Three conceptual models of restorative environments based on user’s restorative experience in the COVID-19 pandemic period

ABSTRACT The COVID-19 pandemic is not just a medical phenomenon, for it impacts not only physical health but the world’s increasing psychological burden. One of the efforts that can be made to reduce this burden is visiting places than can be considered capable of supporting the restoration process. Therefore, this study aims to understand the restorative experience felt by a group of people in Indonesia towards a particular environment or place during the COVID-19 pandemic. Data was collected through an online-based cross-sectional survey from 11th September to 13 October 2020 with 261 participants. The data were then analyzed using the content analysis method, which consisted of three stages: open coding, axial coding, and selective coding. The results show that there are three types of restorative environments commonly visited during the COVID-19 pandemic: Dependent Restorative Environment (D-REn), Conventional Restorative Environment (C-REn), and Enclosed Restorative Environment (E-REn). Each type of environment has a unique characteristic that cannot be found in other types. D-REn depends on routine and interaction, while C-REn heavily relies on the power of attraction, especially the value of visual appeal. On the other hand, E-REn exists because of the lack of choice, making it unpredictable and out of the box.


Introduction
The emergence of the global pandemic caused by the COVID-19 outbreak in various countries had severely affected all aspects of society (Wu, Chen, and Chan 2020;Pediatr 2020).Cases continued to increase exponentially over the few months since the virus first appeared, placing unprecedented pressure on the world.However, the increasing number of infections had exhausted the limited services that could be offered, and combined with the lack of staff and communication; it soon led to a significant and terrifying rise in medical emergencies (Luo, et al., 2020;Xiong et al. 2020;Ye et al. 2020).In addition to this problem, a pressing issue concerning handling and caring for those who were suffering not only physically, but also mentally, emerged.Several studies shown that the number of people affected by psychological disorders was higher than people infected by the COVID-19 virus (Benke et al. 2020;Marko, Košec, and Brecic 2020;Li, et al., 2020).The most significant indicators of psychological disorders developing were anxiety and depression, with a prevalence of 33% (28%-38%) and 28% (23%-32%).Apart from anxiety and depression, other indicators reported were distress, stress, insomnia, and symptoms of post-traumatic stress disorder (PTSD) (Luo et al. 2020b).
It's safe to assume that the primary source of the high prevalence of psychological disorders did not only come from the pandemic itself but also the indirect impact of the government policies and containment efforts (Lahav, et al., 2020;Wang et al. 2020;Benke et al. 2020;Brooks et al. 2020;Xiong et al. 2020).The rules were strict and limiting, which in turn caused irreparable problems, such as the destruction of the economy and dramatically altered people's life (Gloster et al. 2020).According to many sources, while effective for preventing the further spread of the virus, these different levels of containment measures (e.g., the prolonged quarantine, social distancing, and self-isolation) were bound to have their consequences, and one of them was tremendously related to stress (Facciolà, Laganà, and Caruso 2021;Lima, et al., 2020;Ye et al. 2020;Fiorillo, et al., 2020).
Stress can be described as the involuntary reaction felt by humans when they were unable to meet the demands of their surroundings.If a situation or circumstance is considered dangerous, threatening, or challenging, stress would present itself (Brannon and Feist 2007).In this situational framework, the uncertainty and fear associated with the transmission of the virus, combined with the social exclusion of the prolonged quarantine, created a period of stress that kept climbing up the notch (Presti et al. 2020).It was not surprising that many people were left vulnerable, whether it be infected persons, healthcare workers, or those who were not even infected but unfortunately suffered the repercussion of the worsening situation of the COVID-19 pandemic (Xiang et al. 2020;Wang et al. 2020;Xiong et al. 2020).
