The literature on the effects of telework outside the COVID-19 pandemic has revealed contradictory findings on employees’ well-being and job performance, either positive or negative (Gajendran et al., 2015; Boell et al., 2016; Charalampous et al., 2019; Golden and Gajendran, 2019; Vayre, 2019; Wang et al., 2019, 2020). Some studies have examined the positive effects of telework and identified that teleworkers are less interrupted and have more concentration on the task (Vittersø et al., 2003; Windeler et al., 2017), and all these factors contribute to increased job satisfaction, employees’ well-being and job performance (Gajendran et al., 2015; Delanoeije and Verbruggen, 2020). It can be noted that studies on the effects of telework have focused mainly on well-being and very few on performance. In a field study (Vega et al., 2015), teleworking was also found to have positive implications on ‘objective’ creativity (e.g., finding an innovative solution in the context of urban development). Similarly, in an experiment aiming to examine differences of performance between repetitive and creative tasks performed remotely, Glenn Dutcher (2012) showed that teleworkers increased their performance in creativity tasks only (e.g., finding as many unusual uses for ordinary objects as possible). Some explanations have been provided as to why employees are creative in remote working, such as greater concentration on the tasks (Biron and van Veldhoven, 2016), and thinking alone and independently, two necessary conditions for producing novel ideas (Nouri et al., 2015). Teleworking also increases the employee’s sense of control over their working time, with greater flexibility in the way they organize their work (Kossek et al., 2006), less travel time and thus overall a better work-life balance (Golden et al., 2008; Anderson et al., 2015; Grant et al., 2019). The sense of autonomy is higher and increases motivation, organizational commitment, and job satisfaction (Charalampous et al., 2019; Grant et al., 2019). Overall, telework can be seen more as a resource, both for the company to support job performance (Gajendran and Harrison, 2007), and for the workers to enhance their work engagement (Masuda et al., 2017) and well-being (Anderson et al., 2015).

Other studies have examined the negative effects of telework, and at least three risk factors for employees have been identified. Firstly, it increases pressure on teleworkers regarding constant connectivity and responsiveness. It results in intensifying work, distracting from the family members and creating work-family conflict (Hammer et al., 2005; Allen et al., 2015; Derks et al., 2016; Grant et al., 2019). Telework can also lead to higher stress and depression (Song and Gao, 2020; Afonso et al., 2021), technostress (Suh and Lee, 2017; Vayre and Vonthron, 2019; Ghislieri et al., 2022), or workaholism (Vayre, 2019; Song and Gao, 2020). Secondly, due to a greater physical distance from teamwork, a feeling of social isolation, alienation and loneliness can emerge (Zhou, 2018; Charalampous et al., 2019). The need-to-belong theory (Baumeister and Leary, 1995) can explain how teleworkers’ physical and psychological isolation may reduce their affective connections to their coworkers (Wang et al., 2020). There can also be a lack of support and visible leadership, and there may be less social interaction when isolated and detached from the workplace. The perceived distance of teleworkers from their team can have a negative impact on performance and can increase job dissatisfaction (Golden et al., 2008; Ozcelik and Barsade, 2018). Remote employees should initiate frequent communications with their coworkers, and organizations have to provide adequate social support by providing an enabling organizational context (Ruiller et al., 2019). However, although remote workers perceive themselves as having more autonomy and more flexibility thus reducing their stress, at the same time, they can feel socially isolated and disconnected from their peers and supervisors, leading to an ambivalent impact on well-being (Ruiller et al., 2019; Pulido-Martos et al., 2021). Finally, another risk of telework concerns the difficulty for the teleworker to find his or her own work organization and to motivate himself or herself in the absence of formal teamwork. The use of information and communication technologies also leads to counterproductive behavior, such as cyberslacking, i.e., the voluntary use of the Internet for non-work-related purposes during working hours (O’Neill et al., 2014b).

