1 Introduction

The COVID-19 (SARS-CoV-2) pandemic declared in March 2020 (World Health Organization, 2020) entailed the establishment of regulations such as social distancing, quarantines, and lockdowns. The literature has shown that pandemic disasters can cause post-traumatic stress disorder (PTSD) and the onset of general stress-related symptoms (Sprang & Silman, 2013). Additionally, a review of the psychological impact of quarantines on the population indicates that most studies reported negative psychological effects such as post-traumatic stress, confusion, and anger (Brooks et al., 2020), as well as symptoms of depression, anxiety, and distress (Harper et al., 2020; Huang & Zhao, 2020).

In the current context of the COVID-19 pandemic, one of the most vulnerable groups is adolescents, who have faced a range of challenging scenarios that increased their vulnerability to emotional, behavioral, and wellbeing problems (Wiguna et al., 2020). Adolescence is a stage with significant changes that involve specific biological and psychosocial factors that promote the cognitive development and life skills of adolescents, such as peer interaction and outdoor activities (World Health Organization, 2018). Therefore, lockdown has had a negative psychological impact on the emotional wellbeing of this population, increasing the already high rate of mental health problems reported before the pandemic (Sadler et al., 2021). A recent systematic review assessed the impact of the pandemic on the mental health of adolescents around the world, showing that adolescents present higher rates of anxiety, depression, and stress (Jones et al., 2021), particularly those with psychiatric or psychological conditions (Mensi et al., 2021).

As mentioned previously, several studies have focused on the consequences of the pandemic for the mental health of adolescents, but little is known about the impact of the pandemic on adolescent wellbeing across countries (Zacher & Rudolph, 2021).

2 Hedonic and Eudaimonic Wellbeing in Adolescents

Wellbeing is a complex construct related to the psychological experience and the optimal functioning of people, as well as one of the most important indicators of adolescent adjustment (Casas, 2016). Scientific evidence has demonstrated that wellbeing is a multidimensional phenomenon that comprises two approaches: a) the hedonic approach, which focuses on happiness and conceives of wellbeing as the pursuit of pleasure and the avoidance of pain; and b) the eudaimonic approach, which is centered on meaning and self-fulfillment, and refers to the extent to which a person is fully functional (Ryan & Deci, 2001). Consequently, hedonic wellbeing expresses cognitive-affective subjective experiences, while eudaimonic wellbeing is related to a positive psychological functioning when facing the challenges of life (Huta & Waterman, 2014). According to Diener’s model (1984), the most relevant and widespread in the field of subjective wellbeing, the construct is divided into two components: an affective one that refers to the experience of positive and negative affect, and a cognitive one that is associated with the overall appraisal people make of their own life (life satisfaction). In recent years, articles about subjective wellbeing during childhood and adolescence have increased significantly, and Diener’s approach has been followed in various countries thanks to the data provided by the Children’s World project (for a review, see Dinisman et al., 2015; Casas, 2019). However, Diener et al. (2018) recently conducted a review in which they addressed the criticism of some authors such as Keyes et al. (2002) concerning the relevance of eudaimonic wellbeing in subjective wellbeing. Currently, there is widespread consensus on the need for specialized scales to capture the different perspectives of wellbeing in both children and adolescents (International Wellbeing Group, 2013).

Authors such as Keyes (2006) suggest that the assessment of subjective wellbeing includes three components: emotional wellbeing, associated with positive and negative feelings; psychological wellbeing, related to optimal individual functioning according to normative psychological standards; and social wellbeing, which refers to optimal social functioning in society. Keyes himself developed the Mental Health Continuum (MHC) scale to measure subjective wellbeing in adults and adolescents based on this definition (Keyes, 2009). To assess the state of wellbeing in adolescents, he specifically proposes the categories of flourishing (high levels of wellbeing), languishing (low levels of wellbeing), and moderate mental health. Subsequently, Keyes (2007) defined mental health (or flourishing) as a combination of high levels of hedonic and eudaimonic wellbeing. Following this perspective, several studies from different countries have assessed wellbeing through the MHC scale, as this is one of the few scales that cover the psychological, emotional, and social dimensions of adolescent wellbeing (for a review, see Rose et al., 2017).

Studies in different cultural contexts have demonstrated that gender and age are related to the emotional, social, and psychological subjective wellbeing of adolescents. Regarding gender, some studies report that male students have higher levels of wellbeing (Didaskalou et al., 2018; Guijarro et al., 2021; Heiman et al., 2018; Moksnes et al., 2021). As for age, some studies indicate that perceptions of wellbeing show a significant decrease as people grow older (Didaskalou et al., 2018). Recently, several studies conducted based on data from the Children’s World project, which assesses subjective wellbeing in adolescents from various countries, have shown that subjective wellbeing decreases progressively during adolescence beginning at 11–12 years of age, and that this is more evident in girls (Casas & González-Carrasco, 2019; González-Carrasco et al., 2017).

