (Weiss & Marmar, 1996) used to assess subjective distress caused by traumatic or stressful events. The short version was recently validated in Italy (Giorgi et al., 2015). Two items (e.g., “Since the beginning of the COVID-19 epidemic, I have felt unable to control the important things in my life,” “I have been angry”) were derived from a scale developed by Liu and Wang (2021) to assess how adolescents’ perceived stress following the COVID-19 pandemic. This scale derived from the CPSS-14 developed by Cohen et al. (1983), and consisting of two subscales: perceived stress and perceived lack of control, was revised by Liu and Wang. Three items measuring fears (“fear of becoming infected,” “fear of becoming ill and dying,” “fear of gaining weight”) were derived from Herbert et al. (2021), four items derived from Droit-Volet et al. (2020) measuring different negative and positive emotions (“I felt sad,” “I felt confused,” “I felt calm and peaceful,” “I felt joyful”) and other items listed, of which we are including only the more statistically relevant (considered negative or positive for at least 50% of the respondents in each of the three different waves); which were: “I thought nostalgically to the past before the COVID-19 pandemic,” “The feeling that time had stopped,” “I was afraid we’d never get out of this pandemic,” “I had the feeling that I could not remember the faces of my friends.”

EFA was implemented for all different scales measuring emotions, resulting as follows: in the first wave (explained 69% of total variance 45.33%, α = .86), two variables are not included (saturated in several factors). These are: first factor was composed of nineteen variables, factor loadings from .30 to .74; second factor included three variables, factor loadings from .40 to .55; third factor, one variable, factor loading .34. For the second wave (variance 43.30%, α = .86) two variables are not included: first factor, eleven variables, factor loadings from .30 to .69; second factor, three variables, factor loadings from .30 to .52; third factor, four variables, from .36 to .71; fourth factor, two variables, from .72 to .73; fifth factor, three variables, from .34 to .72. For the third wave (variance 44.23, α = .86) a variable was excluded (saturation in different factors): first factor, nine variables, from .35 to .60; second factor, five variables, .39 to .72; third factor, four variables, .33 to .73; fourth factor, four variables, .28 to .75; fifth factor, two variables, .56 to .81. Before the pandemic (variance 42.70, α = .82): first factor, nine variables, .28 to .59; second factor, four variables, .40 to .75; third factor, four variables .30 to 58; fourth factor, three variables, .28 to .63, fifth factor three variables, .28 to .63, sixth factor two variables .57 to .61.

CatPCA analysis was implemented on the three different waves and pre-pandemic. First wave, two factors were extracted and together explained 10.3% of the total variance, where α = .94: first factor, component loadings from .37 to .77 on all negative emotions; second factor from .34 to .63 on positive emotions. Second wave, total variance 10.4%, α = .94 were extracted two factors: first factor, component loadings from .37 to .78 in all negative emotions, second factor from .65 to .66 in all positive emotions. Third wave, total variance 10.4%, α = .91, first factor from .38 to .80 on negative emotions, second factor from .31 to .66 on positive emotions. Finally, emotions already experienced pre-pandemic, total variance 10.4%, α = .94, first factor from .37 to .80 in negative emotions, from .56 to .75 on positive emotions.

Adaptive and Maladaptive Coping Measures

We assessed the strategies to cope with negative emotions by using ten items derived from Brief COPE (Carver, 1997; α = .55), an abbreviated version of the COPE Inventory (Carver, 1989). This consists of 14 subscales, with only two items per scale, and it measures cognitive and behavioral coping mainly referred to as: “adaptive” or “maladaptive” (Folkman & Moskowitz, 2000). Adaptive strategies are “active coping,” “planning,” “positive reframing,” “acceptance,” “humor,” “religion,” “use of emotional support,” and “use instrumental support.” Maladaptive coping includes “self-distraction,” “denial,” “venting negative emotion,” “substance use,” “behavioral disengagement,” and “self-blame” (Carver, 1997). We used the following items taken from the Italian version (Caricati et al., 2015): “I’ve been taking action to make the situation better” (active coping); “I’ve been saying to myself ‘this isn’t real’” (venting); “I’ve been getting emotional support from others” (emotional social support); “I’ve been getting help and advice from other people” (instrumental support); “I’ve been giving up trying to deal with it” (behavioral disengagement); “I’ve been trying to see things in a different light, to make it seem more positive” (positive reframing); “I’ve been making fun of the situation” (humor); “I’ve been accepting the reality of the fact that it has happened” (acceptance); “I've been trying to find comfort in my religion” (religion); “I’ve been blaming myself for things that happened” (self-blame). Multi Correspondence Analysis (MCA) was implemented, resulting in three factors being identified, total variance 48.8, α = .89. First factor: component loadings from .24 to .42, with greater loadings in all variables of adaptive coping. Second factor: component loadings from .33 to .45, where some variables of adaptive coping (“use of emotional social support,” “use of instrumental support,” “positive reframing,” “humor”); have the most amount of loadings; and the third factor, component loadings from = .09 to .25, where the variables of maladaptive coping have the greatest loads: “venting negative emotion,” “behavioral disengagement,” “self-blame.”

