Change in mental health, physical health, and social relationships during highly restrictive lockdown in the COVID-19 pandemic: evidence from Australia

The impact of COVID-19 lockdown on mental health around the world, and in Australia

Government policies enforcing restrictive lockdown to curb the spread of COVID-19 have the potential to negatively impact on fundamental psychological needs. According to self-determination theory (SDT) three core psychological needs are competence, autonomy and relatedness (Ryan & Deci, 2000). Mandated social distancing during a pandemic may detract from an individual’s sense of autonomy (i.e., sense of personal control and freedom) and relatedness (i.e., sense of closeness to other people). The flow on negative impact on the economy may further exacerbate stress as people feel a threat to their competence (i.e., sense of self-efficacy) as jobs are threatened or lost. Similarly, the same kind of logic can be applied to Maslow’s hierarchy of needs that includes physiological, safety, social, esteem, and self-actualisation elements (for a discussion see: Ryan et al., 2020).

Quarantine (lockdown) during prior pandemics such as the severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS) were reported to negatively affect psychological well-being (Brooks et al., 2020). Research on people’s experience during COVID-19 lockdowns around the world have produced consistent findings of increased levels of stress, anxiety and depression (Baloch et al., 2021; Benke et al., 2020; Bruno et al., 2021; Ebrahimi, Hoffart & Johnson, 2020; Fiorenzato et al., 2020; Gao et al., 2020; Hamadani et al., 2020; Holingue et al., 2020; Holman et al., 2020; Huang & Zhao, 2020; Kalaitzaki, 2020; Lee, 2020; Marashi et al., 2020; Mazza et al., 2020; Ozamiz-Etxebarria et al., 2020; Patrick et al., 2020; Pierce et al., 2020; Qiu et al., 2020; Roy et al., 2020; Sameer et al., 2020; Saraswathi et al., 2020; Twenge & Joiner, 2020; Wang et al., 2020; Zacher & Rudolph, 2020).

Factors reported as being associated with higher levels of emotional distress during lockdown are being younger, female, having lower education, having pre-existing medical conditions, increased social media use, rumination on COVID-19 (Baloch et al., 2021; Benke et al., 2020; Bruno et al., 2021; Bu et al., 2021; Ebrahimi, Hoffart & Johnson, 2020; Fiorenzato et al., 2020; Gao et al., 2020; Holman et al., 2020; Hsiang et al., 2020; Huang & Zhao, 2020; Kalaitzaki, 2020; Lee, 2020; Mazza et al., 2020; Ozamiz-Etxebarria et al., 2020; Qiu et al., 2020; Roy et al., 2020; Wang et al., 2020), whereas keeping physically and socially active have been reported as protective factors (Brand, Timme & Nosrat, 2020; Bu et al., 2021; Ebrahimi, Hoffart & Johnson, 2020; Galle et al., 2020; Marashi et al., 2020).

Some Australian surveys have been conducted showing that the lockdown experience, similar to other parts of the world, had a broadly negative impact on mental health (Biddle et al., 2020a; Brand, Timme & Nosrat, 2020; Fisher et al., 2020; Li et al., 2020; Newby et al., 2020; Phillipou et al., 2020; Rossell et al., 2021; Stanton et al., 2020; Titov et al., 2020; van Agteren et al., 2020; Westrupp et al., 2021). In a survey of 5,070 Australians in the first week of April 2020, Newby et al. (2020) found that 55% of respondents felt their mental health had worsened a little, and 23% a lot. Primary worries underlying this were that 50% or more indicated worrying about catching the virus, feeling moderately-extremely lonely, and feeling uncertain about their financial situation. In a survey with 13,829 respondents, Fisher et al. (2020) also found overall elevated levels of psychological distress.

The potential impact of COVID-19 lockdown on social relationships

One underlying psychological mechanism to explain how COVID-19 lockdown might reduce mental health is via negative impacts on social relationships. Social relationships require maintenance where the extent of required maintenance can vary a great deal depending on the closeness of the relationship, and personal qualities and preferences (Blieszner & Ogletree, 2017; Fehr, 2004; Mesch & Talmund, 2006; Ogolsky & Bowers, 2012). COVID-19 lockdown experiences have the potential to significantly impair core aspects of relationship maintenance by disrupting the frequency and quality of shared experiences, and interpersonal communication patterns. This could be associated with psychological consequences such as a decreased sense of closeness and a reduction in perceived relationship quality (Aleman & Sommer, 2020; Lardone et al., 2020).