In the context of the COVID-19 pandemic, some studies have shown the impact of the social dynamics generated by the pandemic on subjective wellbeing. Specifically, a study examined the changes in the subjective wellbeing of people in Germany between December 2019 and May 2020. The findings revealed that on average, life satisfaction, positive affect, and negative affect did not significantly change between December 2019 and March 2020 but decreased between March and May 2020 (Zacher & Rudolph, 2021). Regarding studies with adolescents, the work by Soest et al. (2020) showed that restrictions due to the pandemic caused a decrease in the subjective wellbeing of adolescents. Another recent study examined the role of resilience in the victimization experiences of students and their subjective wellbeing, exploring the effects related to gender and age. The findings on wellbeing indicated that males from all grades reported higher levels than females, and that students who languished were at higher risk of suffering serious victimization (Andreou et al., 2020).

To summarize, there are studies that have reported on COVID-19’s impact on different measures of subjective wellbeing, but there is a lack of research that allows for characterizing this wellbeing in other countries, considering different wellbeing traditions, as well as age and gender groups.

3 Victimization and Adolescent Wellbeing during the COVID-19 Pandemic

Natural catastrophes and pandemics like COVID-19 could make adolescents more prone to different forms of victimization; for example, domestic violence (Phelps & Sperry, 2020), or an increase in peer violence via social networks, such as cyberbullying (Fraser, 2020).

Studies conducted in recent years have shown some of the effects that different forms of victimization can have on the wellbeing measures of adolescents. For example, bullying and cyberbullying victims show lower levels of subjective wellbeing than non-victims, and victimization is negatively related to wellbeing at these ages (Chai et al., 2020; Miranda et al., 2019). When observing the effects of victimization on adolescent subjective wellbeing in different cultural contexts (Chile and South Africa), bullying is observed to be a significant predictor of subjective wellbeing at the individual and school levels in both cultural contexts (Varela et al., 2020). It should be noted that the negative relationship between bullying and subjective wellbeing is seen across age groups and geographical populations (Savahl et al., 2019). In a wide variety of definitions from bullying studies, this form of peer aggression occurs between a bully and a victim, who are in an asymmetrical relationship in which the bully repeatedly exerts intimidation over the victim (Olweus, 2006). From this definition, several works have characterized bullying as different forms of aggression such as verbal aggression, physical assault, or relational aggression (Vivolo et al., 2011). They have also distinguished between traditional bullying and bullying through electronic devices (Miranda et al., 2019).

There has also been growing interest in observing the effects of the co-occurrence of victimization forms in adolescents (Finkelhor et al., 2007), as many adolescents have been observed to suffer different forms of victimization at the same time (Ford & Delker, 2018). For example, some studies indicate that adolescents who report having suffered cyberbullying also express having been victims of one or more forms of face-to-face bullying (Beran et al., 2015). In a recent study conducted by Wang and et al., (2019), 48.7% of students who were involved in cyberbullying had also experienced traditional bullying. In addition, it should be noted that poly-victimization is more frequent in middle-low-income countries than in high- and upper-middle-income countries because adolescents are subject to more risk factors in the different developmental contexts (Le et al., 2018). The co-occurrence of forms of victimization exacerbates the impact on the subjective wellbeing of adolescents, in that adolescents who have experienced poly-victimization are more prone to lower levels of subjective wellbeing (Oriol et al., 2019). A recent systematic review also showed that poly-victimization is associated with more externalizing and internalizing of problems and total psychological distress (Haahr-Pedersen et al., 2020). Therefore, the use of instruments to assess different forms of victimization simultaneously has been suggested (Skrzypiec et al., 2019).

When examining the literature about the impact of the COVID-19 pandemic on adolescent wellbeing, studies are scarce. However, there is some evidence recently gathered by a review that comprised 22 studies that include high-income, middle-income, and low-income countries, conducted through November 2020. Among the results specifically related to the effects of COVID-19 on wellbeing, countries such as Australia, Spain, and China reported an increase in depression symptoms and a decrease in life satisfaction. Croatia and Bosnia & Herzegovina had a significant reduction in physical activity and a rise in the consumption of unhealthy foods; in addition, reports on child abuse and neglect decreased almost a third due to the closing of schools in Florida (Rajmil et al., 2021). Meanwhile, a study conducted in Spain suggests that the overall wellbeing of children was at an intermediate level during lockdown, and the lowest levels of wellbeing were obtained for physical activity, together with creative and recreational activities (Berasategi Sancho et al., 2021). In turn, the results of the study conducted in Australia revealed that physical distancing and social restrictions had an impact on the health and wellbeing indicators associated with chronic disease among young people (Munasinghe et al., 2020). Regarding Latin America, a study conducted in Brazil concluded that the uncertainty about COVID-19 in a context of socioeconomic vulnerability, along with the increase of unhealthy behaviors and isolation from immediate social circles, has negatively affected the emotional state of adolescents during the COVID-19 pandemic (Szwarcwald et al., 2021). Additionally, a study on adolescents from India, Malaysia, Mexico, and the United Kingdom found that, in general, adolescents have increased their use of internet and social networks. Furthermore, adolescents who scored higher in game addiction, compulsive use of the internet, and social network use also reported high scores on depression, loneliness, aversion, poor sleep quality, and anxiety related to the pandemic, which indicates a significant effect of the pandemic on the use of internet and psychosocial wellbeing of adolescents (Fernandes et al., 2020).