Personal Deviance Measures

Personal deviance was measured by asking respondents how often they had performed each of the sixteen deviant acts since the beginning of the pandemic and not before the epidemic (factor loadings from .22 to .83). These were the possible outcomes: 1 = never. 2 = rarely, 3 = sometimes, 4 = “often,” 5 = “usually.” The inventory of delinquent behaviors was taken from Blaya and Gatti (2010) and represents a list of the deviant behaviors commonly committed by Italian youths. The analyzed deviant acts were the following: theft and robbery, assault, consumption of light drugs, consumption of hard drugs, sexual performances on webcam (camgirls, camboys), vandalism, gang fights, alcohol abuse, drug selling, sexual assault (also online, e.g., revenge porn), driving under the influence of alcohol; bingeing and purging, depression and anxiety, self-injury; face to face bullying, cyberbullying; sending or sharing intimate or sexually explicit messages, images, photos or videos via Web or WhatsApp.

Social Control Measures

Attachment to parents was measured by five items derived from Labos (1994) asking how the respondents considered their relationship with them had evolved over the last year (separately questions were asked regarding father and mother), α = .77. “Excellent: I felt (him or her) very close, I am very attached to him/her, there is mutual caring, understanding and respect”; “Good: I felt her/him close, I’m attached to her/him, there is a certain understanding and respect for one another”; “Acceptable. Sometimes we were close, I haven’t always felt bonded to her/him, confidence and respect are sometime present”; “Bad: I didn’t feel her/him close to me, even though we lived under the same roof. We have had various misunderstandings and we do not respect each other”; “Very bad. We are emotionally distant. We don’t understand each other. She/He doesn’t respect me.”

Social Learning (Peer Deviance) Measures

Peer deviance was measured by asking respondents how many of their close friends had performed the same sixteen deviant acts that they had performed during epidemic, the answers being: “none,” "few of them,” “some of them,” “many of them,” “all of them.” Exploratory Factor Analysis (EFA) identified two main factors. The first factor was comprised of (factor loadings from .23 to .85 in all deviant variables): theft and robbery, assault, consumption of light drugs, consumption of hard drugs, sexual performances on webcam (camgirls, camboys), vandalism, gang fights, alcohol abuse, drug selling, sexual assault (also online, e.g., revenge porn), driving under the influence of alcohol; bingeing and purging, self-injury; face to face bullying, cyberbullying; sending or sharing intimate or sexually explicit messages, images, photos or videos via Web or WhatsApp. The second factor, with the exception of “sending or sharing intimate or sexually explicit messages, images, photos or videos via the Web or WhatsApp”: .02, the loading of almost all the remaining variables is negative.