Engaging in shared activities/experiences has been identified as one key aspect of relationship maintenance (Daniels, Watson & Gedikli, 2017; Girme, Overall & Faingataa, 2014; Jolly et al., 2019; Mesch & Talmund, 2006; Rossignac-Milon & Higgins, 2018). The closure of sources of social activities (e.g., restaurants, bars, sporting clubs, gyms, movie theatres, etc.) during COVID-19 lockdown arguably has the potential to impair social relationships via disrupting engagement in shared activities for relationship maintenance. While engaging in shared activities can facilitate social bonds, so too can the simple act of conversing with others (Rossignac-Milon & Higgins, 2018). Communicative acts such as self-disclosure (Sprecher & Treger, 2015), reminiscing (Roberts, 2018), sharing news (Reis et al., 2010) and gossip (Yucel et al., 2020), all have the potential to serve as bonding experiences. Therefore, social distancing policies associated with COVID-19 lockdown has potential to reduce the amount or frequency of social interaction, as people spend more time at home.

Lockdown also has the potential to impact upon the type of social interaction, with an increased amount of communication via technology (e.g., phone, email, social media) to compensate for reduced opportunities for face-to-face interaction. Research suggests that communication via technology can be considered as impoverished compared to face-to-face communication via a reduction in certain verbal (e.g., tone of voice) or non-verbal (e.g., eye contact, gestures, or body posture) communicative cues (Colvin et al., 2004; Kock, 2002; Kock, 2004; Lieberman & Schroeder, 2020; Walther, 1996; Walther & Parks, 2002). As a consequence it is not surprising that people generally have a preference for face-to-face communication over technology-mediated communication in most instances (Flaherty, Pearce & Rubin, 1998; Wilson et al., 2020). Therefore, even if a person tries to compensate for reduced face-to-face interaction via a corresponding increase in technology-mediated communication, theoretically there remains potential for a negative impact on perceived relationship quality if the subsequent technology-mediated communication is experienced as a lesser form of interaction.

Research investigating social relationships during COVID-19 lockdown

As discussed in the prior section, it is reasonable to expect that social relationships might be negatively impacted during COVID-19 lockdown. However, recent studies that have investigated loneliness during the pandemic have been reporting mixed findings. Some studies have reported an increase in loneliness (Bu, Steptoe & Fancourt, 2020; Killgore et al., 2020; Van Tilburg et al., 2020), whereas others have reported there has been no significant increase in loneliness during lockdowns compared to pre-pandemic levels (Groarke et al., 2020; Luchetti et al., 2020).

Few studies have directly asked targeted questions about perceived relationship quality during the pandemic (Biddle et al., 2020b; Bulow et al., 2020; Cooper, Pauletti & Di Donato, 2020; Philpot et al., 2021), and even less about the use of technology-mediated communication (Monin et al., 2020). In a recent Australian study Biddle et al. (2020b) asked their 3,249 respondents from the general public “How has the quality of your relationships with other people/family members in your household changed since the spread of COVID-19?”. Responses were: 2% a lot more difficult/strained, 15% a little more difficult/strained, 55% no change, 23% a little closer/stronger, and 5% a lot closer/stronger. Biddle et al.’s (2020b) study provides some evidence to suggest that during the lockdown in April there may not have been any broad large negative impact on social relationships in the Australian population.

The present study

Our study aims to contribute to the emerging literature on the mental and physical health impacts of government mandated lockdown policies on the public during COVID-19. We also take an in-depth look at how communication patterns changed regarding face-to-face and technology-mediated communication to examine how these different modes of communication are associated with self-reported quality of social relationships. Our survey was conducted from 14th–30th April in 2020 when all non-essential businesses were closed, and people were not supposed to congregate in groups larger than two people outside of their immediate household members. We were anticipating that like other Australian studies investigating the same period in time, we would find a substantial portion of participants self-reporting deterioration across their mental health, physical health, and financial situation (Fisher et al., 2020; Newby et al., 2020; Phillipou et al., 2020).