Likewise, the number of studies specifically about domestic violence against children and adolescents during the COVID-19 pandemic is still limited. However, it seems that pandemic contexts are of particular risk for children and adolescents as these times create conditions for violence and abuse against children (Martinkevich et al., 2020; Pereda & Díaz-Faes, 2020). The European Parliament reported that domestic violence cases increased by one-third in some countries of the European Union after the lockdown was declared due to the pandemic (Delaleu, 2020). Similar results were found in the USA, where an increase of domestic violence was observed in 2020 (Boserup et al., 2020). A recent review that comprised 48 studies also produced similar results, with a reported rise in violence against children and adolescents during the COVID-19 pandemic (Cappa & Jijon, 2021). LGBTQ + adolescents also seem to be simultaneously experiencing universal and specific stress factors regarding their identity as a consequence of being confined at home with family members who do not support them, as well as difficulty with accessing support resources (Fish et al., 2020). A field that has received little attention is physical and emotional violence between siblings. A study conducted by Perkins et al. (2021) reveals an increase in this type of violence during the COVID-19 pandemic, which has not received the attention required. It should be noted that in Latin America and the Caribbean, the precarious situation has reduced the opportunities to develop protections during the pandemic and this suggests that the rise in violence against children and adolescents is even higher (CEPAL, 2020; OMS, 2020).

To summarize, there has been an increase in studies on the relationship between victimization and subjective wellbeing in adolescents over recent years. However, since the COVID-19 pandemic seems to be a risk factor for adolescent victimization, studies that allow for understanding these relationships during the pandemic are increasingly necessary.

4 Contextual Factors of Chile and Mexico during the COVID-19 Pandemic

In children and adolescents, indicators like the GDP of a country have a relatively small influence on the perception of wellbeing, as opposed to other factors such as relationships with family and peers (for a review, see Newland et al., 2019). However, in periods of crisis, like the current pandemic, changes in the life circumstances of adolescents due to lockdowns and social distancing can cause more fluctuations in adolescent wellbeing (Kapetanovic et al., 2021).

In Latin America and the Caribbean, the impact of COVID-19 on children and adolescents has been particularly hard, with an estimated increase of 7.6% in poverty among children and adolescents of the region as a consequence of the pandemic (ECLAC-UNICEF, 2020). As mentioned previously, some studies warn about the risk of increase in violence against children and adolescents under lockdown (Sadler et al., 2021; Pereda & Díaz-Faes, 2020). This may be a key factor in Latin American and Caribbean countries, where high levels of violence against children and adolescents are reported more frequently than in other regions (for a review, see Garmendia-Lorena, 2011). Therefore, the probability of adolescents experiencing some form of violence increases dramatically during pandemics like COVID-19 (Pereda & Díaz-Faes, 2020). Chile, for example, is one of the countries with the highest GDP per capita in the region, but remains one of the countries with the most unequal income distribution (World Bank, 2018), which has affected the containment of the pandemic. For instance, the pandemic has had differing impacts across Chilean communities depending on their socioeconomic situation, which has led to long lockdowns and delays in the return of children and adolescents to school (Asahi et al., 2021). In a recent study by Pinto-Cortez et al. (2020) with a representative sample of adolescents in Chile, a high prevalence of poly-victimization was observed, particularly in boys. The adolescents exhibited a higher prevalence of different forms of victimization (poly-victimization) than in Europe and the USA (Pinto-Cortez et al., 2018).

Mexico is one of the countries in Latin America and the Caribbean with high levels of structural violence in its society (Charles-Leija et al., 2019). A high prevalence of violence directed at women from men is often observed due to the machismo ingrained in the different spheres of Mexican society (Moral & Ramos, 2016). However, in addition to the high rate of violence against women (Romero Mendoza et al., 2021), there is also a high prevalence of poly-victimization in the adolescent population (Méndez-López et al., 2021). According to the Mexican Survey on Household Dynamics (ENDIREH, 2016), 26% of women aged 15 to 28 report having suffered some form of abuse (physical, psychological, or sexual violence) during childhood.

5 Present Study

The COVID-19 pandemic that hit most countries in the world has affected the wellbeing of adolescents. Several studies have reported on the effects that lockdowns and social distancing may be having on adolescent wellbeing (Soest et al., 2020; Zacher & Rudolph, 2021). Research has also warned about an increase in exposure to violence during adolescence and its repercussions on the mental health and wellbeing of teenagers (Pereda and Díaz-Faes, 2020; Soest et al., 2020). In addition, exposure to violence seems to be much more common in middle-low-income countries, where the virus has struck more violently (Cappa & Jijon, 2021).

This study aims to contribute to the reduction of the gap in the literature by characterizing wellbeing during the COVID-19 pandemic in adolescents from Mexico and Chile, delving into the relationship between exposure to violence and hedonic and eudaimonic wellbeing. The study has the following objectives: 1) To characterize the subjective wellbeing of adolescents from Mexico and Chile during the COVID-19 pandemic by adolescence stage and gender. Late adolescents are expected to exhibit lower levels of flourishing, and boys to score higher on this indicator. 2) To observe the prevalence of exposure to violence in Mexico and Chile considering adolescent stage and gender, with a high prevalence expected for both countries. 3) To observe the effect of being or not exposure to violence over the wellbeing categories of flourishing, languishing, and moderate mental health, taking into consideration the age groups in both countries. Non being exposure to violence are expected to have a higher probability of being in the flourishing group than victims. 4) To observe the effect of being or not exposure to violence on hedonic and eudaimonic wellbeing in both countries by age group. Not being exposure to violence group are expected to present higher levels of both hedonic and eudaimonic wellbeing.