To summarize, the effects of telework on employees seem to be inconsistent or even divergent. Several factors can explain the inconsistent effects of telework (Oakman et al., 2020): the intensity of telework, characteristics of the telework environment, the nature of work (e.g., task interdependence, autonomy), the organizational culture and leadership management, the technical and social supports (e.g., supervisor support, colleagues support), social relationships external to work (e.g., family support, friends support) and the nature of outcomes studied (e.g., job satisfaction, mental health, self-rated or objective performance, productivity or creative performance). As the COVID-19 pandemic imposed telework on all, it also seems important to know the profiles of teleworkers, notably to identify those at risk, in order to manage better the implementation of WFH. To date, we ignore which psychological profiles of employees benefit or suffer from teleworking. Individual differences or personality characteristics of teleworkers could contribute to explaining the inconsistent effects observed, as “one size does not fit all” (Charalampous et al., 2019, p. 69).

Although most studies have focused on situational or organizational variables that influence the effects of telework, some research has tried to identify individual differences that might influence the acceptability and the effects of telework on employees’ well-being and performance (Allen et al., 2015; Anderson et al., 2015; Donati et al., 2021; Wang et al., 2021).

Concerning the effects of personality traits on employees’ well-being during telework, some personality variables related to the Big Five, such as agreeableness, openness to experience, or consciousness appear to be positively related to telework acceptability and employees’ well-being, while mixed-results were found for extraversion and neuroticism. A high level of openness to experience would thus moderate the positive correlation between telework and positive emotions, making it more pronounced (Anderson et al., 2015). The authors explain this by the creative, curious, eager for novelty and change aspects of this personality type, which would allow employees to cope better with the many adaptations and flexibility that telework requires (Anderson et al., 2015). During WFH conscientious people are able to set non-immediate goals, and are well organized, self-disciplined and autonomous at work (Wilmot et al., 2019; Kniffin et al., 2021). Concerning the effects of extraversion and neuroticism traits on employees’ well-being during WFH, the findings are not consistent. In general, extraversion is associated with lower work-stress and higher job satisfaction (Udayar et al., 2020). Thus, individuals who are more sociable and attracted to developing relationships with others tend to report more adaptability to telework and less deleterious effects of social isolation (O’Neill et al., 2009; Anderson et al., 2015; Wilmot et al., 2019). Recent studies during the COVID-19 pandemic revealed that extraversion is associated with more active seeking of socioemotional support and adaptative strategies of coping (Volk et al., 2021). On the contrary, a study conducted in Norway with police officers did not find a significant association between extraversion and general satisfaction with the home-office arrangement (Langvik et al., 2021). Another study with Canadian adults recently showed that both higher neuroticism and extraversion scores were associated with high levels of stress during the pandemic (Liu et al., 2021). In a recent longitudinal study, Entringer and Gosling (2021) showed that loneliness increased significantly compared to pre-pandemic levels, and to a greater extend for women and younger people, but also for extraverted, neurotic and conscientious individuals. Thus, the effects of extraversion on the acceptability of WFH and its effects on employees’ mental health need further investigation. Concerning the effects of neuroticism on teleworkers’ well-being, the results are also inconsistent. Some studies suggest that the deleterious effects of telework on mental health are greater with higher levels of neuroticism (Wilmot et al., 2019). Employees who are more emotionally stable are more able to establish good interpersonal relationships with others and to capitalize on positive emotions. By contrast, other studies suggest that individuals high on neuroticism reported positive attitudes toward remote work (Clark et al., 2012), possibly because they saw this as an opportunity to exert less effort and as a way to avoid difficult relationships with others in the workplace. Thus, studies on the effects of personality traits on attitudes to telework and on employees’ well-being do not always give consistent results.