Based on background psychological literature on relationship maintenance, and the impoverished nature of technology mediated communication (when compared to face-to-face communication), we were initially predicting that COVID-19 lockdown would be associated with a negative impact on social relationships in our sample. However, based on Biddle et al. (2020b) results (that was published after our study commenced), we were subsequently anticipating that social relationships might not be greatly impacted for most individuals. Extending upon Biddle et al. (2020b), we further explore potential changes in participant behaviour regarding the extent of face-to-face and technology-mediated social interaction, and whether (or not) changes in social interaction are associated with changes in perceived relationship quality.

We also asked participants about a range of concerns they were experiencing about COVID-19 to compare how concerns about their personal relationships compares with other concerns. If we find that social relationships are not being heavily impacted, then we also expect that self-reported concerns about social relationships will overall be lower compared to other concerns. Prior research has indicated that concerns regarding catching the virus and the economy have typically been high during the pandemic (Ebrahimi, Hoffart & Johnson, 2020; Newby et al., 2020; Saraswathi et al., 2020; Timming, French & Mortensen, 2021; Wang et al., 2020). We also look at other factors that have been examined in other parts of the world but not yet given much attention by Australian researchers (i.e., frequency of exercise, COVID-19 rumination, and social media exposure). Based on what has recently been reported in the literature, we hypothesize that in our sample higher frequency of exercise will be positively associated with emotional well-being (Bu et al., 2021; Ebrahimi, Hoffart & Johnson, 2020; Galle et al., 2020), whereas increased rumination on COVID-19 and increased use of social media will be negatively associated with emotional well-being (Bu et al., 2021; Gao et al., 2020; Hsiang et al., 2020; Lee, 2020; Roy et al., 2020).

Participants and procedure

An online survey was disseminated between 14th-30th April 2020 via social media channels (primarily Facebook). We received 1599 responses, 78% female. Age of participants in years: 18–21 (6%), 22–25 (4%), 26–30 (11%), 31–35 (13%), 36–40 (11%), 41–50 (22%), 51–60 (18%), 61–70 (11%), 71+ (3%). Employment status: Full-time (44%), part-time (19%), casual (12%), stay at home parent (8%), retired (9%), unemployed (8%). The industry worked in: Health care and social assistance (24%), Education and training (22%), Professional, scientific and technical services (9%), retail trade (6%), administration (5%), with the remaining 44% spread across other industries in smaller proportions. Most of our participants resided in Western Australia (81%), with the remainder of the sample coming from the various other Australian states and territories. Most participants (75%) were in a committed relationship at the time of the study. The lockdown period began from the start of April 2020, where gatherings were restricted to two people (except for immediate members of households). Participants in our sample had been in lockdown for an average of 18.84 days (SD = 4.62, min = 14, max = 30). Prior to running the survey, ethics approval was obtained from Edith Cowan University (Ref: 2020-01305-ROGERS). Informed consent was obtained from a dedicated question embedded at the beginning of the survey asking participants to consent for their data to be used for research purposes.

Survey emasures

Perceived impact of COVID-19 lockdown on emotional well-being, physical well-being, relationships, and work

We asked participants a series of questions about their emotional well-being, physical well-being, and exercise reflecting over the prior month, and for the same time last year. The results from these measures are presented in Fig. 1. Both emotional (t (1598) = 24.45, p < .001, r = .52) and physical (t (1598) = 19.58, p < .001, r = .44) well-being were found to be lower during the COVID-19 lockdown period compared to participant estimates of their health the same time last year. More specifically, while 13% of participants estimated greater negative than positive emotion for the prior year (i.e., BEES score <0), this increased to 41% of participants for the prior month. Similarly, 19% of participants reported greater negative than positive physical well-being for the prior year (i.e., BEEPS score <0), increasing to 42% for the prior month. As indicated by the effect sizes (r), and as can be seen in Fig. 1, these represent statistically moderate effects. When comparing estimates of last year with the present month, we also found an overall decrease in self-reported frequency of exercise (t (1598) = 9.80, p < .001, r = .24). However as indicated by the effect size (r), and can be seen in Fig. 1, this represents a relatively small effect.