Results

COVID-19 as a Subjective Strain

We stated (hypothesis 1.1.) that there would be a difference in the subjective reaction to strains (emotions and sensations) during the three waves of the pandemic. This occurred between people for whom it was perceived as more “adverse/negative,” and those for whom COVID-19 has had “advantageous/positive” aspects. We expect respondents to affirm that they have retained the same evaluations and experienced the same emotions during the three epidemic waves. The data in Table 1. (correlational matrix) provides a first measure of the association between the variables of interest. As can be seen, there are generally positive statistically significant association between measures of strain and emotions. “Adverse/negative evaluation” has a set of positive, statistically significant, association between them and emotions during all three waves, except with what we call “emotions before pandemic,” where the association is not statistically significant. The “advantageous/positive evaluation” has positive, statistically significant, association between them and between emotions in all three waves. The association between the “emotions experienced in the different waves” are positive, statistically significant and constant throughout all the phases of the pandemic. The association between pre-COVID-19 emotions and those experienced during the pandemic are statistically significant and they grow over the three phases, with the above-mentioned exceptions. There is modest but statistically significant association between “attachment to parents” and the evaluations of the emotions experienced during the three phases. The “coping” (third factor) has a series of modest but statistically significant association with the emotions of all three phases, with those experienced before the pandemic, and with “adverse/negative” perceptions of strain in the second and third waves. Regarding deviant subjects (“individual post-pandemic deviance”), all correlations among evaluations and the emotions experienced during the three waves of the pandemic are not statistically significant, while a modest, statistically significant, association with “before COVID-19” emotions exists. A mild, but statistically significant association between the “attachment to parents” and the “individual deviance post pandemic” is observed. Differently from what Agnew suggests, there is no statistically significant association between “attachment to parents” and “peer deviance.” Among deviant subjects, contrary to non-deviant ones, the association with “maladaptive copings” (third factor) is modest but statistically significant. Finally, between “individual deviance” and “peer deviance” the association (first and second factor) is rather important.

Table 1.

Correlations Between Variables Used in the Analysis.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Adverse/negative first wave	1
2. Advantage/positive first wave	.358	1
2. Advantage/positive first wave	.000	1
3. Adverse/negative second wave	.733	.312	1
3. Adverse/negative second wave	.000	.000	1
4. Advantage/positive second wave	.281	.831	.364	1
4. Advantage/positive second wave	.000	.000	.000	1
5. Adverse/negative third wave	.424	.224	.806	.308	1
5. Adverse/negative third wave	.000	.000	.000	.000	1
6. Advantage/positive third wave	.216	.713	.334	.902	.350	1
6. Advantage/positive third wave	.000	.000	.000	.000	.000	1
7. Emotions first wave	.540	.306	.477	.239	.379	.212	1
7. Emotions first wave	.000	.000	.000	.000	.000	.000	1
8. Emotions second wave	.404	.276	.544	.249	.482	.223	.774	1
8. Emotions second wave	.000	.000	.000	.000	.000	.000	.000	1
9. Emotions third wave	.300	.241	.434	.210	.460	.198	.657	.876	1
9. Emotions third wave	.000	.000	.000	.000	.000	.000	.000	.000	1
10. Emotions before pandemic	.069	.253	.107	.239	.159	.205	.332	.410	.452	1
10. Emotions before pandemic	.099	.000	.010	.000	.000	.000	.000	.000	.000	1
11. Attachment to parent	.051	−.058	.082	−.094	.070	−.096	.151	.187	.205	.230	1
11. Attachment to parent	.228	.165	.052	.026	.098	.022	.000	.000	.000	.000	1
12. Coping (third factor)	.116	.015	.200	.078	.222	.063	.205	.226	.217	.174	.200	1
12. Coping (third factor)	.006	.721	.000	.064	.000	.133	.000	.000	.000	.000	.000	1
13. Peer deviance (first factor)	.067	.107	.056	.162	.106	.146	.051	.069	.052	.114	.098	.224	1
13. Peer deviance (first factor)	.111	.010	.185	.000	.012	.001	.224	.102	.214	.006	.019	.000	1
14. Peer deviance (second factor)	.034	.080	.014	.059	.021	.047	.064	.064	.052	.107	.080	−.039	.358	1
14. Peer deviance (second factor)	.419	.055	.734	.157	.615	.267	.131	.130	.213	.010	.056	.358	.000	1
15. Individual deviance post pandemic (first factor)	.053	.072	.029	.107	.049	.088	.074	.091	.080	.147	.230	.223	.549	.125	1
15. Individual deviance post pandemic (first factor)	.205	.088	.488	.011	.240	.036	.079	.030	.057	.000	.000	.000	.000	.000	1

In order to better test the association between strain, evaluation and emotion, we sharpened the analysis by using a linear regression. We elaborated three models, whose data is reported in Table 2. The first model shows how the factors “adversity/negativity during the first wave,” “emotions before pandemic,” and “advantages/positivity during the first wave” (not significant) regressed to “emotions of first wave.”

Table 2.

Linear Regression of Adversity/Negativity on Emotions During Three Waves.