Concerning the effects of personality traits on employees’ performance during telework, very few studies have investigated the influence of certain personality traits or individual differences (O’Neill et al., 2009; Allen et al., 2015; Wang et al., 2021). However, these studies have generally not focused on ‘objective’ performance of employees, or notably on creativity, except in recent studies conducted during the COVID-19 pandemic which reveal that personality and individual difference related to the preference for solitude (Burger, 1995) may moderate the effects of social isolation on performance on creative tasks during lockdown (Michinov and Michinov, 2021). In a study carried out in France during the first lockdown, Michinov and Michinov (2021) showed that the level of loneliness, stress and anxiety varied according to the level of preference for solitude and certain traits related to the Big Five (extraversion, neuroticism, openness). Individuals with an “Affiliative” profile (i.e., individuals with low preference for solitude, extraverted and emotionally stable) as well as an “Emotionally Unstable Lonely” profile (i.e., individuals with high levels of neuroticism and openness, and moderate level of extraversion) expressed high stress and anxiety during lockdown, and the latter generated a greater number of creative ideas than the former. By contrast, those with an “Emotionally Stable Lonely” Profile (i.e., individuals with a high level of preference for solitude, low levels of extraversion and neuroticism) expressed a lower level of loneliness during lockdown and performed better on a creativity task requiring a correct solution to be found through insight. However, these data have been observed on the general population, and deserve to be studied on a specific population of teleworkers.

Thus, it appears that not all people are equal when faced with social isolation and distance from peer groups, whether friendly or professional groups. Moreover, prior studies were conducted with individuals from the general population and did not specifically examine a sample of employees working from home during the COVID-19 pandemic. In this perspective, it seems necessary to extend research to: (1) examine the relationships between the intensity and familiarity with telework, employees’ well-being, employees’ creativity and personality variables; (2) identify psychological profile of employees at risk from social isolation during mandatory WFH, and (3) examine if distinct personality profiles affect employees’ well-being and creativity during mandatory WFH.

Based on research demonstrating the moderating role of personality variables on acceptability of telework and performance (O’Neill et al., 2009; Allen et al., 2015; Anderson et al., 2015; Donati et al., 2021; Wang et al., 2021), and the role of personality variables and individual differences on individuals’ mental health and creativity during lockdown (Entringer and Gosling, 2021; Götz et al., 2021; Liu et al., 2021; Michinov and Michinov, 2021; Volk et al., 2021), we expected that personal characteristics of teleworkers would affect well-being and creativity during mandatory WFH. More specifically, we could expect the effects of WFH to be related not only to the characteristics of WFH (experience or intensity) but also to teleworkers’ psychological profiles based on their preference for solitude and personality variables related to the Big Five dimensions (Hypothesis 1). Although it is difficult to determine a priori the psychological profiles that can emerge from a sample of teleworkers, based on a recent study by Michinov and Michinov (2021) on a wider population than those of employees, we could expect that “Affiliative” and “Emotionally Unstable Lonely” individuals should express higher loneliness at work and stress, but less job satisfaction and work engagement, and the latter should produce a greater number of creative ideas than the former those with an “Emotionally Stable Lonely” profile (Hypothesis 2).

Participants completed an anonymous online survey using the LimeSurvey software (Limesurvey GmbH, Limesurvey GmbH), after having read the written consent form and explicitly agreeing to participate. The study was conducted in accordance with the Declaration of Helsinki. Ethical approval was granted by the Ethics Committee of the University of Rennes 2 (No 2021-030). There was no monetary or credit compensation. The survey was shared via social media from November 13th to December 15th 2020 during the second lockdown in France.

Participants were recruited via snowball sampling. The authors of the article and students from Masters in work and organizational psychology, and human resources management posted an advertisement to recruit participants on their social media profile and disseminated the advertisement directly to personal and professional contacts working in diverse occupations. This survey was entitled “My experience of teleworking,” and was addressed to teleworkers over 18 years old living in France. A total of 1,155 responses were collected, and 209 responses were deleted because of incomplete responses to the scales (108 returns with no response for the stress, well-being and creativity scales, 101 with no response for the creativity scales or socio-demographic data). The final sample therefore included the responses of 946 participants. The valid response rate (complete data) was thus 81.90%.

