According to the conservation of resource theory, stressful situations could lead to difficulties with a person’s physical or MH (Mao et al., 2021). The COVID-19 pandemic’s primary characteristics as a novel infectious illness are that it is extremely contagious and dangerous, developing quickly, and has no effective medications for either prevention or therapy. Students’ MH during the pandemic outbreak was impacted by a number of stressful pandemic-related factors, such as personal daily routine disruptions and individual physical health under threat from the COVID-19 sickness.

Mental illness is a result of social exclusion and ongoing worry about contagion. The COVID-19 situation has had significant negative effects on people’s health (Pfefferbaum and North, 2020). However, researchers are concentrating more on how this quickly developing worldwide catastrophe may affect the MH of the aging population. The psychological impact of COVID-19 on teenage MH has gotten very little consideration (Rawat and Sehrawat, 2021). The psychological strain of COVID-19 among teenagers should be a prominent aspect in the COVID-19 study due to the severe and long-lasting consequences of MH, which lead to poor physical health outcomes, such as the rise of cardiovascular disorders, and poor MH outcomes. Teenagers struggle more than adults do with the mental burden of this influence since they lack the adult’s coping mechanisms and physiological development (Rawat and Sehrawat, 2021). Teenagers who already have MH problems face even more MH issues during times of crisis (Gavin et al., 2020).

The effects of COVID-19 on family well-being include a quarantine that restricts movement, financial difficulty, decreased income, a lack of job, obesity, and worsening non-communicable diseases. Nature and range of risks that have emerged after COVID-19 have expanded and increased. The risks involve physical as well as psychological aspects such as the risk of hospitalization or even losing one’s life, failing to access food, vulnerability related to family, and psychological health problems. Furthermore, the chances of maltreatment of children also increased in the wake of social isolation (Rapoport et al., 2021). A survey revealed increased instances of child treatment during school breaks, summer, and amid natural disasters (Rapoport et al., 2021). The only seemingly beneficial aspect of COVID-19 might be the benefit of having more family time that one could spend as the restrictions on the movement either slowed down or stopped the mobility of people confining them to their residences.

Additionally, family interactions have also been severely affected by COVID-19(Pan, 2020). In difficult times like COVID-19, a family’s ability to manage difficulties and deal with risks, stress, and crises is referred to as family support. The capacity to not only survive but also recover from the negative impact of any troublesome situation can be referred to as FS. Some key factors in developing this capacity of FS are adaptation, acceptance, and management in the wake of untoward incidents (Shin et al., 2021). Parents are essential in building the family’s resilience during a widespread public health crisis. Children’s and teenagers’ resilience is influenced by parents’ resilience, including how well they look after their extended family members while still being able to take care of themselves.

Children can adapt well to the epidemic if their parents do so in a positive way. It’s critical to have parental assistance available to lessen the effects of COVID-19. Support from family members during the epidemic may lessen concerns about depression and anxiety (Yang et al., 2020). According to studies, a large number of people experience tension and worry while there are extended lockdowns because of the pandemic (Budimir et al., 2021). The level of concern in people regarding contracting COVID-19 themselves is lower than this apprehension for their family members (Gavin et al., 2020). In addition, elderly people tend to experience less anxiousness.

Anxiety lessens with an increase in age. Core reasons lie in the fact that coping mechanisms become more and more effective with an advance in age, moreover, better financial stability also plays its part (Drentea, 2000). Adjustment, social integration, greater communication, and proper financial management are significant FS factors. Maintaining family resilience requires flexibility, which is the ability to pay attention to changing circumstances and adjust one’s approach to meet those demands.

Due to school closures caused by the COVID-19 pandemic, students now take their classes online. Physical activity, travel, and other logistical stresses may be lessened as a result of home learning. On the other hand, stress can also result from home education. Some students may find home learning to be very difficult owing to scholastic issues, limited internet connection, and insufficient resources (Soetisna et al., 2021). During the pandemic, families might lack the tools and information necessary to manage at-home education. In addition, parents struggle to raise their children while juggling domestic chores, working remotely in the wake of economic insecurity, and fear of loss of work. Parents struggle to achieve a work-life balance and contend with several shifting circumstances while engaging in physical distance, which increases stress (Dinh, 2020).

