Impacts of social distancing on marital life during Covid-19 pandemic

Abstract The aim of this study was to assess the impact of social distancing on different areas of marital life during the Covid-19 pandemic. 1121 Brazilians, who were living with romantic partners, answered an online survey about their social distancing practices during the pandemic and aspects of their relationship. Most participants did not report impairment in marital behavior and 68% did not report experiencing willingness to divorce during pandemic. However, binary logistic regressions showed that the Odds Ratio of reporting willingness to divorce was lower for individuals with longer cohabitation (OR = 0.998), while higher when there was a decrease in positive marital behaviors (OR = 1.8 - 3.13), and also reported an increase in the conflicts, arguments and fights (OR = 6.12 - 6.43). Indicators of higher confinement at home during the pandemic were associated with higher chances of reporting willingness to divorce.

Social distancing is considered one of the most important measures to contain the spread of infectious diseases such as COVID-19.Generally, it can be understood as a set of measures which aims to reduce physical contact among people in a community (Aquino et al., 2020).The definition used in this study was: "the period of social distancing is the time you stopped doing your routine activities due to the Coronavirus pandemic".Social distancing measures have been adopted in all countries affected by the pandemic, causing a ban of public events; closure of schools, commercial establishments, and religious places; the imposition of a curfew; and even restricting the circulation of people on the streets, a measure known as "lockdown" (Aquino et al., 2020).In Brazil, all states adopted social distancing measures since March 2020.Some of them adopted these measures even before having confirmed cases (Silva et al., 2020).Even though they are important to prevent the spread of the disease (Ayuso et al., 2020;Bo et al., 2021;Flaxman et al., 2020), social distancing measures have side effects on all sectors of the economy (Nicola et al., 2020) and also have a social harm, especially for the most vulnerable populations (Rocha & Pires, 2020).Thus, given the importance of these measures in controlling the pandemic, it is essential to identify their possible side effects, which can contribute to the planning of psychosocial and socioeconomic interventions for their mitigation.
Considering that marital life is influenced by the context in which the spouses are inserted (Walsh, 2016), it is supposed that economic and social consequences of the pandemic, associated with the imposed confinement on the families, would present effects on couple life.In a Spanish study (Ayuso et al., 2020), 49% of participants stated that family relationships improved during confinement.Among those who reported a worsening in relationships, people aged between 25 and 34 were the most affected, which is the period of life considered by the authors as the most important for the consolidation of life projects.Elderly people reported less family damage, while participants who lived in large cities reported greater relationship impairment compared to those who lived in small towns.The study also reported an increase in difficulties to balance work and household chores, and an increase in domestic violence against children.Thus, confinement imposed by the pandemic and the other related consequences of social distancing measures have the potential to impair families, as well as to contribute positively to relationships.