Participants were 73.7% females, and they had a mean age of 35.4 (SD = 11.5). The majority were married or with a partner (69.9%), 55.71% had no children, 13% one children, 24% two children, 6.6% three children, and 0.7% more than four children. Most of them had a bachelor’s or master’s degree (90.4%). Most of them had a full-time job (87.6%) and a permanent contract (74.9%). On average, their organizational tenure was 4.83 years (SD = 6.04). They came from diverse occupations, for example, banking, insurance, finance (13.3%), health and social work (11.6%), education (9.7%), the industrial sector (9.5%) and business services (9.5%). Among the participants, 48.2% worked in large organizations (300 or more employees), 18% in medium organizations (101–300 employees), 18% in medium-small organizations (21–100 employees) and 14.8% in small organizations (1–20 employees).

A majority of respondents were teleworking 5 days per week (49.2%), with the rest working 4 days (15.8%), 3 days (12.9%), 2 days (12.4%) or 1 day (6%) per week. Most of them reported that telework was not a chosen situation (76.8%) and 51% declared that telework had been formalized by a contractual agreement during the COVID-19 lockdown. Outside this period of lockdown, the experience of telework was varied among respondents (38.5% had never teleworked, 32.8% rarely, 21.4% often, 7.4% very often), and we distinguished between high and low experience teleworkers. The most common telework space was at home with a dedicated room or office (52.2%) and for 47.6% at home without a dedicated room. The equipment was provided entirely by the employer (47%) or only in part (37%). The vast majority of participants said they had help from colleagues or a helpdesk in case of technical problems (91.9%).

The questionnaire consisted of brief measures to be administered in large online surveys which are known to have good validity and reliability across different samples. All the scales had already been largely used and validated in French, except for the loneliness at work scale. For this scale, the items were adapted using a back-translation procedure. The first author translated the scale. Masters’ students in Psychology and Organizational Management also reviewed each item of the questionnaire to ensure the face validity. A bilingual professional translator then checked the scale and ambiguous or complex terms were either removed or rephrased. The details of items used are presented in Supplementary Material.

The data analyses were performed using Jamovi (The Jamovi Project, 2021) and the ‘Machine Learning’ module (Clustering) of the JASP software (JASP Team, 2022) from the datasets available in our OSF project (Michinov et al., 2021). The descriptive statistics of the quantitative variables were reported as frequencies and percentages. Relationships between study variables were examined with Pearson’s correlations. Multiple regression analyses were conducted to examine the relative contributions of control variables, preference for solitude and personality variables on indicators of well-being (i.e., loneliness at work, stress, job satisfaction and work engagement) and creative performance (i.e., ‘subjective’ and ‘objective’ creativity).

To explore the data thoroughly, we used a person-centered approach to identify the psychological profiles of employees which fit WFH. Since we could not specifically predict the type of teleworker profiles during this period of lockdown, we chose to use exploratory data clustering techniques with a combination of hierarchical (Ward’s method) and non-hierarchical (K-means analysis) methods on a standardized dataset without outliers (Hair et al., 2010). These common methods have been used in several studies with a large number of data sets. It is also important to assess interpretability and theoretical meaningfulness of the latent profiles when determining the optimal solution. For this, we used common graphical representations (i.e., the dendrogram of cluster analysis, elbow method, and t-SNE cluster plots) when determining the optimal number of clusters.

To determine whether identified clusters significantly differed on the preference for solitude and personality variables (extraversion, openness, conscientiousness, agreeableness and neuroticism), multivariate analyses of variance (MANOVA) were conducted. Finally, we examined the effects of the profiles on well-being and creative performance measures with analysis of variance (ANOVAs). Post-hoc tests were corrected for multiple comparisons with the Bonferroni procedure when appropriated. The final sample for analyses consisted of 946 participants. A sensitivity power analysis was conducted using G*Power (Faul et al., 2007). The sample size gave us a power of 95% to detect effects of at least Cohen’s f²= 0.02 for the linear multiple regression, 0.07 for the MANOVA, and 0.12 for the ANOVA.