Families experience distress when schools are closed. To boost children’s and parents’ readiness to accept homeschooling during the epidemic and ensure that kids get good grades, families have to be better educated on how to assist one another. Households with special needs children have to cope with additional stress during a pandemic. Children with special needs spend all of their time living with family members as a result of the closure of the school. Children have little time for social activities outside of their homes during the pandemic. The need for an easy environment at home becomes even more important and pertinent particularly for the family which has special needs child. In order to create a setting that fosters family resilience, parents must learn to control their stress and obtain social support. The pandemic has drastically changed the family routines that affect both physical as well as MH of family members (Ameis et al., 2020). The family can deal with the pandemic together, though, by creating routines and engaging in activities that involve all members, such as scheduling mealtimes together, participating in household chores together, fostering effective communication, engaging in thrilling activities, getting regular physical workouts, being vigilant regarding cleanliness, and developing good sleep habits, which is crucial, particularly in the wake of COVID-19.

These habits may enhance the family’s capacity to care for each other and provide better support to one another. Families are better able to help one another when there are protective factors (which foster the capacity for adaptation or recovery in times of crisis) and recovery variables (which assist in development growth; Semenov and Zelazo, 2019). These factors may make the family more competent to deal with the crisis and struggle as a unit. The main factors in FS are a wholesome amalgamation of useful and healthy family activities that help a family create and maintain a strong trust-based bond enabling all family members to sustain during adverse circumstances (Semenov and Zelazo, 2019). From all the above research findings about FS, MH, and coping strategy it is assumed that FS has a positive relationship with coping strategy and MH.

How people handle stress has a significant psychological impact on how stressful life events like COVID-19 affect people’s MH. The broad definition of coping is the mental and behavioral strategies people use to control their stress (Obbarius et al., 2021). When faced with challenging circumstances, like the pandemic, it is typical to use coping mechanisms or behaviors that encourage adjustments and solutions to the problems being faced. People employ a variety of coping mechanisms to deal with difficult situations or times (Budimir et al., 2021). The two types of coping models, according to Folkman and Lazarus, are the problem-and emotion-focused models (Obbarius et al., 2021). When faced with a stressful scenario, both coping mechanisms are employed, however, their efficacy varies. Self-distraction, stress management, exclusion, medication, moral support usage, keeping informed, and behavioral adjustments are just a few of the many coping mechanisms that have been proven extremely effective. However, the question of whether one strategy is better than another is hotly contested. More successful than other methods for reducing the effects of stress and promoting MH can be those that include addressing and resolving stresses (Nshimyiryo et al., 2021).

Avoidant tactics may be effective in lowering short-term stress, but they are typically viewed as damaging in the long run since they do not directly address the stressor, exposing people to high levels of stress over extended periods. A person’s stress response may have an impact on their health over time. It is therefore likely that coping with COVID-19-related trauma inappropriately may lead to MH issues in the future. To improve family wellbeing, it’s critical to establish a regular appreciation practice. To foster a sense of community, loyalty, harmony, and happiness among family members, it is critical that the members maintain open lines of interaction and find enjoyable activities to do regularly.

Additionally, when parents or other family members ask for assistance, forming positive reinforcement with them is crucial. Examples of these people include siblings, acquaintances, extended families, and colleagues. Technology use for support networks and interaction amid social isolation is crucial (SayinKasar and Karaman, 2021). Utilizing technological tools to avail social support helps keep friends and family connected and supportive of one another through a trying time. Resilience during the pandemic is positively impacted by supportive families, good parenting, and effective coping mechanisms. During a severe disruption scenario like a pandemic, family activities are crucial for greater well-being (Petersen et al., 2021). The above-mentioned facts from the available literature help us devise our hypotheses 3 and 4 as coping strategy influences MH positively and mediates the relationship between FS and MH.