	Standardized coefficients Beta	t	Sig.
Model 1^a (R Square = 0.380)
Adversity/negativity first wave	.520	15.657	.000
Emotions before pandemic	.298	8.965	.000
Advantages/positivity first wave	0.055	1.508	.132
a. Dependent variable: emotions of first wave
Model 2^b (R Square = 0.669)
Emotions of first wave	.599	20.605	.000
Adversity/negativity second wave	.238	8.633	.000
Emotions before pandemic	.184	7.159	.000
Advantages/positivity second wave	−.032	−1.183	.237
b. Dependent variable: emotions of second wave
Model 3^c (R Square = 0.780)
Emotions of second wave	.801	32.798	.000
Emotions before pandemic	.116	5.356	.000
Adversity/negativity third wave	.056	2.464	.014
Advantages/positivity third wave	−.027	−1.262	.207
c. Dependent variable: emotions of third wave

In the second model, “emotions during the first wave,” “adversity/negativity during the second wave,” “emotions before pandemic,” “advantages/positivity during the second wave” (not significant) regressed to the “emotions of the second wave.” Finally, in the third model, “emotions during the second wave,” “emotions before the pandemic,” “adversity/negativity during the third wave,” and “advantages/positivity during the third wave” (not significant) regressed to the “emotions of the third wave.” The linear regression describes that adversity has a positive and statistically significant effect on emotions as a dependent variable of the first, second and third waves, whose association decreases in the three models. In all three models, emotions “experienced before the pandemic” have a positive and statistically significant association with emotions, but this association decreases over the course of the three waves. Emotions are always positively related to each other and, contrary to what happens to other variables, they grew over the course of the three waves. Finally, “advantages/positivity” are statistically not significant throughout the model. Regression results suggest that it is important to consider individuals’ subjective reactions to COVID-19’s objective strain. When strains are subjectively evaluated as high in negativity, they are also associated with high levels of negative emotions. This seems to confirm our Hypothesis no.1, 1.1. The collected data shows us a trend that is different from what we expected for emotions (hypothesis 1.2.): in contrast to the evaluations, they grew during the three waves.

Path Analysis

We theorized that maladaptive coping methods are conditioning factors for negative emotions during all three waves of the pandemic (hypothesis 2.) and that the emotions caused by COVID-19 related, repeated strain are coped with deviance when young people associate with peers involved in deviant behavior and when they have a weak attachment to parents (hypothesis 3.).

Based on the results of the previously reported measurements (except for EFA used in order to validate strain and emotion scales), we set up a path model, as shown in Figure 1. This analysis was performed using the PROC CALIS in SAS System Software (SAS Institute, 2013). It is to be noted that all the variables considered in this model have had a modest effect, except “peer deviance.” As you can see both the “emotions of first wave” (0.17, t = 3.87, SE = .04, p = .00), and emotions which had been experienced “before the pandemic” (0.12, t = 2.67, SE = .04, p = .00), have had little statistically significant, direct effect on “maladaptive coping of third factor”; which has in turn barely any direct statistically significant effect on the first factor of “post COVID-19 individual deviance” (0.07, t = 2.03, SE = .03; p = .04.). This data provides a kind of confirmation to our second hypothesis prediction: “maladaptive coping exerts a conditioning role on emotions and, subsequently, on individual deviance.” In the path model, only emotions from the first wave are associated with maladaptive coping, even if they are related also to the second and third wave emotions (see correlation matrix at Table 1). The role of coping of the third factor on emotions is moderate, as its impact on “individual deviance of first factor” is, which seems to indicate that these coping strategies do not play a decisive role when it comes to other variables in the path. However, the effect of “coping of third factor” on “deviance of first factor” is still weak evidence of the relationship/association between COVID-19 related strain and deviant behaviors.

Figure 1.

Path model.*

The GST affirms that repeated strains increase deviant behavior when the subject associates with deviant peer and when he/she has a low attachment to the parents, thus we expected a correlation between the two factors; although, our data showed a prevalence of the “peer group” compared to “attachment to parents.” The first factor of peer deviance has a major direct and significant effect on individual deviance (0.55, t = 17.26, SE = .03, p < .0001). On the other hand, the second factor of peer deviance (“sending or sharing intimate or sexually explicit messages, images, photos or videos via Web or WhatsApp”) has a lower direct effect on deviance (0.08, t = 2.16, SE = .03, p = .03). Finally, “family attachment,” as opposed to “peer group,” has had little direct effect on individual deviance during COVID-19 (0.14, t = −4.03; SE = .034, p < .0001). These associations: (1) seem to confirm Agnew’s statement according to which repeated strain is coped with deviance when young people associate with peers involving in deviant behavior. (2) A predominant association between peer deviance of the first factor and individual deviance of the first factor seems to suggest a concurrence between individual and group deviant behaviors. (3) Contrary to what was hypothesized based on the GST thesis in our data, family attachment does not appear to have a significant effect on individual deviance.