To check the potential problem of common method bias when self-reported data is used, we performed some of the statistical remedies suggested by Podsakoff et al. (2003). First, we used Harman’s one-factor test (Fuller et al., 2016). We loaded all items of all measures into an exploratory factor analysis (EFA), using maximum likelihood extraction, and examined the solution. No single factor accounting for the majority of the covariance among measures emerged. We obtained a factor solution with the first factor explaining 20.4% of variance. As the first factor accounted for less than 50–60% of the variance among variables (Saris and Gallhofer, 2014), we can assume that there was no common method variance problem. Then, we conducted a CFA on the same dataset, in which all items loaded on a single factor. The solution found had mediocre fit indices: χ²(1080) = 10770, p < 0.001, CFI = 0.493, TLI = 0.470, RMSEA = 0.097, SRMR = 0.103. These preliminary statistical analyses suggest that common method bias was not a serious problem in this study. We also followed the methodological recommendations (Podsakoff et al., 2003) in building the questionnaire by assuring participants that their responses were anonymous and treated confidentially, by separating the blocks of measures from each other and by varying the response scales.

Table 1 reports the values for means and standard deviations, skewness and kurtosis, and intercorrelations between study variables. All significant correlations are small (0.20), or moderate (0.50) in size according to Cohen’s criteria (Cohen and Cohen, 2003). The strongest correlations (>0.65) concern the relationships between the sub-dimensions of some scales (i.e., emotional deprivation and companionship, vigor, absorption and dedication). Moreover, some correlations indicate that the control variables (i.e., gender, age, number of children, WFH experience and WFH intensity) were related to the measured variables, and thus needed to be controlled for in the analyses. On the other hand, as expected, significant relationships were found between the variables related to preference for solitude and associated personality variables and measures of well-being and creative performance.

Multiple linear regressions, controlling gender, age, number of children, WFH experience and WFH intensity, were calculated in order to assess to what extent preference for solitude and related personality variables could predict negative indicators of well-being (i.e., loneliness at work and perceived stress), positive indicators of well-being (i.e., job satisfaction and dimensions of work engagement) and creative performance (i.e., ‘subjective’ and ‘objective’ creativity). The results of these analyses are summarized in Tables 2–4.

For the sub-dimension of loneliness at work related to emotional deprivation (Table 2), the results yielded an overall significant effect, F(11,910) = 19.81, p < 0.001, R²= 0.19, with a significant contribution of gender, preference for solitude and personality variables. Indeed, male teleworkers tended to report more loneliness at work than female teleworkers (β = −0.18). Moreover, emotional deprivation was higher among individuals with a higher preference for solitude (β = 0.06) and with neuroticism traits (β = 0.24). In contrast, emotional deprivation was negatively related to individuals with higher levels of extraversion (β = −0.12), agreeableness (β = −0.18) and conscientiousness (β = −0.25). For the lack of social companionship, the results yielded an overall significant effect, F(11,910) = 12.55, p < 0.001, R²= 0.13, with a significant contribution of gender (β = −0.29). Moreover, lack of social companionship was higher among individuals with a higher preference for solitude (β = 0.07) and with neuroticism traits (β = 0.17). In contrast, lack of social companionship was negatively related to individuals with higher levels of extraversion (β = −0.11) and conscientiousness (β = −0.17). For the symptoms of stress, the results yielded an overall significant effect, F(11,910) = 25.09, p < 0.001, R²= 0.22, with a major contribution of neuroticism traits (β = 0.48), indicating that individuals with high levels of neuroticism expressed more symptoms of stress. Moreover, stress was lower among individuals with a higher preference for solitude (β = −0.03) and among older employees (β = 0.009).

For job satisfaction (Table 3), the results yielded an overall significant effect, F(11,910) = 19.45, p < 0.001, R²= 0.19. Among control variables, only WFH experience was positively associated with job satisfaction (β = 0.17). As expected, preference for solitude was positively related to job satisfaction (β = 0.17) and also extraversion (β = 0.25), agreeableness (β = 0.16), conscientiousness (β = 0.72) and openness (β = 0.19), whereas neuroticism was negatively associated with job satisfaction (β = −0.26). For the components of work engagement, the results showed an overall significant effect, F(11,910) = 26.20, p < 0.001, R²= 0.24 for vigor, F(11,910) = 16.05, p < 0.001, R²= 0.16 for absorption, and F(11,910) = 11.16, p < 0.001, R²= 0.12 for dedication. Among control variables, only WFH intensity was negatively associated with the sub-dimension of vigor at work (β = −0.06). As expected, personality variables related to the Big Five traits were associated with sub-dimensions of work engagement, but not the preference for solitude.