6 Methodology

6.1 Participants

A total of 3,275 children and adolescents from Chile and Mexico responded to the survey, of which 1,186 were from Mexico (36.2%) and 2,089 from Chile (63.8%). Regarding gender, 43.4% were males (n = 409), 56.2% were females (n = 530), and 0.4% were non-binary (n = 4) in the Mexican sample, while males accounted for 43.5% (n = 721), females for 54.9% (n = 910), and non-binary for 1.6% (n = 26) of the Chilean sample. The non-binary category refers to people who identify as a gender different from male or female.

The age interval of the sample was 10 to 18 years. Analyses were performed based on the age groups proposed by the World Health Organization (2018), namely early adolescence (10 to 14 years) and late adolescence (15 to 19 years). In the case of Mexico, the mean age of participants was 15.27 with a standard deviation of 1.66; while for Chile, mean age was 13.62 and standard deviation 1.67.

6.2 Data Collection

This study is part of a wider research protocol, an international collaboration that originated from the Global Research Alliance in the context of the COVID-19 pandemic that aims to study wellbeing and victimization in adolescents. Questionnaires were translated from Spanish to English and then retranslated from English into the local language. With respect to Chile and Mexico, data was gathered from June to September 2020 during the COVID-19 lockdown. The research plan was first approved by the Ethics Committee of Universidad de Flinders and Universidad de La Frontera (Chile). Then, it was presented to the respective public education institutions (representatives of the Ministry of Education) in each country for their approval before reaching out to schools and introducing the project to the families of students. Due to the pandemic, invitations were issued by e-mail and through the social network accounts associated with the schools and public education institutions. The invitation e-mail included the link to an electronic platform that contained: (a) the informed consent for parents/guardians of students, and (b) the informed assent form for students. Students completed the survey anonymously and under strict confidentiality. Estimated response time was 15 to 20 min.

6.3 Measures

The instruments were administered during the lockdown period from June to September 2020, when online education had been adopted by both countries and therefore students were at home with their families.

Exposure to Violence during COVID-19

The instrument was adapted from the Student Aggression and Victimization Questionnaire (Skrzypiec, 2015), which included 8 types of exposure to violence (for example, "other person spread rumors (fake stories) about me", “I was beaten, kicked or pushed”, and “someone was mean to me”). Since this project was carried out during lockdown, the instrument was adapted to inquire about forms of violence suffered during that period without considering the source of violence. Students from both countries were confined to their homes with their families, but some of them continued to have contact with classmate through online lessons and social networks. Therefore, the variable was denominated exposure to violence suffered during lockdown. Each item was measured through a dichotomous answer (“yes” or “no”). Likewise, to assess the severity of violence exposure, a global indicator was calculated, with 0 if none of the 8 types of violence had been exposure and 1 if at least one type of violence had been exposure during the pandemic.

Student Wellbeing

The subjective wellbeing of students was assessed by Keyes's (2009) Mental Health Continuum. Students were asked to answer 14 questions on a 6-point Likert scale that ranged from “never” to “every day”. This questionnaire is divided into 3 subdimensions: 3 items represent emotional wellbeing (happy, interested in life, and satisfied), six items represent social wellbeing (That you had something important to contribute to society; That you belonged to a community [like a social group, your school, or your neighborhood]; That our society is a good place, or is becoming a better place, for all people; That people are basically [generally] good; That the way our society works made sense to you), and five items assess psychological wellbeing (That you liked most parts of your personality; That you are good at managing the responsibilities of your daily life; That you had warm and trusting relationships with others; That you had experiences that challenged you to grow and become a better person; Confident to think or express your own ideas and opinions; That your life has a sense of direction or meaning to it). Only emotional wellbeing was considered for the hedonic wellbeing indicator, while for eudaimonic wellbeing both psychological and social wellbeing were used. For Mexico, the CFA adjustment indicators were a CFI of 0.98, a TLI of 0.97, and a RMSEA of 0.06; the reliability indicator values were Ω = 0.87 and α = 0.87. For Chile the indicators were a CFI of 0.99, a TLI of 0.98, and a RMSEA of 0.05, and the reliability indicator values were Ω = 0.88 and α = 0.88.

Likewise, this scale classifies mental health into three groups: flourishing, languishing, and moderate. People in the flourishing group experienced at least one of the three signs of hedonic wellbeing and at least six of the eleven signs of positive functioning ‘every day’ or ‘almost every day’ during the pandemic. Individuals with languishing mental health exhibited low levels (i.e., ‘never’ or ‘once or twice’ during the past month) of at least one measure of hedonic wellbeing and low levels of at least six measures of positive functioning. Those who were neither flourishing nor languishing were classified into the moderate mental health group.