Correlations among self-reported emotional well-being, physical well-being, and exercise, during lockdown and last year, are presented in Table 1. Emotional well-being (BEES) was found to be positively associated with physical well-being (BEEPS) during lockdown (Pearson r = .60, p < .001), and last year (Pearson r = .57, p < .001). Only weak positive associations between current month and last year values were found for both the emotional wellbeing and physical wellbeing suggesting that participants were able to dissociate last year from the current year when making their appraisals. Frequency of exercise was found to positively associate with physical well-being, during lockdown (Pearson r = .40, p < .001), and last year (Pearson r = .48, p < .001). Whereas there were only weak positive associations between frequency of exercise and emotional well-being. This replicates prior work by Skead & Rogers (2016) that suggests exercise may have an indirect benefit for mental well-being via enhancing one’s sense of physical well-being.

Perceived impact of COVID-19 lockdown on emotional well-being, physical well-being, relationships, and work

We directly asked participants to estimate the impact of COVID-19 lockdown regarding their mental health, physical health, finances, work, and relationships, see Fig. 2. A substantial proportion of participants indicated at least some deterioration of their mental health (54%), physical health (41%), financial situation (39%), and productivity with work (41%). Note that for most individuals reporting a deterioration, the response is somewhat rather than a lot (see Fig. 2). Also, except for mental health, the most common response for physical health, financial situation, and work productivity was no change. Interestingly, there appears less impact on social relationships, as only approximately 17% indicated deterioration across different types of relationships, while approximately 66–72% reported no change, and 8–16% reported improvement. As will be shown later in the results section, many participants appeared to compensate for a decrease in face-to-face contact using technology to keep in contact with others.

Inter-correlations among the variables in Fig. 2 are presented in Table 2. Participants’ present month emotional well-being score was positively associated with participants’ self-rating of COVID-19 lockdown impact on their mental health (Spearman r = .64, p < .001), and present month physical well-being was positively associated with their rating of COVID-19 lockdown impact on their physical health (Spearman r = .49, p < .001). There are only weak positive associations among most other variables.

Concerns related to COVID-19

We asked participants the extent they were concerned about a range of social and economic issues due to COVID-19, see Fig. 3. We found a substantial proportion of participants to have concern across all issues, however the two stand outs are concern about friends and family catching COVID-19 (63% very/extremely concerned), and a negative impact on the economy (73% very/extremely concerned). Note that concern over personal relationships suffering is much lower in comparison (20% very/extremely concerned).

Inter-correlations among the different concerns, and relationships with emotional well-being measures, are presented in Table 3. All concerns were found to be negatively associated with emotional well-being. Associations among concerns varied depending on the similarity of the concerns. As an example, a strong association was found between self-concern catching COVID-19 and concern for friends/family catching COVID-19 (Spearman r = .60, p < .001). Whereas these concerns about catching the virus were only weakly positively associated with all other concerns.

Amount of personal attention given to COVID-19

We expected that participants would report giving a lot of personal attention to COVID-19 considering the extent of concern associated with it. Indeed, as can be seen in Fig. 4, between 35–66% of participants were thinking about COVID-19, monitoring news information, and talking with others about COVID-19 multiple times per day.

Inter-correlations among items are presented in Table 4. Across the items the clearest negative association with emotional well-being was the item ‘thinking about COVID-19’ (Spearman r =  − .27, p < .001). Therefore, we found a moderate association between rumination about COVID-19 and lower emotional well-being.

Perceived changes in social behaviour due to COVID-19

Many participants reported a substantial shift in their social behaviour in response to the social distancing policies in place at the time of the survey, see Fig. 5. Many participants reported more distancing from others (88%), isolating themselves from other people (75%), and spending more time on social media (63%). Spending more time on social media was found to be negatively associated with BEES emotional well-being (Spearman r = -.28, p < .001), and perceived change in mental health (Spearman r = -.25, p < .001). Participant’s face-to-face interaction time typically increased for time with one’s partner (52%), and decreased for time with family (54%), friends (86%), and work colleagues (64%). Whereas technology mediated interaction time typically increased for family (57%), friends (56%), and work colleagues (49%). Therefore, results suggest that people were compensating for the decreased face-to-face interaction time with increased technology mediated interaction time.

Associations between change in communication and change in relationship quality

Correlations between perceived change in relationship quality and change in the extent of both face-to-face and technology-mediated communication are presented in Table 5. A consistent weak-moderate positive association between change in face-to-face interaction and change in each type of relationship was found. Weak-moderate positive relationships were also found between change in technology mediated communication and perceived change in quality for relationships with family, friends, and work colleagues (but not partner).