Dr. Samantha K. Brooks (2020) established seven types of stressors during the COVID-19 pandemic: 1) Duration of quarantine; 2) Fears of infection; 3) Frustration and boredom; 4) Inadequate supplies; 5) Poor information; 6) Financial loss, and 7) Social stigma.It's important to note that these stressors, as insignificant as they may seem, were proven to be very determinant to an individual's endurance and survival during this challenging time (Brooks et al. 2020).Suppose a person could not access environments that support the sufficiently rapid or complete restoration of needed resources; it is feared that these accumulated stressors' impact would increase the risk of suffering from severe physiological disorders.Therefore, visiting places considered capable of supporting restoration becomes especially important.
Restoration denotes a process or set of procedures through which one or more individuals renew or reestablish some adaptive resources or capacities that have become diminished (Lindern et al., 2016).Restorative approaches are not necessarily specific to a particular environment but may proceed more readily or smoothly in some intrinsic activities and environments.Restorative environments can accordingly be defined as environments that both permit and promote restoration (Hartig 2004).Kaplan and Kaplan (1989) describes a restorative environment as a place to rest and recuperate from mental fatigue.Korpela (1991) depicts it as a place that allows someone to relax and clear their thoughts from a stressful situation.In general, restorative environments can support restoration in two ways.The first is linked to the relative absence of perceived social and physical demands or stressors like crowding, noise, or in the case of the COVID-19 pandemic exceptional circumstances, reminders of stressful predicaments such as the risk of infection and monetary loss.On the other hand, the second refers to a certain quality of the environment that could promote restoration (Mittlemark, et al., 2017).
Empirical research has revealed that restorative experience appears in four different responses consisting of emotional, cognitive, physiological, and behavioral response.Han's (2003) Short-version Revised Restoration Scale (SRRS) provides an in-depth understanding of these responses.SRRS demonstrates how environments that support restoration change people's negatively toned emotions into more positively toned ones reduce physiological arousal, improve specific cognitive functioning, and influence people's action intentions or actual behaviors (Han 2003;Serin et. al., 2018;Memari, Pazhouhanfar, and Nourtaghani 2017).It also explains how these four responses may vary when compared with one another.Still, unfortunately, SRRS lacks diversity, for it only examined the quality of the natural aspect of the environment, just like many of its predecessors.It is a fact that most of the well-known literature on the restorative environment, such as Attention Restorative Theory (ART), Stress Reduction Theory (SRT), and Perceived Restorativeness Scale (RPRS), have significantly favored nature/green settings.And despite the ever-increasing contributions from many sources around built and indoor environments, little to no circumferential theory exists based on the COVID-19 pandemic's exceptional situation and its effects on the overall evaluation of restorativeness (Subiza-Pérez et al. 2021;Bornioli andSubiza-Pérez, 2022, Staats et al. 2016;Kaplan and Kaplan 1989;Hartig 2004;Han 2003;Ulrich, 1991;Ulrich, 1983;Collado et al. 2016).
Furthermore, there is also a concern that most evidence related to the restorative environment comes from Western societies.In contrast, interest in other parts of the world, such as Asian communities has been scarce.According to empirical data, Indonesia is one of the countries with multitudes cases of COVID-19, ranked 3 rd in Asia and 17 th in the world (Worldometers, 2021).A study in Indonesia also reported a prevalence of moderate to extreme psychological disorders (e.g., depression, anxiety, and stress).It was comparable to psychological disorders in China, Japan, the United Kingdom, and Bangladesh (Izzatika, Syakurah, and Bonita 2021).By conducting this research, it can also offer another vital contribution to promote more effective restoration in addition to improving urban public health in Asian communities.Thus, this study aims to understand the restorative experience felt by a group of people in Indonesia towards a particular environment or place during the COVID-19 pandemic.This research will also identify behavioral groups that are formed so it could provide critical insight to tackle similar cases in the future, especially on issues related to the restorative environment during the pandemic.