6.4 Statistical Analysis

The SPSS 22.0 software was employed to calculate the descriptive statistics and the ordinal and linear multiple regression, while AMOS 22.0 was used for running the confirmatory factor analysis. For the latter, the robust maximum likelihood estimator was calculated. Likewise, the Tucker-Lewis index (TLI) and the comparative fit index (CFI) were computed, which should be above 0.90, as well as the root mean square error of approximation (RMSEA), which should be lower than 0.08, as suggested by Marsh et al. (2004) and by Hu and Bentler (1998). Scale reliability was measured through Cronbach’s alpha (α) and McDonald’s omega (Ω), and their performance was assessed based on Kline’s criteria (2013).

A series of logistic ordinal regressions were carried out in order to measure the effect of exposure to violence at the global level, i.e., after having suffered any form of violence during the pandemic, and by reported violence type over the resulting wellbeing category. Odds ratios and p values were calculated for all associations, using a lack of violent experiences as a basis to observe how each violence exposure form affects the wellbeing groups, with gender and age as control variables.

Likewise, a series of multiple regressions were calculated taking the hedonic and eudaimonic wellbeing indicators as dependent variables and the different forms of violence reported as independent variables. For each proposed regression, gender and age were considered dependent variables. The independence assumptions for observations over the linear relationships between the dependent variable and independent variables, homoscedasticity, multicollinearity, and normal distribution of residuals were tested prior to analysis.

7 Results

7.1 Descriptive Analysis

Table 1 presents the descriptive results by country, gender, and adolescent age group. Results show that almost all respondents spent lockdown at home with their families. Access to the internet was above 90% in Chile for all groups, whereas in Mexico, reported access was 75% for the non-binary group and 83.7% for the early adolescence group.

Table 1 Descriptive analysis of the sample
Full size table

Finally, regarding the question about whether they had contact with any friends during lockdown, it is observed that for Chilean students, males (12.5%) and late adolescents (11.3%) represent the groups that had the least contact with friends. In the case of Mexico, the groups that reported less contact with friends were males (12%) and late adolescents (10.3%) as well.

As for the wellbeing indicators by country, Table 2 shows that in the case of Chile, 49.5% of males were in the flourishing category, while the non-binary group presented a higher percentage in the languishing category (50%). Considering age groups, 50% of early adolescents were classified as flourishing, whereas only 33.6% of late adolescents belonged to this group. This difference in wellbeing is also reflected in the languishing category; 12% of the early adolescence group fell into the category, and this increased to 18.9% for the late adolescence group. In Mexico, the flourishing category also has a higher percentage of males (53.1%). In turn, languishing exhibits a higher percentage (43.3%) in the non-binary group, followed by females (13.6%) and males (8.9%).

Table 2 Comparison of wellbeing groups and violence by gender and age group
Full size table

Regarding age groups, 52.9% of the early adolescence group is classified as flourishing, while the late adolescence group is at 43.6%. Likewise, the late adolescence group has a higher percentage of individuals (12.8%) in the languishing category than the early adolescence group (11.3%).

As for the prevalence of violence exposure in the Chilean sample, the non-binary group reports the highest prevalence (68.4%), followed by females (52%) and males (44.7%). In other words, all groups reported having exposure at least one form of violence during lockdown. Regarding age groups, late adolescents report the highest percentage of violence exposure experienced (52.9%).

When comparing all forms of violence exposure, the group that presents the highest prevalence is the non-binary group for all forms. Contrasting males and females, prevalence is higher in females for all forms of violence exposure, except for threats. Likewise, considering form violence exposure, suffering ‘teasing’, being ‘left out’ and ‘called names’ are the most commonly reported types of violence. Regarding age groups, prevalence in the late adolescence group (52.9%) was higher than in the early adolescence group (47.9%). Comparing the forms of violence exposure reported, prevalence in the late adolescence group is higher for all forms of violence, except for ‘hit’, which had a higher prevalence (3.4%) in the early adolescence group.

In the case of Mexico, the female group presented the highest prevalence of violence exposure considering all the forms of violence reported. Of this group, 42.5% report having exposure at least one form of violence, compared to the male group, at 34%. Regarding age groups, both early and late adolescence exhibit similar percentages. Additionally, when analyzing all types of violence exposure, ‘rumors’ and ‘teased’ are observed to account for 8% and present the highest prevalence among both males and females.

Regarding age groups, ‘teased’, ‘called names’ and ‘rumors’ have high prevalence in both early and late adolescence, with the late adolescence group having the highest prevalence.

7.2 Ordinal Logistic Regression

A series of ordinal regressions was calculated to assess the effect of violence exposure at the global level and for each form of violence exposure, taking the reported wellbeing category as a result variable. In addition, age and gender were considered control variables, while each form of violence exposure was taken as a base variable for the ordinal regression variables.

As observed in Table 3, in the case of Mexico, most forms of violence exposure reported have a significant effect on the wellbeing group for early adolescents. Although there is no effect at the general level, it is observed that not being exposure to ‘teased’ increases by 5.10 times the probability of being in the flourishing group (OR = 5.10, p < 0.001). Likewise, OR values are significant for the forms of violence exposure ‘picked-on’ (OR = 2.97, p = 0.02), ‘called names’ (OR = 6.05, p < 0.001), ‘threatened’ (OR = 7.77, p = 0.02), and ‘mean’ (OR = 3.78, p = 0.01).