Social support is defined as “support made available to an individual through social relationships to other people, groups, and the greater community” (Heaney and Israel, 2008). There are multiple ways to measure social support, however, one of the most commonly used indexes of social support has been PSS. It is not only easy to measure the PSS but the accuracy aspect is also of utmost importance as it depicts a relatively precise state of MH. Numerous studies have shown how PSS improves MH in stressful circumstances (Wang and Eccles, 2012).

Though PSS and received social support are entirely distinctive from each other, however, it is generally believed that the relationship between these otherwise distinct supports should be high, particularly in scenarios when the demanded support is equivalent to the provided support (FasihiHarandi et al., 2017). Similar to this, other authors contend that the recollection of helpful behaviors displayed can be used to gauge perceived support (Özmete and Pak, 2020). We believe that social support received by an individual during the COVID-19 pandemic equals the required social support by an individual, therefore, we measured that the received social support reflected the level of PSS. However, social support and social support resources may not be mistaken as similar but they should be understood as distinct.

According to a study one’s subjective perceptions of the availability of social support from other users on social networks are represented by perceived support (Noret et al., 2020). According to various studies, perceived support quality is more closely correlated with MH than real personal network structure (Kandeğer et al., 2021). Numerous factors have an impact on people’s MH. Previous research has established that the availability of social support along with effective coping strategies helps maintain good MH (FasihiHarandi et al., 2017). A correlation between social support provided by the family and stress symptoms related to post-trauma was found. However, no such correlations were found when social support was provided by the friends. In a different study, peer social support was found to be more effective at preventing psychological distress than social support provided by family members (Özmete and Pak, 2020).

Recently, there has been a rise in interest in new approaches such as latent profile analysis (LPA) for studying the many types of networks of social support (Mai et al., 2021). Furthermore, some scholars utilized cluster analysis to analyze typologies of the social network, however, these analyses are based on different notions. More significantly, the studies that have already been published showed that various support profiles would have varying relationships with outcomes related to MH (Wang et al., 2018). In particular, one study used LPA and discovered that four different profiles of social support have different correlations to elder migrants’ well-being (Burholt et al., 2018).

According to one study, older persons who participated in the social support profile that was locally integrated had a lower risk of dementia than those who participated in the social support profile that was family-dependent (Yang et al., 2020). Researchers discovered different patterns of relations between different profiles of social support and outcomes of MH. The susceptibility of the youth was found to be higher vis-à-vis pessimism and stress irrespective of sources of social support. Social support particularly from family and friends plays a significantly important role in managing MH menaces such as anxiety and depression. LPA directly estimates the membership probability from the model, making it more plausible to examine the outcomes of individual clustering than the conventional method of clustering. LPA being a person-centric strategy helps specifically in identifying different subgroups of people who have similar social support behaviors (Mai et al., 2021). A regression mixed model can be constructed using the LPA model and additional outcome variables to confirm the association between profiles and results. In order to investigate the profiles of sources of social support and the association between these profiles and MH in this study, we did an LPA and used the regression mixture model (BCH technique).

According to research, social support serves as a buffer against depression or the unfavorable consequences of stress (Wilson et al., 2020). It also serves as a buffer between stressful life events and depression. Keeping in view the above facts from the literature the current study assumes that PSS from friends and others positively moderates the relationship between FS and MH it is also hypothesized that PSS from friends and others positively moderates the relationship between FS and coping strategy.

Research framework

The prospective respondents were the students studying in various universities of China. We selected the Henan province of China for this purpose and 10 large universities of the province were the main target. Because of Covid, it was expected that most of the students stayed at their homes and with their families faced some short of mental disorders. Therefore, to collect data about the study variables, we used a validated survey questionnaire which was circulated to the selected universities’ help desks/information desks for anonymous self-rated responses. As the data collection was related to highly sensitive and personal matters, i.e., MH, family/friends support and coping strategies, and because the population was unknown, we opted for a combination of snow ball and purposive sampling technique (non-probability sampling). In addition to this, for identifying the target respondents, we also sought help from the staff of student affairs.