For the “subjective” measure of creativity (Table 4), the results yielded an overall significant effect, F(11,910) = 39.20, p < 0.001, R²= 0.32. The gender variable explained perceived creativity when entering the preference for solitude and personality variables in the regression. Males perceived themselves more creative than females (β = −0.44). WFH experience was also positively associated with perceived creativity (β = 0.10). Preference for solitude and personality variables were associated with perceived creativity.

Finally, concerning the “objective” measures of creativity (Table 4), as expected openness was positively associated with the number of ideas produced (β = 0.28), and extraversion was positively associated with the originality of ideas produced (β = 0.08). The other personality characteristics were not related to fluency or originality scores. Moreover, age was positively associated with the number of ideas produced (β = 0.04), and weakly associated with their originality (β = 0.007).

As expected in our first hypothesis, the experience or intensity of telework alone did not explain the effects of WFH, variables referring to the preference for solitude or Big Five traits often appeared to be related to loneliness at work, stress, job satisfaction, work engagement and creative performance. To explore the data thoroughly, we used a person-centered approach to identify the psychological profiles of employees which fit WFH.

Hierarchical clustering analysis (Ward’s method) and k-means non-hierarchical clustering analysis were performed on the five standardized scores for the preference for solitude variable and personality scores (i.e., extraversion, agreeableness, conscientiousness, neuroticism and openness). First, the results showed that the best number of clusters was two. We then compared different solutions (k = 2 and k = 3) to form clusters with k-means analysis. The three-profile subgroup model fitted the data slightly better (AIC = 4140.27, BIC = 4227.61) than the two-profile subgroup model (AIC = 4635.32, BIC = 4693.54) considering the smaller fit indices. However, the increment in model fit of the three-group model from the two-group model was not significant. Complementary visual techniques based on the inspection of the dendrogram (Figure 1), elbow method plot (Figure 2), and t-SNE cluster plots (Figure 3) confirmed that the solution with two clusters was better than the solution with three less-distinct clusters. This analysis is not consistent with a prior study in which three profiles were identified among a wider population than employees working from home (Michinov and Michinov, 2021).

Figure 4 describes the two profiles retained in the present study. Profile 1, named “Affiliative,” was represented by low levels of preference for solitude and neuroticism, high levels of extraversion and agreeableness, and moderate levels of conscientiousness and openness. This profile concerned 53.91% of the sample (N = 510). Profile 2 was labeled “Solitary” and was characterized by high levels of preference for solitude and neuroticism, low levels of extraversion and agreeableness, and moderate levels of conscientiousness and openness. It concerned 46.09% of the sample (N = 436). A chi-square analysis showed that gender was equally distributed across the two profiles, χ²(1, N = 946) = 1.18, p = 0.276.

To see whether the two clusters identified were distinct and whether each variable contributed significantly to cluster formation, a MANOVA was conducted with cluster membership as the independent variable and the five personality variables as dependent variables. The overall model was significant, revealing that each personality variable appeared to vary significantly between the two profiles [Wilks’ Λ = 0.32, F(6,939) = 328, p < 0.001]. The two profiles differ significantly on the preference for solitude and the Big Five personality variables, except on openness with a marginal difference between the two profiles. The details of means in each cluster are shown in Table 5.

Then, in order to identify the profile of teleworkers at risk, ANOVAs were conducted with cluster membership as the independent variable and well-being and creativity measures as dependent variables. With regard to well-being measures, results showed significant differences between the two clusters (see Table 5). Post-hoc tests showed that the individuals with a “Solitary” profile reported more loneliness at work (i.e., emotional deprivation and social companionship), higher levels of stress, and lower levels of job satisfaction and work engagement (i.e., vigor, absorption and dedication) than the individuals with an “Affiliative” profile. As regards the effects of profiles on creativity measures, results showed that individuals with a “Solitary” profile perceived themselves as less creative and actually produced fewer ideas (fluency score) than those with an “Affiliative” profile. For the originality of ideas, the difference was also significant between the two profiles indicating that according to the assessment of external coders the ideas produced were more original among individuals with an “Affiliative” than a “Solitary” profile.