Table 3 Ordinal regression analysis of forms of violence exposure over wellbeing categories
Full size table

Likewise, for late adolescence, a significant odds ratio is observed for prevalence in general. In other words, not experiencing any form of violence exposure increases by 1.70 times the probability of being in the flourishing category compared to not suffering any form of violence exposure (OR = 1.70, p = 0.09). When analyzing the results by form of violence exposure, not experiencing being ‘called names’ (OR = 1.68, p = 0.04), ‘left out’ (OR = 2.59, p < 0.001), ‘rumors’ (OR = 1.84, p < 0.001), or ‘mean’ treatment (OR = 2.51, p < 0.001) significantly increases the probability of being in the flourishing group compared with suffering these types of violence.

In the case of Chile, the odds ratio for the early adolescence group is significant for all types of violence exposure separately but not for the global indicator (OR = 1.30, p = 0.28). Comparing the OR values, higher values are observed for ‘mean’ in forms of violence exposure; in this case, not experiencing violence exposure increases by 4.66 times the possibility of being in the flourishing wellbeing category compared with having suffered this form of violence. ‘Mean’ is followed by ‘threatened’ (OR = 3.39, p < 0.001), ‘hit’ (OR = 3.29, p < 0.001), ‘called names’ (OR = 3.25, p < 0.001), ‘left out’ (OR = 2.80, p < 0.001), ‘teased’ (OR = 2.48, p < 0.001), ‘rumors’ (OR = 2.25, p < 0.001), and ‘picked-on’ (OR = 2.48, p < 0.001).

Finally, at the general level in the late adolescence group, not having experienced any type of violence exposure increases by 2.08 times the probability of being in the flourishing wellbeing group compared with those who suffered some type of violence exposure. When analyzing each type of violence exposure, ‘called names’ presents an OR of 2.23 (p = 0.02), ‘left out’ has an OR = 1.95 with p = 0.01, and ‘mean’ presents an OR = 3.49 with p < 0.001.

7.3 Multiple Linear Regression

Table 4 presents the results of a multiple linear regression, considering the hedonic and eudaimonic wellbeing indicators as dependent variables for both the early and late adolescence group. Likewise, gender and age of participants were used as control variables. As shown in Table 3, the indicator of violence exposure at the general level and each form of violence exposure mentioned previously were considered independent variables.

Table 4 Multiple linear regression taking reported violence over wellbeing indicators
Full size table

In the early adolescence group from Mexico, ‘threatened’ (B = -1.37, p = 0.01) has an effect on the hedonic coefficient of wellbeing. In other words, suffering this form of violence exposure reduces the hedonic wellbeing indicator by 1.37 points, and this effect is stronger than with other forms of violence exposure. The only forms of violence exposure that did not prove to be significant were ‘hit’ (B = -0.1, p = 0.75) and ‘rumors’ (B = -0.42, p = 0.10).

Experiencing any form of violence exposure has no significant effect on eudaimonic wellbeing as a dependent variable (B = -0.28, p = 0.29). Except for ‘rumors’ (B = -0.33, p = 0.2), all types of violence exposure proved significant, with ‘threatened’ having the highest regression coefficient (B = -1.03, p = 0.04). In other words, having experienced this form of violence exposure reduces the score of the eudaimonic wellbeing indicator by 1.03.

Regarding the late adolescent group, the indicators of violence exposure significantly affected the hedonic wellbeing indicator, with the form of violence ‘threatened’ (B = -1.45, p < 0.001) having the highest coefficient. Taking the eudaimonic component of wellbeing as a dependent variable, there was significant effect (B = -0.26, p = 0.14) at the global level. However, considering each form of violence exposure, only being exposure to ‘hit’ does not result in a significant regression coefficient (B = -0.35, p = 0.24). In turn, consistent with the previous results, being ‘threatened’ represents the form of violence that has the strongest effect on the eudaimonic component of wellbeing (B = -1.03, p < 0.001).

The results for Chile show that all violence exposure indicators have a significant effect on hedonic wellbeing for the early adolescence group. Specifically, a B = -0.31, p = 0.05 is reported at the global level; that is to say, suffering at least one form of violence exposure reduces the hedonic wellbeing indicator by 0.31. Likewise, for each form of violence exposure, the resulting coefficients are significant, with ‘threatened’ (B = -1.23, p < 0.001), ‘hit’ (B = -1.20, p < 0.001), and ‘mean’ (B = -1.02, p < 0.001) having the greatest effect on hedonic wellbeing.

In turn, taking eudaimonic wellbeing as a dependent variable, the exposure to violence indicator at the global level does not present a significant effect on the eudaimonic wellbeing indicator (B = -0.14, p = 0.32). Likewise, considering the forms of violence, only being victim of ‘rumors’ is not significant (B = -0.22, p = 0.12), while ‘hit’ (B = -0.81, p < 0.001), ‘mean’ (B = -0.79, p < 0.001), and ‘left out’ (B = -0.79, p < 0.001) are the forms of violence exposure with the strongest effect over this type of wellbeing.

For the late adolescence group at the global level, the regression coefficient is significant (B = -0.71, p = 0.02); that is to say, the hedonic wellbeing indicator decreases by 0.71 points when suffering any violence exposure. For each form of violence exposure, only being ‘threatened’ (B = -0.34, p = 0.11) and a victim of ‘rumors’ (B = -0.34, p = 0.95) were not significant. On the other hand, being ‘hit’ (B = -1.34, p < 0.001) or the victim of ‘mean’ treatment (B = -1.08, p < 0.001) are the two forms of violence exposure that most affect reported hedonic wellbeing.