Materials and methods
This research was carried out by using a qualitative method which involved the use of questionnaires.Qualitative research is commonly used to seek the potential antecedents and factors about which little has been known and explored.One of the approaches for this method is grounded theory.Grounded theory is derived from data systematically gathered and analyzed through the research process (Glaser and Strauss 1967;Strauss and Corbin 1998;Corbin and Strauss 2008).This approach was chosen because of its action-oriented and interaction-oriented inclination to uncover a current phenomenon during the COVID-19 pandemic.
A cross-sectional study was conducted among 263 participants through an online questionnaire from 11 th September to 13 October 2020.The sample selection is non-random sampling in the form of a snowball sampling technique.The questionnaire was distributed through social media to the general public using Google Forms.It was designed in two sections.The first section was designed to collect the attributes of the participant, while the second contained the core questions.The collected data from the first section shows that most participants are women (178 people, 67%), then followed by men (87 people, 33%) in which 187 participants are not married (71%), 3 participants are married without children (1.3%), and 73 participants married with children (27.7%).The age ranged from 12 to 68 years old, with the average occupation being students (42.7%) and workers (47.9%).The level of education varies from junior high school to post-graduate level, with an average income between Rp. 1.5 million -3.5 million per month.Most of the participants live on the islands of Sumatra, with the highest concentration in Palembang City.
After filling in their attributes, participants were asked to answer a series of open-ended and close-ended questions focused on the place they had visited to recuperate during the COVID-19 pandemic (Table 1).There were no boundaries for open-ended questions, so the participants were free to complete the questionnaire according to their respective experiences, as detailed as possible.
The data obtained were then analyzed using content analysis.Content analysis is the commonly used qualitative research technique to analyze written, verbal, or visual documentation systematically.The content is analyzed by breaking it up into conceptual chunks, which are then coded or identified (Wilson 2011).There are three stages in the content data analysis method: 1) Open coding, 2) Axial coding, and 3) Selective coding (Corbin and Strauss 2008).Open coding is the process of identifying a segment of meaning and code-naming to categorize and describe a phenomenon found within the text.Axial coding is the stage when the obtained codes and categories are linked to one another by drawing a connection.The final stage, selective coding, is selecting one category that connects all the codes and captures the essence of the research.
The result was summarized and put into several categories in the form of a contingency table, where the entries are nonnegative measures of association between the row and column entities.Using Correspondence Analysis (CA), these compiled data were analyzed further.CA is a statistical technique proposed by Herman Otto Hartley (Hirschfeld) and later developed by Jean-Paul Benzécri.CA is suited to explore relationships among qualitative variables (or categorical data) because it provides a solution for summarizing and visualizing data sets in two-dimension plots.The results of CA are one or more sets of scale values for the rows and columns, values that have a geometric interpretation leading to visualizations of the similarities between rows and between columns, as well as the row-column associations.(Greenarce, 2015;Mellinger, 1987;Hesier, 2001;Blasius, et al., 2015) Since the focus of this study is the restorative environment, the primary category would be the Places to Recuperate or, as described by the experts, "places to rest and recuperate from mental fatigue" (Kaplan and Kaplan 1989;Korpela 1991;Han 2003).This is the most critical category which connects all categories; therefore, it would be used as the focus of the correspondence analysis.Other sub responding categories which would be paired with this primary category are: 1) Motives; 2) Desired Things; 3) Activities; 4) Level of Concern; 5) Intensity of Visit; 6) Duration of Visit; 7) Choice of Partner; 8) Type of Space; and 9) Proximity.The results of the correspondence between these categories thus would be concluded in the form of a conceptual model.

Places to recuperate
The first stage of analysis began with summarising open coding analysis in tables to prepare for Correspondence Analysis (CA).A cross-tabulation or contingency table is the primary example suitable for CA.However, the method extends smoothly to analyzing almost any table of nonnegative numbers where the relative values in each row and each column are of interest (Greenacre 2015).The first category to be summarized is the Place to Recuperate as shown in Table 2.

Motive
Explain your motivation to visit that place.

Desired Things
What did you look for the most when visiting that place?Activities What kind of activities did you do in that place?

Level of Concern
What was your level of concern when visiting that place?

Intensity of Visit
How many times have you visited that place so far?Choose the one that comes closest.

Duration of Visit
On an average basis, how much time did you spend there?Choose the one that comes closest.

Choice of Partner
With whom did you usually visit that place?Choose the appropriate one.