The existence of different outcomes can be explained by the several factors that contribute to the coping and/or the adaptation to stressful situations.The Vulnerability Stress Adaptation (VSA) model, developed by Karney and Bradbury (1995), gathers concepts from different theoretical frameworks in order to understand what influences marital relationships' success over time.This model predicts that three factors are of importance to marital quality and stability: a) durable or persistent vulnerabilities (i.e., educational, personality, and psychopathology aspects of the spouses); b) stressful experiences (i.e., unemployment and health issues); and c) adaptative processes (i.e., interactions between spouses to deal with daily life problems).When applied to the context of the COVID-19 pandemic (Pietromonaco & Overall, 2021), this model forecasts different consequences in the couple's life, depending on the spouses' psychological, relational, and social conditions.For instance, Balzarini et al. (2021) observed that for 57 couples from countries (including Brazil), loneliness, stress and financial strain associated to the pandemics seem to affect marital quality over time.However, the effect of these stressors was moderated by perceived responsiveness, which represents how much a person feels understood, validated, and cared by their partner.
In addition to perceived responsiveness, Delatorre (2019) proposed that marital quality can be measured by five dimensions: 1) satisfaction, 2) commitment, 3) intimacy, 4) attraction and sex, and 5) affection.In fact, the expressions of positive affection in couple interactions seem to be an important predictor of marital satisfaction over time and in stressful situations.This applies to affectivity expressed during the couple's verbal interactions (Johnson et al., 2005), in daily life (Caughlin & Huston, 2002), as well as non-verbal expressions of affection, including kissing (Floyd et al., 2009) and other affectionate touches performed by couples (Jakubiak & Feeney, 2017).On the other hand, negative communication behaviors can perpetuate marital conflict and lead to divorce (Birditt et al., 2010).In the Brazilian population, relationship with their children, time spent together, money, household chores, and sex are the most common reasons for marital conflicts (Mosmann & Falcke, 2011).Confinement during the pandemic could contribute to the resolution of some of these conflicts due to the increasing time that spouses spend together, but it could also exacerbate conflicts regarding children's education and household chores, for instance.Therefore, confinement during the pandemic could provide benefits or impairments to marital life.
Relationship longevity seems to be associated with its stability (Rosenfeld, 2014).Although, the associations between longevity and marital satisfaction are more complex, since satisfaction can decline over time (Lavner & Bradbury, 2010) or even have different paths for men and women.For instance, in a cross-sectional study of approximately 1000 couples from five different countries (including Brazil), Heiman et al. (2011) observed that the probability of being in a happy relationship seems to decrease for women in relationships of up to 15 years, increasing in relationships of 20 years or more.On the other hand, the probability of reporting a happy relationship is progressively higher with increasing marital longevity for men.However, the relationships' duration has great collinearity with the age of the spouses, making it difficult to separate their effects in most studies (Lyngstad & Jalovaara, 2010).
Better comprehending the side effects of social distancing on marital life is essential for the development of effective intervention strategies for its mitigation.Investigating the impacts of this measure might also offer a better insight on the effects of acute or chronic stressors on these relationships, preparing couples to better cope with future crises.Thus, in this study, we sought to understand the influence of time of cohabitation, social distance (duration and intensity), and interactions of the couple (affection, conflicts, arguments and fights, and sexual relations) on the willingness to divorce in Brazilian couples who lived together during the data collection period, in the COVID-19 pandemic.We hypothesized that: 1) the longer the cohabitation period, the lesser would be the willingness to divorce during the pandemic; 2) The higher the rates of social distancing, the greater the willingness to divorce; and 3) people who reported an increase in conflicts, a decrease in affection and in the frequency of sexual intercourse, would be more likely to have willingness to divorce.