Finally, considering eudaimonic wellbeing as a dependent variable, as in the case of hedonic wellbeing, there is a significant effect of suffering any form of violence exposure (B = -0.67, p < 0.001) at the global level. Regarding each form of violence, only being ‘picked-on’ (B = -0.24, p = 0.22) and ‘threatened’ (B = -0.95, p = 0.60) had no significant effect over eudaimonic wellbeing, compared to being ‘hit’ (B = -0.72, p = 0.02) and the victim of ‘mean’ treatment (B = -0.79, p < 0.001), which have the strongest impact on eudaimonic wellbeing.

8 Discussion

The descriptive data of the study reveals the particular situation experienced by adolescents during lockdown in both countries. First, the results show that almost all adolescents spent lockdown with their families. In addition to this, adolescents had little contact with their friends during this period in both Chile and Mexico, particularly in the case of late adolescents, who reported spending less time with their peers. These data warn of the risk factors that lockdown situations such as the ones associated with the COVID-19 pandemic may create for the adolescent population. Social identity develops during adolescence as a product of interaction with the peer group, and therefore lack of contact with peers may have serious repercussions on the mental health of adolescents (Blakemore, 2019). Social isolation and new forms of interpersonal relationships that take place within the family due to lockdown are some of the risk factors that may have the greatest influence on the increase of stress and presence of violence, especially in vulnerable families (Pereda and Díaz-Faes, 2020).

The first objective of this study aimed to characterize the types of wellbeing in adolescents from Mexico and Chile, considering the mental health categories proposed by Keyes (2009). According to this classification, flourishing indicates adequate adolescent mental health and therefore implies the presence of wellbeing. The results show that during the pandemic, late adolescents are the age group that presents the lowest scores in flourishing. This is in agreement with previous studies indicating that as adolescence progresses, the perception of wellbeing decreases (Andreou et al., 2020; Didaskalou et al., 2018; Guijarro et al., 2021; Heiman et al., 2018; Moksnes et al., 2021); likewise, data confirm that males score higher, as reported by the literature (Casas & González-Carrasco, 2019; González-Carrasco et al., 2017). Following this perspective, girls were observed to have higher scores on languishing than boys in both countries; a higher prevalence of languishing is also observed in late adolescents. Finally, the high presence of languishing (around 50%) in the non-binary group from Chile is noteworthy. These data warn about the strong impact the pandemic may have had in groups that are usually less accepted socially and therefore susceptible to increased risk of mental health problems (Lassiter et al., 2020). Languishing does not imply mental illness but, according to Keyes (2009), it may generate feelings of stagnation and emptiness that can have serious repercussions on mental health in the medium and long term.

With respect to the second objective of this study, a high prevalence of exposure to violence was expected both in Mexico and Chile due to the risk factors associated with lockdown in both countries. The results show a high prevalence of poly-victimization, with the highest prevalence reported by girls and late adolescents; however, the early adolescence group also shows a very high prevalence. When observing general data, almost half of boys and more than half of girls have experienced some form of violence in Chile. It should also be noted that in the non-binary group, 68.4% of adolescents report having suffered some form of victimization. Data indicate that the prevalence in Mexico is slightly lower, but the numbers are still high.

Data from both countries confirm that girls are more prone to experiencing violence due to the patriarchy, which is observed in many countries in Latin America and the Caribbean (Garmendia-Lorena, 2011). The patriarchy also explains why the non-binary group presents a high prevalence of exposure to violence in the case of Chile but is practically nonexistent in Mexico, where gender diversity is socially condemned (Baruch-Dominguez et al., 2016). Data on violence exposure from Chile during the pandemic, in particular, confirm the high presence of adolescent poly-victimization (Pinto-Cortez et al., 2020). This confirms the importance of developing programs and strategies to tackle the social-emotional effects of the pandemic on this population. In addition to launching comprehensive policies for mitigating violence against adolescents at the national level, initiatives for promoting wellbeing and healthy coexistence that focus on respect for diversity should be launched.

Similar data are observed in Mexico, with a high prevalence of adolescent poly-victimization that affects girls to a greater extent. It should be noted that recent data indicates that almost 80% of the Mexican adolescents sample reported at least one victimization experience over the last year, and that girls experienced a wider spectrum of victimization than boys (Méndez-López & Pereda, 2019). As in the Chilean case, this implies that in countries where there is often a high prevalence of adolescent violence exposure, lockdown and social distancing measures may become a major risk factor for mental health. When broadening the discussion on victimization types, it is noteworthy that the forms of violence that affect women the most are relational and psychological. The Chilean adolescents in the sample are mainly affected by the categories of ‘left out’, ‘called names’, ‘mean’, and ‘teased’, while ‘rumors’, ‘teased’, and ‘left out’ were predominant in Mexican adolescents. Relational or passive victimization is aimed at damaging peer relationships and include exclusion, withdrawal of friendship, and gossiping (Pontillo et al., 2019), representing a high risk factor for the wellbeing of adolescents, which is in line with other studies that identified gender differences, with female students reporting more challenges for their wellbeing (Didaskalou et al., 2018). The types of violence that affected men to a larger extent are also relational and psychological, with similar percentages for both countries. In the case of Chilean adolescents, ‘called names’ and ‘left out’ were the most reported forms of violence, while for Mexican adolescents these were ‘rumors’ and ‘teased’. As mentioned previously, these data are especially relevant if it is considered that adolescents present significantly higher levels of depressive and anxious symptoms when they are exposed to or have participated in forms of violence (Garthe et al., 2021).