There is a high proportion of participants reporting no change in perceived relationship quality across all relationship types. Therefore, the relatively low correlations in Table 5 might represent an under-representation of the strength of association between the variables due to a lack of spread across the response options, and/or non-linear patterns of association. In Fig. 6 we graphically present the associations between the variables which helps to provide a clearer indication of how both face-to-face and technology-mediated communication are associated with perceived change in relationship quality. For example, Fig. 6A shows that most of the participants in a relationship that felt their relationship with their partner had improved stated their face-to-face communication had increased (78%). Whereas this percentage is substantially lower in participants reporting no change (49%) or a deterioration in relationship quality with their partner (39%). A similar pattern is also evident across the other types of relationships such as (B) family, (C) friends, and (D) co-workers, see Fig. 6. This indicates that the positive association between more face-to-face communication and improved relationship quality might be stronger than the correlation values suggest.

A similar pattern is also evident for technology-mediated communication across relationship types (except for relationship with partner). For example, the proportion of the sample that reported an improvement in friendships quality typically also reported an increased level of technology-mediated communication (85%). Whereas the amount of participants reporting an increased level of technology-mediated communication was comparatively lower for participants reporting no change (59%) or a deterioration of friendship quality (40%), see Fig. 6G. This indicates that the positive association between more technology-mediated communication and improved relationship quality might be stronger than the correlation values suggest. Overall, our findings indicate that both face-to-face interaction and technology-mediated interaction had a role to play in maintaining relationships during the April lockdown period.

Emotional and physical well-being, exercise, concerns, social media use, and COVID-19 rumination

We found that during COVID-19 lockdown 41% and 42% of participants were experiencing greater negative than positive emotional and physical well-being, respectively. This contrasts with only 13% and 19% reporting greater negative than positive emotional and physical well-being when estimating how they were feeling the same time last year. Therefore, our results are consistent with other recent findings from around the world that the COVID-19 lockdown experience is associated with a negative impact on well-being for a substantial proportion of the public (Biddle et al., 2020a; Ebrahimi, Hoffart & Johnson, 2020; Fisher et al., 2020; Gao et al., 2020; Huang & Zhao, 2020; Lee, 2020; Mazza et al., 2020; Newby et al., 2020; Ozamiz-Etxebarria et al., 2020; Phillipou et al., 2020; Pierce et al., 2020; Qiu et al., 2020; Roy et al., 2020; Twenge & Joiner, 2020; Wang et al., 2020; Westrupp et al., 2021).

We only observed a small decrease in frequency of exercise. One potential reason for only a minor reduction in exercise is that our sample appears to have been relatively sedentary prior to the lockdown, see Fig. 1C. Another reason is that some individuals may have increased exercise (with more time on their hands) that cancels out those exercising less. Consistent with other recent studies (Brand, Timme & Nosrat, 2020; Bu et al., 2021; Ebrahimi, Hoffart & Johnson, 2020; Galle et al., 2020; Marashi et al., 2020), we found a link between frequency of exercise with both physical and mental well-being suggesting that keeping physically active can potentially act as a buffer for stress during lock-down. Codella et al. (2020) have also pointed out that being physically active has benefits for maintaining a healthy immune system, which is obviously very relevant during a pandemic.

We acknowledge that relying on participant retrospective accounts of their mental health, physical health and exercise from the previous year is not ideal. This approach has the potential to exaggerate differences between pre- and post-lockdown ratings. This is because the reflection on last year might be biased by current feelings. However, the correlations between lockdown and last year estimates for mental health (r = .18) and physical health (r = .34) were not strong, which does provide some evidence to suggest that participants were able to dissociate current feelings from prior feelings. Comparatively, current estimate of frequency of exercise was more strongly related to prior year estimates of exercise (r = .56).