Type of Space
According to your experience, what kind of space was it?Outdoor, indoor, or a combination of the two?Choose the appropriate one.

Proximity
If measured from where you lived, was that place considered close or not?Choose the appropriate one.

Conclusion
If you were given a choice to visit a brand-new place, what would be the ideal place to help you recuperate during the COVID-19 pandemic?Describe/list your criteria as much as possible.

Motives
The second category to be summarized was the reason for choosing a specific place to recuperate from mental fatigue or the motives.Based on the results, several unique answers are obtained: 1) It is near; 2) Novelty; 3) Suitable for gathering; 4) Not crowded; 5) Want to enjoy nature; 6) Safe; 7) Bored; 8) Fresh air.Some codes, such as wanting to enjoy nature and fresh air, were then put into one big category called natural nuance, while not crowded and safe were grouped into the same category: safety.These were the results mapped in a broad category in Table 3

Correspondence between place and motivation
Pearson's correspondence analysis (CA) was then carried out to explore relationships between these categories and suggest the best model possible to explain said relationships.Correspondence analysis has long been used to visualize a contingency table's association between categorical variables (Sourial et al. 2010).After a test run, a high-significance result with a p-value of < 0.00001 emerged.It showed that there are five clusters divided based on the participants' experiences.
To make interpretation more accessible, some adjustments were made.The differences in colors (red and black) indicated two separate categories: the blackcoded words referred to the primary category, Place to Recuperate, while the red-coded words referred to responding sub-categories.In the case of Diagram 1, the responding sub-category would be Motives.On the other hand, the linking segments represented the correspondence between the primary category (code-red) and responding sub-categories (code-black).It also depicted the behavioral groups formed, which then translated into several clusters describing the similarity found in distinctive characteristics of motives visiting a particular place or environment.
Diagram 1. Correspondence between places to recuperate & motives (p < 0.0001).In the first cluster, participants did not visit purely to recuperate.Other needs or purposes had to be fulfilled, such as buying necessities, doing assignments, and buying things such as food, drinks, etc.This motive made the commercial public place such as malls, supermarkets, traditional markets, and minimarkets have the highest visiting rates.A small proportion of participants also chose formal establishments such as workplaces, campuses, and offices as their prior visits.Participants in this cluster appeared to receive restorative impacts from places they have visited or had previously visited.A sense of familiarity played a big part in constructing this cluster because it not only gives comfort but also enhances wayfinding and reduces wandering and topographical disorientation in individuals suffering from mental fatigue or stress (Lopez, et al., 2020).
The second cluster consists of a large population of participants whose motive was to enjoy the atmosphere, quality, and facilities offered in public spaces such as the neighborhood's open court, jogging track, open green space, amusement parks, sports fields, and public parks.New scenery, comfortable, spacious, and not crowded are the dominant codes in each category.This shows that most participants visited hoping to get away from the boring and mundane daily task of staying home, but at the same time, not wanting to risk going to a place that could not guarantee their safety.
The third cluster comprises places visited by participants who wished to interact with other people -be it family, friends, acquaintances, or even strangers.Policies after policies for social distancing are constantly being updated; some of them can be exceedingly difficult to navigate.Social distancing orders been enacted worldwide to slow the COVID-19 pandemic.People of all ages were encouraged to reduce physical contact to control the infection.Still, many also chose to play it safe and avoid any contact altogether (Luo et al. 2020).This condition made it difficult to find the social interaction needed to create the illusion that we are not alone in this trying time.This could lead to a long period of social exclusion, which is believed to be the main reason for the high prevalence of psychological disorders currently developing in the community (Xiong et al. 2020;Saltzman, 2020).
The second and third clusters appeared to be heavily affected by the stress level resulting from the forced social exclusion.Living in the same place for a long time and losing social and physical contact has often been shown to lead to boredom, frustration, and feelings of isolation from the outside world (Brooks et al. 2020).These negative feelings were then exacerbated by the fact that most participants could not carry out activities that previously could be done without any preamble, such as meeting friends, traveling, going to karaoke, etc.These reasons pushed them to their limits, and in the end, they could not resist venturing out of their safe zone and visiting some precarious places such as attractions and recreational spots, cafes, coffee shops, and restaurants.These types of places are known to be common social stops, which is why they are perfect for those who desperately wished for a semblance of everyday life and real social interaction.
The fourth cluster is very dependent on the natural aspect of the environment.Fresh air, natural settings, and beautiful scenery are the dominant codes of the participants' motives.Parks, beaches, gardens, and green open spaces were considered with medium to high-intensity visits.According to Berto (2005), the natural environment does not require an addiction or special attention, allowing tired people to rest their minds.James (1892) also observed that some aspects of the natural environment are also elementary to attract attention.Along with this basis, Kaplan et al. (1995) also considered the experience of the natural environment to be a very effective means of restoring mental fatigue (Kaplan (1995;Memari et al., 2017).Though the natural environment is considered one of the most compatible places to recuperate, it has a lower probability of being visited in the pandemic because of the many restrictions being laid outwards and how much time is usually spent inside rather than outside.
The last cluster shows that due to the severity of the COVID-19 pandemic, which was not showing any improvement, participants had no choice but to stay at home and avoid any unnecessary contact with the outside world.Some participants did not even have the slightest motivation to visit any place because they were pretty comfortable with their "safe zone" that usually catered to their home or neighborhood.This situation indicated that most considered public space dangerous and unpredictable because they could not control their environment as effectively as they did back home.There was also the concern of crowding.Crowding occurs when the number of individuals exceeds the space available, resulting in adverse health outcomes, such as infectious diseases.For highly contagious agents such as COVID-19, even temporary crowding in public areas, such as parks, trains, and buses, can propagate a chain of infection rates (Von Seidlein et al. 2021).