Overall, these results confirm that the psychological profiles related to a preference for solitude and personality variables associated with the Big Five traits influence employees’ well-being and creativity performance during WFH.

The large sample of teleworkers from various organizations, the use of validated measures of ill- and well-being, and the inclusion of ‘objective’ indicators of creativity represent strengths of the present study. Although our research has revealed insights into the effects of distinct psychological profiles of teleworkers on well-being and ‘objective’ performance measures of creativity during WFH, this study has some limitations. First, concerning the indicators of well-being (loneliness at work, stress, job satisfaction and work engagement) the study relied only on self-report measures. Additionally, it is not possible to assess whether well-being and performance actually changed for these people during the lockdowns. Future studies could test whether similar profiles can be replicated at another time and monitor how they evolve over time. Second, the study has limitations that relate to the individual differences and personality variables measured. In the present research, we examined the effects of individual differences in tolerance to social isolation and preference for solitude. Other individual differences may influence telework attitudes such as procrastination, self-discipline or segmentation preferences (Kreiner, 2006; Allen et al., 2021; Kerman et al., 2021). For example, segmentation preferences refer to the degree to which an employee prefers to keep work- and private-life separate (Allen et al., 2021). A relationship was found between the employees’ segmentation preferences and work-related conflict, indicating that ‘segmentors’ could be able to manage their work-life balance more effectively during mandatory WFH in the context of the COVID-19 pandemic (Allen et al., 2021). Finally, the study is based on a large sample of French teleworkers, and future studies could investigate a more diversified population to generalize the findings in different organizational contexts and in other countries. Unfortunately, this study does not include measures of social contact besides colleagues, and it would be relevant to include measures of social activities and the use of videoconferencing systems. Despite these limitations, this study has several implications for organizations and managers who develop telework practices and virtual teams.

The COVID-19 pandemic has dramatically changed social interactions and work settings. Virtual teams and WFH have rapidly increased and forced workers to adapt to these new forms of work. The pandemic has revealed some groups of workers who are more vulnerable than others, for example, based on their age, ethnicity, gender or personality. The present study provides evidence of employees’ profiles related to preference for solitude and the Big Five personality traits (“affiliative” and “solitary” profiles), and encourages reflection on the needs of teleworkers. As suggested by Kramer and Kramer (2020, p. 2), WFH seem to “require selection of workers who are better suited to work from home, training of such workers on more efficient methods of remote work, and greater monitoring of the quality and productivity of those assigned to work from home.” First, based on the detected profiles, organizations and managers could try to detect the employees’ profiles and organize telework practices in order to guarantee employees’ performance, creativity and well-being. Knowing the different levels of tolerance and sensitivity to professional isolation could help managers to propose support practices and adapted communication tools (e.g., formal and informal networks). The support tools during remote working would also be more useful for inexperienced teleworkers or employees who are uncomfortable with digital technology. Employees should be informed of the effects of remote work, and some guidelines adapted to their personality profile would be provided to arrange their telecommuting time. Experienced teleworkers could also provide informational and social support via communication tools or private rooms on telecommuting platforms. Managers must continue to allow employees flexibility in managing telework according to their personality profiles, their working conditions at home (presence of the spouse, children, etc.) and feedback on their remote work experiences. Managers’ consideration of employee profiles should help to avoid deepening the inequalities at work caused by digitalization and new forms of work organization. Of course, although the remote working experience may be influenced by psychological factors, companies have to adopt organizational strategies reflecting their own situation. In the future, practitioners need to reflect on working outside the office in different places, and this organization of where people work should take into account the characteristics of the work and employee preferences. Human resources managers should not forget that the effects of WFH could differ according to the profile of employees. Future research is also needed to improve understanding of what determines employees’ preferences about workplaces.