When asking participants to directly rate the impact they perceived COVID-19 lockdown was having on their mental health, our results are similar with another Australian survey by Newby et al. (2020). In our sample 44% felt their mental health had deteriorated somewhat, and 10% a lot. In Newby et al.’s (2020) study 55% felt their mental health had worsened a little, and 23% a lot. The higher level of distress in Newby et al.’s study (2020) might be attributed to their data collection occurring slightly earlier than ours. When their study occurred, there was less clarity that the lock-down was flattening the infection rate curve. Also, they reported quite a high proportion of their participants (70%) with prior lived experience with a mental health diagnosis. We concede a limitation of our own study was that we did not query this with our own participants, however we expect that if we had such a question in our study the figure would be lower than 70%. Additionally, our sample is primarily from Western Australia whereas their sample contained more people from the Eastern states. As can be seen in Fig. 7 the number of recorded COVID-19 cases in the month of March leading up to the lockdown was higher in the Eastern states of New South Wales (NSW), Victoria (VIC) and Queensland (QLD) compared to Western Australia (WA). When considering population size, the differences were negligible, as WA has a substantially lower population size compared to these other Australian states. However, at the time the media consistently reported in terms of number of cases and communicated that the situation was worse in the Eastern states. We suggest this may have been associated with heightened anxiety in the population of the Eastern states compared to Western Australia and a potential reason for the difference between our findings with those of Newby et al. (2020).

We are hesitant to frame our findings as demonstrating any substantial increase in clinical levels of distress. We argue that diminished well-being during such an experience (i.e., lockdown) that impacts on core psychological needs such as competence, autonomy and relatedness is a natural and understandable human reaction. We expect that most people will possess the resilience to recover quickly. However, as suggested by Gruber et al. (2020) there might be a proportion of participants for whom the lockdown experience acts as a traumatic precipitating event for the development of clinical depression or anxiety. Future research is required to better understand the precise extent of such occurrences.

Another finding of ours consistent with Newby et al. (2020) was that the dominant concerns of our participants were related to friends or family catching COVID-19 (63% very/extremely concerned), and the economy (73% very/extremely concerned). Consistent with research from other parts of the world (Gao et al., 2020; Hsiang et al., 2020; Lee, 2020; Roy et al., 2020), we found that such concern was perhaps influencing people to seek out COVID-19 information, as a large proportion of participants (i.e., 35%–55%) were monitoring sources of information about COVID-19 multiple times per day. Or alternatively, high exposure to sensationalist media coverage might have been exacerbating people’s concerns (Bendau et al., 2020; Bu et al., 2021). Also, we found evidence for a high degree of rumination on COVID-19, as most participants (66%) reported thinking about COVID-19 multiple times per day. In our study we found such rumination was associated with lower levels of emotional well-being. Additionally, we found that many people (63%) reported engaging with social media more often, and consistent with Gao et al. (2020) we found a weak negative association with social media use and emotional well-being.

Perceived change in social relationships and communication patterns during lockdown

In our study most participants reported no deterioration in their social relationships, and most were not concerned about it. More specifically, only 13–26% reported a deterioration in their relationships with their partner, family, friends, or work colleagues, while 66–72% reported no change, and 8–16% reported improvement. Across the relationships of family, friends, and work colleagues, face-to-face interaction time typically decreased (54–86% of participants), while technology-mediated interaction via phone, email, or online correspondingly increased (49–57% of participants). An exception to these findings was type of communication with one’s partner, with most participants responding that they had been spending more time interacting face-to-face with their partner (52%), with technology-mediated communication typically unchanged (64%).

Considering prior literature on relationship maintenance (Blieszner & Ogletree, 2017; Fehr, 2004; Mesch & Talmund, 2006; Ogolsky & Bowers, 2012; Rossignac-Milon & Higgins, 2018), media naturalness theory (Kock, 2002; Kock, 2004), and a general preference for face-to-face interaction (Flaherty, Pearce & Rubin, 1998; Wilson et al., 2020), at the outset of our study we were anticipating that a sudden decrease in face-to-face interaction would be associated with diminished perceived relationship quality. On the one hand, we did find that change in face-to-face interaction was positively associated with change in relationship quality. That is, participants reporting increased face-to-face interaction also tended to be more likely to report an improvement in relationship quality. This pattern was consistent across all relationship types. However, we did also find that most people reported no change in relationship quality, which indicated no systematic large scale negative impact on relationship quality during lockdown across multiple types of social relationships (i.e., partner, family, friends, and work colleagues). This finding is consistent with recent Australian research by Biddle et al. (2020b) whom reported that a majority of participants reported no change in relationship quality with other people/family members in their household.