Discussion

Our data shows some interesting aspects and raises new questions that deserve further attention from researchers.

Firstly, GST affirms that many of the people who experience objective strains do not evaluate them in an adverse way; whether such individuals may have a positive reaction, or no reaction at all, depends on the subjective evaluation of the event (subjective strain). Our data seems to confirm this postulate. We found that, for most subjects, COVID-19 strain is associated with different evaluations for the three waves, and that these evaluations are associated with different emotions (negative emotions for those who evaluate the pandemic negatively, positive emotions for those who give a positive evaluation). Furthermore, the subjects of our sample affirm that during the three epidemic phases the emotions, unlike what happened for the evaluations, increased their intensity. This finding may perhaps suggest a revision of the role of emotions in the strain process. Emotions seem to be not only an effect of evaluation, but also an element that can operate independently from a person’s interpretation of his/her relationship with the environment or with adverse events. Our hypothesis 1.2. was based on the idea of the cyclical process of the dynamics of strain. Emotional responses unfold over an undetermined amount of time that spans between the event, the evaluation, coping and the re-valuation processes during which the implications and consequences for longer-term goals are assessed, as well as the individual’s ability to adjust to, or cope with these consequences. This explanation of dynamics partially fails in our results. Even if the evaluations did not drastically decrease during the three phases, the decreasing process of their strength was slow, but constant. This could have several explanations and there is still considerable controversy regarding the relationship between evaluation and emotion. Some researchers are following new explanatory models according to which the representation of strain deposited in memory would play an important role. For example, Marsella and Gratch (2009) make a distinction between evaluation (they speak about “appraisal”) and inference, arguing for a single and automatic appraisal process that operates beyond a person’s interpretation of their relationship to the environment, assuming that it is a memory-based process that makes quick connections and provides evaluation information based on activated memories, and then quickly associates with the given stimulus. Since emotional responses are often rapid and seemingly automatic, the authors believe that in the stress/strain process emotions are not simply the immediate effects of evaluations, but play a different and more active role in the interpretation of a stressful event.

Secondly, subjects exhibiting deviant behaviors during pandemic had already experienced many of the listed emotions before the pandemic (e.g., anxiety, depression, anger, and so on). This finding is in line with one of the basic assumptions of GST, where Agnew (2006) states that a high, negative emotionality is strongly related to deviance. Research (Mazerolle & Maahs, 2000) has shown that subjects who have high levels of negative emotionality may also respond to strain with deviant behavior. This could in part explain the low levels of emotion intensity of the deviant group (see above), suggesting that the pandemic did not trigger any different or new emotions in them.

Thirdly, coping of the third factor is present in the path analysis, even if with little effectiveness. This finding provides modest support for the idea that there are other coping strategies not listed by GST that can act in deviant behavior and increase the strength of a strain’s impact. However, it is necessary to reflect that we have only used a few of Carver’s COPE Inventory Strategies, making it necessary to investigate further by using more comprehensive measures.

Fourthly, association with peers involved in deviant behavior has a robust impact on individual deviance (.63 two variables combined, “peer deviance first factor” and “peer deviance second factor”), while attachment to parents has a weaker impact. Why did the predicted combined impact between “peer group” and “attachment to parents” on deviant behavior not occur? Why is there this great prevalence of peers over family influence in our data? The importance of the peer group is predominant. Our study confirms its leading role in the context of adolescent development and, as it is known traditionally, as a vehicle for modeling deviant behavior. However, the strong difference in impact on individual deviance between this variable and the others considered in the model could also conceive other explanations.