We targeted 350 deaf students from the concerned universities as the exact no. of students in the target universities remained unknown and not confirmed by the concerned authorities. We therefore, followed the guidelines of various scholars; for example, recommended that “each item should be represented using 5 samples,” who suggested that “sample of 300 will be considered as good,” proposed that “the size of the sample should be 20 times greater than the expected factors” (if factor analysis is to be conducted) and for conducting SEM, N = 100–150 is acceptable (Anderson and Gerbing, 1988). Keeping in view the recommendations from these researchers and an average response rate (if any) we determined a sample size of 350 out of which 260 questionnaires were filled and 210 turned out to be valid (response rate 60%). These questionnaires were used for further analysis.

All the scales used in this study were adapted from previous researches and were already validated by the scholars. Family support was adapted and its sample item was “I get the emotional help and support I need from my family.” The scales of coping strategies and were adapted from Hamby, and its sample item was “When dealing with a problem, I try to see the positive side of the situation.” The questions of perceived support from friends were adapted, and its sample item was “I can count on my friends when things go wrong.” Likewise the scale items of other’s support were adapted the study of, and its sample item was “There is a special person in my life who cares about my feelings.” Moreover, the measurement questions of MH and its three dimensions, i.e., cognitive, emotional, psychology were adapted and its sample item was “I’ve been feeling optimistic about the future.” All the items were measured at 7 Point Likert scale.

To analyze the data, SPSS 21 was used for calculating basic and descriptive statistics while Mplus 7 was applied for hypothesis testing. But prior to moving toward hypothesis testing, we have examined the measurement model in order to ensure the reliability and validity of the constructs, the details of which are discussed in the next section.

First of all, researchers analyzed the measurement model keeping in view the requirements of “convergent validity and discriminant validity,” while structural model was assessed in the next step (Anderson and Gerbing, 1988). This research applied Mplus 7, a frontline software which has the potential to handle both normal and non-normal data (Van Der Linden et al., 2017). Mplus 7 output provides the following model fit indices, i.e., “Chi-square,” “root mean square error of approximation (RMSEA),” “comparative fit index (CFI),” “Tucker Lewis index (TLI)” and “standardized root mean square residual (SRMR)” which were used to determine the model fitness. Accordingly, the criterion for model fit indices given by Hu and Bentler (2009) in Table 1 and it can be observed that all the values are lesser than the threshold levels.

According to Hair (2009), convergent validity is assessed taking into account the standardized loadings of all the items of the construct/scale. It is important for the scale items to have STDYX >0.5, which is this case. Moreover, the research instrument had 34 items in total and the standardize loadings of all the items ranged from 0.694 to 0.894, which signals the existence of convergent validity (see Figure 1). Additionally, to ensure the internal consistency, convergent validity of the scale, we calculated CR and AVE of the constructs. As the study conducted confirmatory factor analysis, that’s why it was important to compute to these values. However, according to the criteria, for a construct to have internal consistency, it is recommended that the values of CR to remain at least 0.7 or greater, while the values of AVE should at least pass the minimum threshold level of 0.5 (Fornell and Larcker, 1981). The results in Table 2 confirmed that the scale met the criteria of internal consistency.

In order to determine whether the scale has discriminant validity or not, comparing the squared root of the AVE with the correlation coefficient was essential (Fornell and Larcker, 1981). For this purpose, diagonal bolded values were compared with the off-diagonal values in Table 3. It can be observed that the diagonal values remained greater than the correlation coefficients, which depicted a strong evidence of validity.

In addition to this, we calculated variance inflation factor to test the existence of multicollinearity. The analysis of the multicollinearity showed that the VIF values were <3 which confirmed the model did not have the issue of multicollinearity.