Study design
The study design was quantitative, a cross-sectional online survey with the Brazilian population.

Participants
The sample was composed of 1121 people aged 18 to 80 years old (M=35.8;SD=11.4), from 21 Brazilian states and the Federal District.The inclusion criterion was to be in a marital relationship (cohabiting with a romantic partner), regardless of marital status and time of cohabitation.Therefore, relationships with or without civil or religious registration, heterosexual, homosexual, and others were included.Even though a sample calculation was not performed, the recommendation of at least 500 participants for logistic regression analysis was followed (Bujang et al., 2018).The initial sample size was 1126 participants, however, five were excluded because they did not answer all of the questions for this study (check below).

Instruments
This study is part of a broader project on conjugality during the pandemic, which collected several data on the couple's life during this period.First, the participant received the orientation: "Think about your romantic relationship with your partner during social distancing in the pandemic and mark the scale below depending on how you have been perceiving the following aspects".
Five variables about marital life and three variables about social distancing were included in this article: "Conflicts/arguments/fights between you and your spouse"; "Affection manifestations"; "The frequency of sexual intercourse"; "Willingness to divorce (during the pandemic)".These questions were answered based on a 5-point Likert scale, with the value 1 referring to "decreased a lot", 3 referring to "neither increased nor decreased" and 5 referred to "increased a lot" in the first three questions.For the question about willingness to divorce, number 1 referred to "not willing to divorce" and number 5 referred to "willing a lot to divorce".The cohabitation time variable was informed in years and months and later transformed into months.The Social Distancing Time was reported in months and days and then turned into days.Two variables were used to verify the intensity of social distancing: "How much did you stop doing outdoor activities during the pandemic" (1 = "I didn't stop doing anything"; 5 = "I stopped doing all outdoor activities I used to do") and "How many times a week did you leave your house during the pandemic" (8-point scale, considering 0 = None, up to 7 = 7 days a week).
The questions for this research were based on previous studies: the variables of positive marital behavior (affection and frequency of sexual intercourse) were chosen from the marital quality indexes proposed by Delatorre (2019); the variable "Conflicts, arguments and fights" was chosen because it is also an important factor for marital satisfaction (Mosmann & Falcke, 2011), and the variable "Willingness to divorce" because it is related to an impairment of the relationship (Birditt et al., 2010).

Data collection
Data were collected between June and July 2020, through a cross-sectional online survey.The sampling method was by convenience and the participants were asked to indicate other possible participants, consisting of a snowball sampling technique (Robson & McCartan, 2016).The initial disclosure was made through the researchers' contact network, through social media and e-mails.

Data analysis
All variables were recoded for analysis since the use of the Likert scale to measure conflict, affection, and frequency of sexual intercourse, presented categorical options for the participants which cannot be interpreted as linear or ordinal variables.In these cases, data were recoded into dichotomous variables, in order to answer the following questions: "Have conflicts, arguments and fights increased during social distancing?" (scale values 4 and 5 = Yes, n = 235); "Have expressions of affection decreased?"(1 and 2 = Yes, n=173); "Has the frequency of sexual intercourse decreased?"(1 and 2 = Yes, n=337).For the willingness to divorce, participants were divided into 2 groups: "Willing to divorce" (scale values 2 to 5, n=355) and "Not willing to divorce" (scale value 1, n= 764).Table 1 shows the operational definitions of the variables included in the analysis.
In the tested models, we sought to predict the willingness to divorce from the self-report of these three marital behaviors: increase in conflicts, arguments and fights, decrease in affection, and in the frequency of sexual intercourses.In addition, variables that measured the time of social distancing (in days) and the intensity of social distancing (weekly outdoor activities and decreased activities outside home) were included.
Data were analyzed using binary logistic regressions.This regression analysis is used to assess the relationship among several independent variables and a dependent variable, when the dependent one has two levels (Forthofer et al., 2007).Jamovi software version 1.6.7.0 was used for the analysis.

Ethical procedures
As determined by Brazilian legislation (Resolutions 466/2012 and 510/2016 of the Health Ministry) all ethical procedures were followed.Before participants could answer to the online questionnaire, the Free and Informed Consent Term was presented to them, and they could only proceed after declaring their agreement.The project was approved by the Research Ethics Committee (Protocol number 4.063.196).
Table 1.Report of the time the participant adopted the social distancing measures recommended by the authorities during the pandemic Decreased outdoor activities How much the participant reports having stopped doing activities that he did outdoors before the pandemic.

Weekly outdoor activities
How many days a week did the participant leave home for any type of activity during the pandemic.