Other correspondences
After obtaining the results of the correspondence analysis between category Place to Recuperate and Motive, other analysis between the primary category and responding sub-categories (e.g., Place to Recuperate & Desired Things, Place to Recuperate & Activities, Place to Recuperate & Level of Concern, etc.) was conducted.Every analysis shown significant index score, and it was decided that more than one correspondence analysis would be used.The significant result is listed below: Diagra As shown above, the average result of every correspondence analysis has aremarkably high significance with ap-value of < 0.00001.This indicates that place or "place to recuperate," as the primary category (Table 2)

Dependent restorative environment (D-REn)
As the name implies, D-REn is a type of restorative environment with an extremely high dependency rate.D-REn can be further divided into two types.The first type depends heavily on routines, while the second is on interaction.D-REn is usually located close enough, making it easy to be accessed at any time (Diagram 9).The type of space can be varied, but most D-REns are identified as indoor spaces (Diagram 8).a) Routine-Dependent Restorative Environment (RD-REn) RD-REn depends heavily on the participant's daily routine.It usually covers moderate indoor public places such as malls, supermarkets, traditional markets, minimarkets, offices, houses of worship, and campuses (Diagram 1).The type of activity that is commonly shared is shopping or buying necessities (Diagram 3).Things most sought after are generally related to personal or family needs -primary, secondary, or tertiarybut daily needs are most often sought (Diagram 2).The prolonged quarantine had caused a high consumption rate, and foods and other daily needs often run out before their predicted time.Having inadequate basic supplies (e.g., food, water, clothes, or accommodation) during quarantine was proven to be one of the primary sources of stress, pushing people to keep restocking every day.
There are also routines associated with financial responsibilities and needs.This kind of routine is probably linked to economic effects but could also be related to the disruption of social networks and loss of leisure activities.It is possible that many participants kept returning to work despite not being required just because they wanted to feel "normal" again.The level of concern regarding RD-REn's safety (Diagram 4) is neither high nor low because they were part of the participants' almost day-to-day routine, so it was almost like second nature to not care much about whether it was safe or not to keep visiting said places.
The duration of the visit is usually relatively short, around 1-2 hours, for routines that do not require extra attention, such as window-shopping, sightseeing, and shopping.Meanwhile, routines that require extra attention, such as doing tasks or working, usually take a long time, which is more than 5 hours (Diagram 5).The intensity of visits is also relatively high, 6-10 times, and usually, those visits are not carried out alone Social interaction plays a crucial role in maintaining well-being.Yet, one of the significant preventative efforts for reducing the spread of the virus involves social distancing -which in turn diminishes the chance to interact with other people (Saltzman, Hansel, and Bordnick 2020).The sudden loss of social and physical contact would make anyone feel trapped and stressed; this is where ID-REn becomes especially important.It's a type of restorative environment that is mostly known to be common social stops located close enough to home (Table 2; Diagram 8 & 9), which is why places in this type of environment are perfect for those who desperately wished for a semblance of normal life and real social interaction.Because ID-REn heavily relies on social interaction, most visits are rarely made alone (Diagram 7).Even when visiting alone, there is a high chance for interactions, whether prompted or not, mostly with the people working there or strangers.Not only physical kind -said interaction can also occur spiritually, which usually can be found in the interaction between the individual and God in several houses of worship.