Some plausible explanations in this regard are not provided by the GST, but can be found in recent analyses of the condition of Italian youth. Some research (e.g., Istituto Giuseppe Toniolo, 2021) shows that in the last ten years, peer groups among Italian adolescents have assumed a different and more important role than in the past. According to Pietropolli Charmet (2018) today peer groups for young Italians represent a kind of “social family” that accompanies, guides and supports the adolescent in carrying out developmental tasks, by performing various functions: security, protection, sense of belonging, avoidance of loneliness, affection and attachment, social reference, identity, and support. The peer group has assumed a strong normative power for young Italians, by dictating the rules of, and shaping almost all deviant behaviors. Some of these functions, except modeling of deviance, but not always, have been typically exercised by the natural family. This “social family,” according to Pietropolli Charmet, can replace the “traditional family.” In a situation such as the one that we all have experienced because of the restrictive measures put in place to contain the pandemic, increased restrictions on personal freedom and limitations on “peer group power” making it likely to have caused an increase in the strength and cohesion of the group. It is possible that our data reflects the changes youth dynamic experts are talking about. According to Agnew, a repeated strain can damage the bonds with parents. Although percentage data have confirmed that deviant individuals are less attached to parents than non-deviants are, we did not find a statistically significant relationship between peer influence and attachment to parents.

A number of questions remain unanswered that further research will be able to answer. An important question concerns the continuity: will COVID-19 as a strain be able to have long-term effects on deviance? Obviously, in the absence of empirical evidence, we do not have an answer. Agnew stated that in the maintenance and in the continuity of deviance over the life course have plays a primary role the association between repeated strain, the influence of deviant peers, weak attachment to parents factors and certain personality traits such as negative emotionality and low restraint. Another important question concerns possible differences between adults and youth in deviant responses to COVID as repeated strain. We do not have empirical data that would allow us to extend the considerations to adults. In this study, we followed the traditional path of youth deviance research, focusing on youth. However, we believe that studies of deviant reactions to objective strain should consider comparisons across different age groups.

It is important to note that our findings have some methodological limitations. One of them is the use of cross-sectional data in the analysis of repeated strain. Cross-sectional studies cannot provide causal evidence due to the potential presence of confounding factors. In addition, relationships among maladaptive coping, social control, social learning and deviant variables can involve reciprocal, rather than unidirectional causation. On the other hand, some researchers (Campedelli et al., 2021) who have been tasked with investigating responses to the ongoing pandemic have proposed the use of other methods, such as the “interrupted time series” typical of quasi-experimental studies, or the longitudinal design, not failing to mention the limitations of these methods in the analysis of the COVID-19 pandemic. Rogowska et al. (2021) stated that the short time lapse between epidemic phases was a methodological difficulty in conducting a longitudinal study, since memory bias might occur. Given the nature of our study, a longitudinal design would have been preferable, yet hardly feasible due to practical obstacles such as the course of the epidemic itself, which has been unpredictable in its evolution since the beginning. Many experts believed the pandemic would be short-lived, and the lockdown was supposed to mitigate the infections. Our data was collected when the pandemic was still active and social and school restrictions were in place, as we tried to capture the reactions of the respondents during the days of the pandemic.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Giacinto Froggio

References

Adler

A. P.

Adler

(2007). The demedicalization of self-injury: From psychopathology to sociological deviance. Journal of Contemporary Ethnography, 36(5), 537–570.

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30(1), 47–84.

Agnew

(2001). Building on the foundation of general strain theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime and Delinquency, 38(4), 319–361.

Agnew

(2006). Pressured into crime. An overview of general strain theory. Roxbury.

Agnew

(2013). When criminal coping is likely: An extension of general strain theory. Deviant Behavior, 34(8), 653–670.

Agus

Mascia

M. L.

Zanetti

M. A.

Perrone

Rollo

Penna

M. P.

(2021). Who are the victims of cyberbullying? Preliminary data towards validation of “Cyberbullying Victim Questionnaire.” Contemporary Educational Technology, 13(3), 1–12.

Bao

W. N.

Haas

Chen

(2014). Repeated strains, social control, social learning, and delinquency: Testing an integrated model of general strain theory in China. Youth& Society, 46(3), 402–424.

Barlett

C. P.

Simmers

M. M.

Roth

(2021). Comparing cyberbullying prevalence and process before and during the COVID-19 pandemic. The Journal of Social Psychology, 161(4), 408–418.

Blaya

Gatti

(2010). Deviant youth groups in Italy and France: Prevalence and characteristics. European Journal on Criminal Policy and Research, 16(2), 127–144.

10.

Broidy

(2001). A test of General Strain Theory. Criminology, 39(1), 9–33.

11.

Campedelli

G.M.

Aziani

Favarin

(2021). Exploring the immediate effects of COVID-19 containment policies on crime: An empirical analysis of the short-term aftermath in Los Angeles. American Journal of Criminal Justice, 46(5), 704–727. https://doi.org/10.1007/s12103-020-09578-6

12.