Results
Data show that most participants did not report impairments in marital behavior: there was no increase in marital conflicts (for 79% of the participants), nor a decrease in expressions of affection (84.3%), or even a reduction in the frequency of sexual intercourse (69.7%).The majority (68.3%) reported not having thought about divorce during the confinement imposed by the pandemic.Data that present the sample characteristics are available in Tables 2 and 3.
In order to address the study hypotheses, three independent models were tested in binary logistic regression, since a single model would lead to excessive collinearity among the social distancing indexes.For the independent models, the obtained FIV were between 1.01 and 1.29, and tolerance between 0.776 and 0.994.These values can be considered adequate, based on the presence of FIV indexes lower than 10 (Forthofer et al., 2007) and tolerances higher than 0.2 (Menard, 2002).
All models included the variable time of cohabitation and variables related to marital behavior as predictors.4 shows the model fit measures of each model.The explained variance (R² N ) ranged from 32% to 34.1%, and all models were significant.
Table 5 shows the results of the three models tested.Time of cohabitation was a significant predictor in all models (OR=0.998),indicating that each month of cohabitation decreases the probability of reporting willingness to divorce during the pandemic by 0.2%.This indicates that for each year of cohabitation, the chance of reporting willingness to divorce during the pandemic decreases by 2%, and for every 10 years the chance is 18% lower (logistic calculation performed using the formula: y=eβ .Δ, considering e=Euler's constant, β = Estimate and Δ=number of months).This result corroborates hypothesis 1 of this study: the longer the cohabitation period, the lesser would be the willingness to divorce during the pandemic.
The second hypothesis about higher rates of social distancing would be predictors of higher rates of willingness to divorce was partially confirmed.Table 5 shows that time of social distancing was not a significant predictor in Model 1.On the other hand, the perception of not having outdoor activities had a significant effect, increasing by 1.23 times (23%) the chance of reporting willingness to divorce for each point on the Likert scale (1 to 5).Therefore, people who reported having suspended all outdoor activities and scored 5 in this variable are 2.27 times more likely to report having felt some level of willingness to divorce than those who had not stopped doing any outdoor activities.The weekly outdoor activities variable also had a significant effect on the model with an increase of 7.6% in the chances of reporting willingness to divorce for each weekly day of confinement.Thus, those who went out home 7 times a week during the pandemic would be 42% less likely to report willingness to divorce than those who stayed at home every day, for instance.
Regarding marital behaviors, hypothesis 3 was also confirmed.The increase in conflicts, arguments and fights was the variable with the greatest weight in all three models.Participants who reported an increase in conflicts, arguments and fights had 6 times higher odds of reporting willingness to divorce in relation to those who did not report an increase in conflicts (Model 1: 6.12; Model 2: 6.43; Model 3: 6.32).The decrease in affection increases the chance of reporting willingness to divorce during confinement by approximately 3 times (Model 1: 3.11; Model 2: 2.96; Model 3: 3.13).Decreased frequency of sexual intercourse increases the chance of reporting willingness to divorce by around 1.8 times (Model 1: 1.8; Model 2: 1.76; Model 3: 1.74).Accuracy, Specificity, and Sensitivity of the tested models are available in Table 6.