The variability of the data was assessed by analyzing the mean values and standard deviation from the mean. The analysis presented in Table 4 revealed the MEAN values of the variables and their standard deviation.

The mean scores of the variables ranged from 5.01 of FS to 6.38 of PSS whereas the standard deviation ranged from 0.587 of COP to 1.023 of FS. The values of standard deviation remained well within prescribed range and the data was found to be good to enough to be handled using Mplus (Van Der Linden et al., 2017). To compliment this, we also calculated Skewness/Kurtosis to make assessment about the normality of the data. It was found that the values of both the statistical measures were well within acceptable range. Table 4 placed above provides variable specific details about normality and data variability.

As the researcher collected survey data at a single point of time and from the same students who suffered some sort of MH issues during COVID-19, therefore, it was important to make sure that the correlation between the said variables was not spurious. For this purpose, Harman’s single-factor was applied to assess the values of CMV. As shown in the Table 5, the single factor explained 33.269% of variance, which was far below the standard limit of 49%. If the result demonstrate that less than half (50%) of the variance is explained by the factor, then it is believed that CMB is not a cause of concern.

The Following section presents the results of direct association, indirect association and the role of the moderator, respectively.

Table 6 contains information about the hypotheses of direct relationship. As pointed out by Henseler et al. (2009), the standardized coefficients are the same as the regression coefficients, that’s why, the decision about the hypothesis were made based upon the standardized coefficients and their significance (value of ps) provided in Mplus 7 output. The results demonstrated that FS was strongly positively linked to MH with β = 0.236 and was significant at 0.00. Therefore, H1 was supported (see Figure 2). Likewise, H2 examined the relationship between FS and coping strategies. The results exposed that FS positively predicted coping strategies with β = 0.520 and p-value 0.00. Hence H2 was also found to be supported (see Figure 3). Finally, for direct relationship, the link between coping strategy and MH was also determined. The researcher found out that “coping strategy” strongly predicted MH with β = 0.534 and p-value 0.000. Therefore, H3 was supported as per the prediction of the researcher (see Figure 4).

Results in Table 7 were about H4 which was related to the mediation analysis. As per the requirement, mediation analysis was performed to confirm whether the intervening/mediating variable (MV) enhanced the impact of IV to the DV (Hair, 2011). The researcher opted for, “bootstrapping procedures” to test the significance of the mediation path (Preacher and Hayes, 2004; Preacher et al., 2010). The analysis revealed that specific indirect effect was significant at 0.05. The STDYX of SIE (specific indirect effect) was β = 0.301 and (p = 0.00) with a 95% CI [0.154–0.148]. Because the upper CI and lower CI values did not include zero, therefore, it was confirmed that “coping strategies” positively and significantly mediated the relationship of “family support” and “MH.” Thus H4 was supported (see Figure 5).

Fore investigating the moderating role of PSS, the study first examined its role on the relationship of FS and MH. Therefore, in order to proceed with the analysis, researcher applied “Xwith” technique for creating the interaction term between concerned IV-MV (Van Der Linden et al., 2017) and for assessing the effect of interaction term (see Figures 6, 7). These results exposed that H5 was not supported as the Table 8 indicated that the interaction effect of PSSXFS (moderating variable * IV) was β = 0.147 (p = 0.228) with a 95% CI [−0.054 to 349].

In Hypotheses 6, it was predicted that PSS will positively moderate the relationship of FS and COP. The results in Table 8 indicated that the direct effect of FS on COP was β = 0.263 (p = 0.00) and its 95% CI [0.135–0.391], the direct effect of moderating variable PSS on COP was β = 0.695 (p = 0.00) with 95% CI [0.536–0.855] and the interaction effect of FSXPSS (moderating variable * IV) was β = 0.357 (0.002) with a 95% CI [0.171–0.544] (see Figure 8). Additionally, Simp-Slope analysis revealed that this moderation effect was greater when the values of PSS were higher and positive, whereas, the effect was lower and positive when the values of PSS were lower. These values implied that H6 was supported (see Figure 9).