Based on this, the third objective of this study sought to delve into the effect of exposure to violence on Keyes’ wellbeing categories (2009). As expected, non-victims have a higher probability of being in the flourishing group compared to victims. These data are in line with previous studies in which adverse experiences were observed to have a negative effect on the flourishing of children and adolescents (Witten et al., 2019). Both countries show that non-victim adolescents have a higher probability of belonging to the flourishing category and perceiving higher levels of wellbeing.

The fourth objective of this study was to analyze the effect of this exposure to violence on Keyes’ well-being categories (2009) while considering the hedonic and eudaimonic types of wellbeing. From this perspective—that is to say, how the two types of wellbeing affect victimization—the global exposure to violence indicator does not seem related to hedonic wellbeing in the early adolescence group from Mexico. However, ‘threatened’ is the victimization form that decreases hedonic wellbeing the most compared to non-victims. A significant negative relationship is found between poly-victimization and eudaimonic wellbeing. For late adolescents, poly-victimization is also significant and negatively related to hedonic and eudaimonic wellbeing.

In Chile, for the early adolescence group, all violence exposure indicators have a significant effect on the hedonic wellbeing indicator. Conversely, no significant effect is observed on the global eudaimonic wellbeing indicator unless violence forms are analyzed separately, with ‘threatened’ having the strongest effect. For the late adolescence group, the poly-victimization global indicator is related to lower levels of hedonic wellbeing. Similar effects are observed on eudaimonic wellbeing.

The results thus confirm that poly-victimization has an effect on the eudaimonic wellbeing of early adolescents in both countries, and on both wellbeing types in late adolescents. These data are important because the influence of forms of victimization on subjective wellbeing is commonly studied, but few works have distinguished between hedonic and eudaimonic wellbeing when dealing with victimization. In turn, the results show that the wellbeing perception of victims of violence decreases as observed in previous studies. For example, victimization by intimidation is significantly associated with lower levels of wellbeing (Savahl et al., 2019), suffering school bullying has an indirect effect on life satisfaction through school satisfaction (Varela et al., 2018), aggression and victimization are negatively correlated with wellbeing indicators (Alcantara et al., 2017), suffering different forms of victimization worsens adolescent wellbeing (Oriol et al., 2019), and victimization during adolescence is a significant risk factor not only for the onset of depression but also for deficient wellbeing as an adult (Armitage et al., 2021).

9 Limitations and Future Directions

This study presents some limitations. The first one is related to the assessment of victimization. Studies on poly-victimization usually differentiate between forms of victimization according to context and aggressor. However, in this case, due to the exceptionality of the pandemic, adolescents from both countries have spent most their time at home and therefore the results reflect whether they have suffered some form of victimization but not from whom. Therefore, the dichotomous indicator was used, inquiring whether they had suffered victimization or not, but there are no exact data on who perpetrated these forms of victimization. Furthermore, for attributing causality to the effect of being a victim on the different types of wellbeing, a longitudinal study with several measurements during lockdown would have been necessary, although the circumstances for collecting this type of sample are very complex. Further studies should compare these results considering the socioeconomic status of the families from the population studied. In this case, this condition could not be determined. It would also be interesting to assess the effects of victimization through other instruments that measure subjective wellbeing in adolescents to enhance the robustness of the results.

9.1 Practical Implications

The COVID-19 pandemic has caused social isolation and lack of physical contact. These factors already imply risk due to the relevance of peer contact and perceived social support to the mental health and wellbeing of adolescents. The results of this study emphasize that, although most adolescents from the samples of both countries have experienced flourishing, some are in groups with moderate mental health and languishing. This is especially noticeable in the late adolescence group of both countries. Furthermore, the high indexes of languishing in non-binary adolescents warn of the need to direct initiatives to more vulnerable adolescents who are prone to suffering different forms of social exclusion.

In Chile and Mexico, there is a high presence of poly-victimization during the pandemic, which can worsen the effects of the pandemic during a developmental stage in which adolescents are susceptible to mental health problems. The results point to the effects of experiencing victimization on hedonic and eudaimonic wellbeing, particularly in late adolescents from both countries. An effect is observed on the eudaimonic wellbeing of early adolescents, but only some forms of victimization affect hedonic wellbeing.

Both countries, as other middle-income Latin-American and Caribbean countries, need to manage the resources allocated for adolescent mental health in a suitable manner. Data points to the importance of implementing policies for offering social-emotional support to adolescents and mitigating the effects of the pandemic, as well as launching programs for the prevention of the high poly-victimization observed. In addition, it is necessary to make advances in strategies for coexistence based on respect for diversity and the generation of high levels of wellbeing in people, as there is robust evidence for their benefits for mental health at the individual and societal levels (Chida & Steptoe, 2008; Diener et al., 2010; Wood & Joseph, 2010).