Caricati

Foà

Fruggeri

Tonarelli

(2015). COPE-NVI-25: Validazione italiana della versione ridotta della Coping Orientation to the Problems Experienced (COPE-NVI). Psicologia della Salute, 2, 123–140.

13.

Carver

C. S.

(1997). You want to measure coping but your protocol’s too long: Consider the Brief COPE. International Journal of Behavioral Medicine, 4(1), 92–100.

14.

Carver

C. S.

Scheier

M. F.

Weintraub

J. K.

(1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56(2), 267–283.

15.

Cohen

Kamarck

Mermelstein

(1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385–396.

16.

Commissione Bicamerale. (2021). Relazione intermedia XX Comitato per la prevenzione e la repressione delle attività predatorie della criminalità organizzata durante l’emergenza sanitaria. Atti Parlamentari, 177, 285–293.

17.

Droit-Volet

Gil

Martinelli

Andan

T. N.

Clinchamps

Parreira

Rouffiac

Dambrun

Huguet

Dubuis

Pereiraet

(2020), Time and COVID-19 stress in the lockdown situation: Time free, “Dying” of boredom and sadness. PLOS ONE, 15(8), e0236465. https://doi.org/10.1371/journal.pone.0236465

18.

Dumitraşcu

M. C.

Şandru

Carsote

(2021). Anorexia nervosa: COVID‑19 pandemic period (Review). Experimental and Therapeutic Medicine, 22, 804. https://doi.org/10.3892/etm.2021.10236

19.

Estévez-Soto

P. R.

(2021). Crime and COVID-19: Effect of changes in routine activities in Mexico City. Crime Science, 10, 15. https://doi.org/10.1186/s40163-021-00151-y

20.

Folkman

Moskowitz

J.T.

(2000). Positive affect and the other side of coping. American Psychologist, 55(6), 647–654.

21.

Gabrielli

Lund

(2020). Acute-on-chronic stress in the time of COVID-19: Assessment considerations for vulnerable youth populations. Pediatric Research, 88, 829–831. https://doi.org/10.1038/s41390-020-1039-7

22.

Gifi

(1990). Non linear multivariate analysis. Wiley.

23.

Gili

Bacci

Aroni

Nicoletti

Gambelunghe

Mercurio

Gambelunghe

(2021). Changes in drug use patterns during the COVID-19 pandemic in Italy: Monitoring a vulnerable group by hair analysis. International Journal of Environmental Res Public Health, 18, 1967. https://doi.org/10.3390/ijerph18041967

24.

Giorgi

Perez

F. S. F.

D’Antonio

A. C.

Mucci

Ferrero

Cupelli

Arcangeli

(2015). Psychometric properties of the Impact of Event Scale-6 in a sample of victims of bank robbery. Psychology Research and Behavior Management, 8, 99–104.

25.

Hammer

A. L.

Hammer

(1988). Manual for the coping resources inventory. Consulting Psychologist Press.

26.

Hasking

Lewis

S. P.

Bloom

Brausch

Kaess

Robinson

(2021). Impact of the COVID-19 pandemic on students at elevated risk of self-injury: The importance of virtual and online resources. School Psychology International, 2(1), 57–78.

27.

Hay

Meldrum

(2010). Bullying victimization and adolescent self-harm: Testing hypotheses from general strain theory. Journal of Youth and Adolescence, 39(5), 446–459.

28.

Herbert

Bolock

Abdennadher

(2021). How do you feel during the COVID-19 pandemic? A survey using psychological and linguistic self-report measures, and machine learning to investigate mental health, subjective experience, personality, and behaviour during the COVID-19 pandemic among university students. BMC Psychology, 9(90). https://doi.org/10.1186/s40359-021-00574-x

29.

Irwin

R.E.

(2021). Deviant swedes in the Global COVID-19 media environment. Medical Anthropology, 40(1), 815–829. https://doi.org/10.1080/01459740.2021.198859

30.

ISS – Istituto Superiore di Sanità (2021). Epidemia COVID-19 Aggiornamento nazionale 18 agosto 2021 – ore 12:00. Bollettino/Bollettino-sorveglianza-integrata-COVID-. Retrieved August 18, 2021, from https://www.epicentro.iss.it/coronavirus/bollettino/Bollettino-sorveglianza-integrata-COVID-19_18-agosto-2021.pdf

31.

Istituto Giuseppe Toniolo. (2021). La condizione giovanile in Italia. Rapporto giovani 2021. Il Mulino.