Discussion
The study results seem to corroborate previous findings in the literature regarding the importance of marital satisfaction for long-term relationships (Røsand et al., 2014;Lawal & Okereke, 2020), especially in stressful events.News regarding the worsening of marital relationships and the increase in the number of divorces during the pandemic was widely disseminated in the media, however the findings presented here demonstrate that this outcome cannot be generalized to most people.On the contrary, most participants did not report relationship problems or willingness to divorce during the pandemic, reinforcing the findings from  other authors who identified high levels of family resilience during this challenging period (Ayuso et al., 2020).This result is theoretically corroborated since marital quality does not depend only on contextual variables, such as social distancing during the pandemic, but it is also related to the spouses' personal characteristics and the adaptive processes in stressful situations (Delatorre, 2019;Karney & Bradbury 1995).Thus, the lack of reports of negative marital outcomes for most couples during the pandemic may be connected to several variables, which contributed to the stability or instability of these relationships.
The report of increased conflicts, arguments and fights, which are associated with dissatisfaction in a relationship, was the variable that most increased the probability of reporting willingness to divorce during the pandemic.The importance of this variable in the models was not unexpected, considering that the frequency of conflicts is a predictor of the marital relationship quality in different contexts (see Kluwer & Johnson, 2007;Uhlich et al., 2022).However, theoretical assumptions (Kally et al., 2003) and previous empirical findings (Birditt et al, 2010) both support that marital satisfaction also depends on the ability of the spouses to constructively solve their conflicts, especially in stressful events.Therefore, this ability may also be one of the factors that differentiate participants who perceived marital impairment during the pandemic (to the point of reporting a desire to separate) from those who did not.
The results on the impact of decreased affection on the willingness to divorce corroborate findings of previous studies that have associated marital quality with the spouse's perception of affection and responsiveness during the pandemic (Balzarini et al., 2021).This result also converges with theoretical assumptions about affection as a component of marital quality (Delatorre, 2019).Findings of this study about higher chance of reporting willingness to divorce in the face of a decreased frequency of sexual intercourse reaffirm what several studies have shown regarding the importance of sexual intercourse for the quality and stability of marital relationships (Rocha & Fensterseifer, 2019;Schoenfeld et al., 2017;Fallis et al., 2016).
The association of long-term relationships with a lower chance of reporting willingness to divorce during the pandemic was also found in previous studies, for both heterosexual and homosexual couples (Rosenfeld, 2014).This lower probability does not necessarily mean that long-term couples have better marital quality.Actually, there is evidence pointing to the opposite, demonstrating that marital satisfaction declines over the years (Lavner & Bradbury, 2010).One of the possibilities that would explain the reduction in the divorce probability over time is that the material and emotional impairments of divorcing become greater as time passes, especially when the couple has children.However, the first years are crucial for marital life, and couples that survive each stage of life acquire more stability to live the subsequent years (Rosenfeld, 2014).The same logic can be applied for the pandemic time.Couples with longer cohabitation may report less willingness to divorce because of the higher losses of this dissolution, or because they have more personal and marital resources to cope with the challenges imposed by social distancing.
Rather than the time extension of social distancing, the association of the intensity with greater odds of reporting thoughts about divorce during the pandemic is surprising, at some level.As Pietromonaco and Overall (2021) highlight, catastrophes that produce chronic stress seem to be associated with worse relational outcomes when compared to catastrophes with acute effects.Therefore, low levels of social distancing would not produce sufficiently high levels of stress to impair the stability of marital relationships, even when the distancing is prolonged.On the other hand, the association of high levels of social distancing with greater marital instability seems to converge with previous findings about the protective role of social support in relationships (see Steiner et al., 2010;Chi et al., 2011), since this support is probably impaired by the need for  (Karney & Bradbury, 1995;Pietromonaco & Overall, 2021).
Given the importance of social distancing measures in controlling the pandemic (Ayuso et al., 2020;Bo et al., 2021;Flaxman et al., 2020), a better understanding of their adverse effects (Ford, 2021;Zolnikov & Furio, 2021) is essential for planning measures and the adoption of mitigation strategies.Considering the catastrophic effects of the uncontrolled pandemic, including the overload in health systems; deaths which could have been prevented; the family stress resulting from the risk of contamination; and the loss of beloved ones, this planning turns out to be essential.