ID-REn offers the main selling points of atmosphere, comfort, and various facilities (Diagram 1).Types of activities are chatting, doing assignments or work, and enjoying food and drinks -which are usually also sold as the main attraction (Diagram 3).The concern for visiting ID-REn is relatively high because they are considered public places where anyone can enter freely.In addition, most activities are also carried out while eating or drinking, so the probability of contracting the virus is almost 1:1 chance for every people (Diagram 3).What is most sought after in ID-REn is togetherness, new scenery, adequate facilities, and entertainment while still paying attention personal and social safety (Diagram 4).The duration of the visit is short, less than 1 hour, accompanied by a low intensity of 1-5 visits per place (Diagram 5 & 6).ID-REn seems to be considered "not safe to visit" because of the respondents' deep concern against the high-rate infection.This shows that despite the building stress and desperation of being trapped at home, most people still remembered to adhere to the public health protocol and kept their visiting rate and hour at the bare minimum.Below (Figure 2) is the distribution of ID-REn:

Conventional Restorative Environment (C-REn)
C-REn is the most common type of restorative environment that can be found in many other previous studies, such as Attention Restorative Theory (ART) and Perceived Restoration Scale (PRS) (Kaplan and Kaplan 1989;Hartig 2004;Han 2003) EB-CEn offers entertainment and recreational activities as the main attraction (Diagram 3).This type of environment garners the highest level of concern (Diagram 4).Unlike other types, visiting EB-CEn is actually being encouraged to keep the economy afloatwhich means the chance of crowding is substantially high.In turn, maintaining the recommended distance is almost impossible.The most sought-after is a different atmosphere or new scenery, which refers to the desire to "get away" from the monotonous routine of being quarantined for so long (Diagram 1 & 2).Partners may vary -some came alone, but others came with family, friends, or a significant other just for fun (Diagram 7).The duration of the visit is quite long, 3-5 hours, because it takes time to enjoy every attraction offered (Diagram 6).However, unlike the prolonged visiting hour, the intensity of visits is not too high (Diagram 5).Below (Figure 3) is the distribution of EB-CEn:     b) Nature-Based Restorative Environment (NB-CEn) As the name suggests, NB-CEn is closely related to natural settings.In this case, it can be described as an open space near or far from home abundant leisure activities (Diagrams 3,8,& 9).NB-CEn has a higher frequency rate than its twin type, which almost triples EB-CEn (Diagram 5).It can be assumed that natural environments may be perceived as more restorative than the built environment.According to Berto (2005), the natural environment does not require dependency or special attention to allow tired people to recuperate and restore their energy.This happens because the visual system processes natural stimuli more effortlessly or fluently than built environments.They contain redundant information that makes them easier to process, which may decrease cognitive load and reduce stress (Twedt, Rainey, and Proffitt 2019).This might be why most respondents labeled these places non-threatening, impacting the high intensity of visits (Diagrams 4 & 5).Below (Figure 4) is the distribution of NB-CEn:

Enclosed Restorative Entertainment (E-REn)
E-REn is the last type of restorative environment found during the COVID-19 pandemic.The enactment of various new policies to curb the spread of the disease i.e., social distancing, working, and studying from home, restrictions on large-scale activities, and restrictions in traveling, compel many to adjust.The fear of being infected and, in turn, infecting others forced most participants to hold themselves back from venturing past their doorsteps.Some of them gradually got comfortable with their respective "safe zones," which usually cater to their indoor living space (Diagrams 8 & 9).
Activities usually carried out tend to be very random, like playing with gadgets and relaxing while watching TV -just something to fill the time (Diagram 3).The level of concern tends to be exceptionally low because the chance of contracting or infecting others in this place is minimal (Diagram 4).The most sought-after are rest, togetherness, facilities, and entertainment.Comfort and tranquility are also the most sought-after when spending time in this type of environment (Diagrams 1 & 2).The time spent staying in the D-REn is the longest of all types, consisting of more than 5 hours with an intensity of visits above 10 times (Diagrams 5 & 6).Below (Figure 5) is the distribution of E-REn: The results of the study indicate that there is a discrepancy between the current restorative environment and what has been conceptualized before.To visualize this novelty, a conceptual model will be used.A conceptual model is a model that describes how the causal relationship of elements with principal factors that have been known in a particular problem or phenomenon.Conceptual models can provide perspectives, or points of view used to view a research topic, so they are often used to define concepts and explain phenomena (Ravitch and Riggan 2017;Trochim 2007).
Below is the conceptual model for three types of environments based on users' restorative experience during the covid-19 pandemic (Figure 6).A few key points should be noted to make interpretation easier.First, the size of the circles refers to the frequencies of each type of restorative environment.The bigger the size, the higher the frequency is, and vice versa.Additionally, the exact frequencies can be seen in Figure 1 until Figure 5 .Second, the color differences indicate several types of restorative environments.The detailed agglomeration of each type can be seen in the description at the bottom of the figure.
Last but not least, the position of the information situated around the circles more or less represents a specific characteristic that could be linked to a particular place or environment.The further away the position is, the more general the representation will be.This is because the characters in the outermost circle represent all types of environments and places inside the innermost circle.