32.

Kaufman

J. M.

(2009). Gendered responses to serious strain: The argument for a general strain theory of deviance. Justice Quarterly, 26(3), 410–444.

33.

Labos. (1994). La gioventù negata. Osservatorio sul disagio giovanile [The youth denied. An observatory on juvenile discomfort]. Edizioni TER.

34.

Lazarus

R. S.

Folkman

(1984). Coping and adaptation. In Gentry

W. D.

(Ed.), The handbook of behavioral medicine (pp. 282–325). Guilford.

35.

Leeper Piquero

Sealock

M.D

. (2000). Generalizing general strain theory: An examination of an offending population. Justice Quarterly, 17(3), 449–484.

36.

Liu

Wang

. (2021). Perceived stress of the COVID-19 pandemic and adolescents’ depression symptoms: The moderating role of character strengths. Personality and Individual Differences, 182, 111062. https://doi.org/10.1016/j.paid.2021.111062

37.

Magson

N. R.

Freeman

J. Y. A.

Rapee

R. M.

Richardson

C. E.

Oar

E. L.

Fardouly

(2021). Risk and protective factors for prospective changes in adolescent mental health during the COVID-19 Pandemic. Journal of Youth and Adolescence, 50, (1), 44–57.

38.

Marsella

S. C.

Gratch

(2009). EMA: A process model of appraisal dynamics. Cognitive Systems Research, 10(1), 70–90.

39.

Mazerolle

Maahs

(2000). General strain and delinquency: As alternative examination of conditioning influences. Justice Quarterly, 17(4), 753–778.

40.

Mazza

Ricci

Biondi

Colasanti

Ferracuti

Napoli

Roma

(2020). A nationwide survey of psychological distress among Italian people during the COVID-19 pandemic: Immediate psychological responses and associated Factors. International Journal of Environmental Research and Public Health, 17, 3165. https://doi:10.3390/ijerph17093165

41.

Nocentini

Palladini

Menesini

(2021). Adolescents’ stress reactions in response to COVID-19 Pandemic at the peak of the outbreak in Italy. Clinical Psychological Science, 9(3), 507–514.

42.

Pietropolli Charmet

. (2018), L' insostenibile bisogno di ammirazione. [The 'unbearable need for admiration]. Laterza.

43.

Ritter

L. A.

Sue

V. M.

(2007). Introduction to using online surveys. New Directions for Evaluation, 115, 47–50.

44.

Rogowska

A. M.

Ochnik

Kuśnierz

Chilicka

Jakubiak

Paradowska

Głazowska

Bojarski

Fijołek

Podolak

Tomasiewicz

Nowicka

Kawka

Grabarczyk

Babińska

(2021). Changes in mental health during three waves of the COVID-19 pandemic: A repeated cross-sectional study among Polish university students. BMC Psychiatry, 627(21). https://doi.org/10.1186/s12888-021-03615-2

45.

SAS Institute (2013). SAS/STAT® 13.1. User’s guide.

46.

Schlegl

Meule

Favreau

(2020). Bulimia nervosa in times of the COVID-19 pandemic-Results from an online survey of former inpatients. European Eating Disorders Review, 28(6), 847–854.

47.

Sun

Bao

Meng

Sun

Schumann

Kosten

Strang

Shi

(2020). Brief report: Increased addictive internet and substance use behavior during the COVID-19 pandemic in China. The American Journal on Addictions, 29, 268–270.

48.

Taub

D. E.

McLorg

P. A.

(2011). Eating disorders as deviance. In Clifton

D. B.

(Ed.), The routledge handbook of deviant behavior (pp. 298–305). Routledge.

49.

Thoresen

Tambs

Hussain

Heir

Johansen

V. A.

Bisson

J. I.

(2010), Brief measure of posttraumatic stress reactions: Impact of Event Scale-6. Social Psychiatry and Psychiatric Epidemiology, 45(3), 405–412.

50.

Weiss

D. S.

Marmar

C. R.

(1996). The impact of event scale–revised. In Wilson

Keane

T. M.

(Eds.), Assessing psychological trauma and PTSD (pp. 399–411). Guilford.

51.

Zhang

Zhao

Lamis

D. A.

Kong

Jia

Zhou

(2014). Developing the psychological strain scales (PSS): Reliability, validity, and preliminary hypothesis tests. Social Indicators Research, 115(1), 337–361.