Regarding practical implications of the present study, the results contribute to a deeper understanding of the consequences of social distancing measures on marital relationships.Even though it is sorely necessary to mitigate the spread of COVID-19, social distancing has side effects on couple life, since it was possible to verify that more intense levels of home confinement were associated with a greater probability of reporting willingness to divorce.This result suggests that, during the pandemic, couples need to be aware of their relationships and invest in positive behaviors toward their partner, in order to keep the relationship healthy.Among the observed marital behaviors, the decrease in the frequency of sexual intercourse and the decrease in affection were important risk factors, suggesting that stimulating these aspects can be beneficial for the stability of the relationship.With the increase in time spent together during confinement, attempts to encourage these behaviors may require some creativity.However, this is a justified effort, since dissatisfaction in these areas can trigger disagreements and increase marital tension.The increase in conflicts, arguments and fights was the variable with the greatest impact on the willingness to divorce, which demonstrates the importance of positive conflict resolution strategies in marital life (Delatorre & Wagner, 2016).
In common sense, it is thought that the skills needed to have a successful married life are learned naturally, however studies show that it is possible to teach these skills and contribute to the improvement of relationships (Lesch et al., 2018;Halford & Petch, 2010).In Brazil, there is a psychoeducational program called "Living as Partners Program" (Wagner et al., 2015), which was developed with and for the Brazilian population.It essentially proposes psychoeducational workshops to optimize marital health in several areas.In addition, it is important that new interventions come along, in order to promote better marital quality, especially in times of crisis such as the current pandemic.According to the results of the present study, conflict resolution skills and the planning of moments of affection can be important mechanisms to be addressed by these interventions.
It is necessary to consider some limitations when interpreting the results of the present study.Firstly, the sample was constituted by participants with a high level of academic education.For example, the number of participants who had completed higher education (69.2%) was approximately 4 times higher than the national index (17.4%)for people aged 25 or more (IBGE, 2019).This aspect should be considered with caution, since higher levels of education lead to lower divorce rates, even though different effects have been reported depending on the country and the existence of differences in the educational level in the couple (Lyngstad & Jalovaara, 2010).It is also important to consider the limitations of the cross-sectional study design, which does not allow a comparison between the situation of the couple before and during the pandemic.Thus, the questionnaire data largely depend on the subjects' estimation of changes in the relationship compared to the previous period.Regarding the used questionnaire, the survey was composed of simple questions to make rapid data collection possible in the initial period of the COVID-19 pandemic, and more complex scales were not used to measure the variables, something that could increase the validity of results.Finally, the study addressed the side effects associated with social distancing as performed in Brazil, but not all possible mediating variables behind these effects were addressed.A good part of them may be associated with socioeconomic variables that were not controlled in this study, since they depend on the analyzed regional and national context and if the public policies adopted them or not for their mitigation.Despite these limitations, the present study contributes to a better understanding of the effects of major environmental stressors on marital relationships.In particular, it helps to comprehend the important effects of social distancing measures.With the support of professionals and organizations, its consideration can make it easier for couples to adopt attitudes that preserve marital quality.
We suggest for future studies to address the variables included in this study longitudinally, allowing more robust inferences regarding the possible causal correlation among them, in the context of the pandemic and similar environmental stressors.In addition, it would be important to further investigate the effects of more and less restrictive and/or prolonged forms of distancing, including more specific indexes of intensity (e. g. school closures, working from home), and their possible interactions with the duration of these measures.This could also shed light on the differences between acute and chronic stress associated to its effect on marital relationships.Lastly, the inclusion of measures of socioeconomic variables, social support, and the subjective experience of individuals regarding distancing (e. g. feelings of loneliness, burden, fear) could help to clarify the most important mechanisms behind the stressful effects of social distancing.
Variables Definitions Willingness to divorce Report of having experienced the desire to divorce during the pandemic Cohabitation time How long the participant reports living together with the current partner Increase in conflicts, arguments and fights Report of increased marital conflicts, arguments and fights with the partner during the pandemic.Decreased affection Report of a decrease in expressions of affection between spouses during the pandemic Decreased frequency of sexual intercourse Report of a decrease in the frequency of sexual intercourse between spouses during the pandemic Distancing time

Notes.
Estimate represents the natural logarithm of the odds "Willingness to divorce = Willing to divorce" vs. "Willingness to divorce = Not willing to divorce".* In months.a The inclusion of the value 1 in the confidence interval results from the rounding of the fourth decimal number.b In days.SE = Standard Error.Z = Z score calculated from the ratio between the regression coefficient and the standard error.p = Statistical significance.OR = Odds Ratio.

Table 2 .
Characteristics of the sample (categorical data)In addition, one of the social distancing indexes was included as a predictor in each model: time of social distancing, decrease in outdoor activities, and weekly outdoor activities.Table

Table 5 .
Logistic regression predicting Willingness to divorce: Three models with different indexes of social distancing

Table 6 .
Accuracy, Specificity, Sensitivity, and Cut-off point of each model