Discussion
The relationships between the physical environment and its impact on human physical, emotional, and social aspects have been a research topic for many years.This study was undertaken with one primary objective: to understand the restorative experience felt by a group of people in Indonesia towards a particular environment or place during the exceptional phase of the COVID-19 pandemic.An exploratory qualitative method with the approaches of grounded theory was used to achieve this.In 2013, two studies focused on stress recovery and well-being were documented using the same method and approach.Similar to this study, both created a substantive theory by developing explanatory and theoretical critical concepts with an interpretation model of open coding recommended by Starrin et al. (1991).Even if some parts are different (e.g., data collection technique), the objective between both studies is the same -which is a conceptualized theory (Adevi, et al., 2013;Völker and Kistemann 2013).By adhering to this as a research method, a theory that was grounded to data obtained emerged.This resulted in obtaining the hidden social and collective patterns and identifying several behavioral groups with the emergence of brandnew categorization of restorative environments.These groups can be identified as follows: 1) Dependent Restorative Environment (D-REn); 2) Conventional Restorative Environment (C-REn); and 3) Enclosed Restorative Environment (E-REn).Distribution analysis shows that D-REn has the highest prevalence compared to other types, with a majority of 62.24%, while C-REn ranks second with a plurality of 22%, followed by E-REn with a prevalence of 15.76%.
If previous studies have always stressed on the importance of environments that do not require a dependency or special attention because it allows people to recuperate and restore their energy more easily, the findings of this study contradictorily demonstrate how dependency (which requires extra attention) is highly effective in promoting and permitting the restorative process.This is fascinating because, unlike previous studies that are always partial to the natural settings, this study put an unconventional group of restorative environments as the most popular and sought-after during the COVID-19 pandemic.
Several essential factors limit this study.First, due to the fact that this study was conducted back in 2020 when the situation of COVID pandemic was on the rise, our results and conclusions do not reflect the emotional, physiological, cognitive, and behavioral changes that have taken place since the study was completed.Nevertheless, our study might serve as a critical insight to tackle similar cases in the future, especially on issues related to the restorative environment during the pandemic.Second, this study merely explored the common aspects of the relationship between physical and non-physical qualities of the built environment during the COVID-19 pandemic.Therefore, a measurable and more comprehensive understanding of how particular qualities of the environment can affect the restorative process is suggested.
It is also recommended that some in depth research should be conducted regarding the restorative potential and benefit of each specific environment (e.g., relaxation, engagement, cognitive rest and etc.)At the current phase of this study, all types of restorative environments (D-REn, C-REn, and E-REn) could offer the same restorative benefit or potential, for example, relaxation.Participants who spent their time at home (E-REn) could achieve the same benefit of relaxation as participants who visited gardens (C-REn) or cafes (D-REn).Another example would be engagement.Participants who spent their time at the mall (D-REn) could achieve the same benefit of engagement as participants who went to the office (C-REn) or watched movies in their bedroom (E-REn).This shows a need for further research to specify these benefits and their relation to each type of restorative environment for a deeper understanding of ever-changing environmental and social circumstances.
This study also has a few important practical implications and managerial insight.First, it can be assumed that places that could offer distractions (e. g., D-REn & E-REn) seem more popular during the COVID-19 pandemic.Second, the disruption of social networks and loss of activities have proven to be one of the main reasons why most people choose public places to recuperate rather than the usual conventional restorative environments such as natural settings.Not only it is easily attainable, places like mall, cafes, and offices are able to offer something natural settings cannot -the opportunity to be engaging.People previously sought a quiet, faraway place with minimal distraction to recuperate from stress.But when faced with this kind of public hysteria, there are very few ways to stay sane, and one of the most obvious is to be productive.This can provide valuable insights into designing and managing the restorative environment and public establishment prospects, especially in preparation for future pandemics.
And last but not least, there is an urgent need to make houses more supportive of restoration and psychological well-being.In the context of a pandemic and similar infectious diseases, a residential home does not only function as a place to live but also designated as a quarantine post, working, studying space, and many more.In this condition, a place that initially only functioned as a shelter must undergo a rapid spatial adaptation to accommodate various functions within an unspecified period.Future lines of research should explore how the improved condition at home could alter one's restorative experience.

Conclusion
Based on the grouping pattern of the correspondence analysis, there are three types of restorative environments commonly visited during the COVID-19 pandemic.These types are the result of emotional, physiological, cognitive, and behavioral responses to the participants' restorative experience.These three types of restorative environment can be identified as follows: 1) Dependent Restorative Environment (D-REn), 2) Conventional Restorative Environment (C-REn), and 3) Enclosed Restorative Environment (E-REn).Each type of environment has a unique characteristic that cannot be found in other types.D-REn depends on routine and interaction, while C-REn heavily relies on the power of attraction, especially the value of visual appeal.On the other hand, E-REn exists because of the lack of choice, which makes it unpredictable and out of the box.
. It is usually located far from the hustle and bustle of the city, with a physical tendency to be an open or semi-open space (Diagram 8 & 9).Unlike D-REn, the motive for visiting C-REn is purely to recuperate and unwind (Diagram 1).Visual attraction is also considered the main selling point of this type, which emphasizes the importance of visual appeal (Pasini et al. 2014; Twedt, Rainey, and Proffitt 2019).Based on its main attraction, C-REn can be divided into two types: Entertainment-based and Nature-based.a) Entertainment-Based Conventional Restorative Environment (EB-CEn)

Figure 6 :
Figure 6: The conceptual model for three types of environments based on user's restorative experience during COVID-19 pandemic.

Table 1 .
Excerpts from the questionnaire Name the place you visited the most to recuperate during the COVID-19 pandemic.