We offer a couple of complimentary potential explanations for this finding. First, while lockdown drastically reduces opportunity for many types of relationship maintenance experiences (e.g., going to see a movie at a cinema), it arguably also provides an over-arching experience that has potential to be a source of relationship bonding. In our study, 94% of participants reported ’talking with people about COVID-19’ at least a few times per week, and 45% reported multiple times per day. Additionally, the pandemic appears to be a largely negative experience for many people around the world (Benke et al., 2020; Ebrahimi, Hoffart & Johnson, 2020; Fiorenzato et al., 2020; Gao et al., 2020; Hamadani et al., 2020; Holman et al., 2020; Huang & Zhao, 2020; Kalaitzaki, 2020; Lee, 2020; Marashi et al., 2020; Mazza et al., 2020; Ozamiz-Etxebarria et al., 2020; Pierce et al., 2020; Qiu et al., 2020; Roy et al., 2020; Twenge & Joiner, 2020; Wang et al., 2020; Zacher & Rudolph, 2020), and we have reported a substantial proportion in our sample of Australians being negatively affected during lockdown. Numerous studies have indicated that shared experiences of pain, adversity, and/or hardship have the potential to act as social bonding experiences (some examples: Bastian, Jetten & Ferris, 2014; Bastian et al., 2014; Bastian et al., 2018; Breslin, 2019; Shaw, Pollio & North, 2020).

The second potential reason why we did not find any substantial negative impact on social relationships is that we found many participants were making good use of communications technology to stay connected during the lockdown period. Across relationship types of family, friends, and work colleagues there was a positive association with change in technology-mediated communication and change in relationship quality. That is, participants reporting an increase in their technology-mediated communication were more likely to also report an improvement in relationship quality. Therefore, while modern communication technology such as social media has potential to reduce well-being via increasing rumination (Ohannessian, Fagle & Salafia, 2020; Parris et al., 2020) and excessive social comparison (Feinstein et al., 2013), such technologies are also arguably particularly useful for helping people keep informed and connected with others during lockdown experiences.

Study limitations

Our sample primarily consists of females (78%), in a relationship (75%), living in Western Australia (81%). Therefore, our sample is not representative of the entire Australian population so generalisations from our findings should be made with caution. Our survey was advertised primarily on Facebook about people’s experience during COVID-19 which may have attracted respondents who were concerned and stressed about COVID-19 to participate. Therefore, our results indicating an overall decrease in emotional/physical well-being, and high concerns about loved ones contracting the virus and concern for the economy, may be over-represented in our specific sample. Also, our sample is limited to people who are actively engaged with social media. Therefore, our findings regarding the extent of social media use, and technology-mediated communication use more broadly, may also be over-represented in our specific sample.

In our study participants had been under COVID-19 lockdown for 2-4 weeks. It could be that a more protracted lockdown would produce quite a different result. While technology-mediated communication appeared to compensate for reduced face-to-face time in the present study, in a lengthier lockdown situation there remains the distinct possibility that reduced face-to-face contact could result in degraded social relationships. There is a need for the further development of communication technologies to be better prepared for such future possibilities. An area of current development is communication in virtual reality via head mounted display (HMD) (Pan & Hamilton, 2018; Rosedale, 2016; Seymour, Reimer & Kay, 2018; Seymour et al., 2019). HMD-based virtual reality has the potential to become a technology-mediated mode of communication that can psychologically feel near identical to face-to-face interaction (Pan & Hamilton, 2018; Rosedale, 2016; Seymour, Reimer & Kay, 2018). Future research is needed to fast-track the development of this new mode of social interaction.

With a cross-sectional design, we are unable to make any confident conclusions regarding the direction of causality for the associations between variables in our study. For example, we have suggested that exercise might be a protective factor against stress, however it may also be that stress reduces motivation for exercise (Stults-Kolehmainen & Sinha, 2014). We have suggested that engaging in greater social media use might be exacerbating stress, but it could also be that stress is what is driving people to seek out more social connection on social media (Luo & Hancock, 2020). We have suggested that engaging in greater social interaction face-to-face and via technology-mediated modes of communication can help to decrease stress, but again it could be that it is stress that is motivating people to increase their interaction with others as they seek social support (Sameer et al., 2020).