Determinants of Perceived Performance during Telework: Evidence from Romania

Abstract

This article confronts telework issues by analyzing how certain factors (motivation, dependence on coordination, self-organizing abilities and stress) impact job performance, as well as some of their interrelations. The research has been carried out with 219 Romanian employees. With the help of Confirmatory Factor Analysis and Structural Equations Modeling, the model led to the following conclusions: employee motivation has a significant positive impact on performance; employees’ dependence on coordination has a significant negative impact on performance; employees’ dependence on coordination has a moderately intense but significant positive correlation with the level of stress perceived during teleworking; and employees’ ability to self-organize their activity is strongly and significantly linked to the level of motivation perceived. The hypothesis that perceived stress has a significant negative impact on performance has not been confirmed. These results add to the specialized literature on telework and can be the basis for future developments of managerial teleworking strategies. The implications are particularly valuable in the context of Corporate Social Responsibility considering the impact of telework on employees, organizations, and society in general.

1. Introduction

In the past decades the unprecedented changes in IT&C have created the premises to stimulate and extend telework even to new professions and categories of staff [1,2]. Relatively recent estimates indicate that 37% of jobs in the USA are fit for telework [3] and in the majority of EU member states it varies between 33–44%, except for five outlier member states [4]. Also, it can be said that telework has benefitted from the impact of the COVID-19 pandemic [2]. Basically, the investments and adjustments made remain useful and have become premises for future developments [5,6]. The future of work is closely tied not only to the technological progress, but also to the new vision that explores the transformative possibilities towards sustainable development, as seen in results from “Transforming our world: the 2030 Agenda for Sustainable Development” [7].

However, just like traditional forms, telework can generate positive effects (e.g., increased working time, flexibility in organizing working time, the possibility for better work-life balance, reduced absenteeism [8,9,10]) or negative effects (e.g., social and professional isolation, enhanced gender inequalities, perceived stress, techno stress, distraction from professional obligations, work overload, self-management issues [8,11,12,13,14,15,16]), which influence worker performance, and consequently financial and nonfinancial organizational performance. Therefore, our aim is to analyze teleworkers’ perceptions concerning a series of factors (motivation, dependence on coordination, self-management skills, and stress) that affected their job performance, but also some of their interrelations [2,17]. The authors of this paper begin with a section presenting details from specialized literature to describe the context of the research and the conceptual framework, then they lay the ground for the research hypotheses, present the research methodology and the results, and conclude with discussions. The model generated with the help of confirmatory analysis and structural equations indicates, among other findings, that motivation has the greatest influence on performance and that there is a very strong, mutually positive relationship between perceived motivation and self-organizing abilities.

The authors believe that this approach can contribute to much more adequate managerial decisions regarding telework, which would lead to more positive outcomes.

Future studies could include other variables to extend the complexity of the model but also to test their validity with other sample groups and in different cultural contexts.

2. Literature Review

2.1. Contextualization

The COVID-19 pandemic was a major challenge with tremendous importance to people’s lives. Great changes were made that impacted people’s professional and personal life. In organizational terms, one widespread solution was the transition to telework [18,19,20]. Obviously, the work itself and the consequences of transitioning to telework vary depending on several factors, including: the type of economic activities and the sector in which they are carried out, the positions and scale at which work is managed, employees’ previous telework experience [21], decisional factors such as work management, performance management, workplace security and health, data security management, digitalization [22] (p. 525), and support from supervisors/managers, colleagues, and friends [23,24,25,26]. Advantages as well as disadvantages, as perceived by teleworkers, were highlighted, some of which are: increased workload correlated with decreased work satisfaction [27], issues concerning the line between professional and extra-professional duties [28,29], increased efficiency, smaller exhaustion risk, decreased promotion opportunities, weakened connection with colleagues and the employer [30], work-life balance or the lack of it [31], productivity and satisfaction growth, flex time, and reduced workplace safety [32].

At the time, telework was not a newly introduced work method, as it had been used before. However, during the pandemic the desired improvements expected from technological progress were not made [18] although the pandemic accelerated the transition towards telework. It is estimated that, before the COVID-19 pandemic, 19% of North American employees worked from home 5 days a week or more, and this number spiked to 44% during the pandemic [33].

In Europe, a JCR study [34] highlights that, at the beginning of 2019, only 5.4% of EU-27 employees were teleworking, most of them in fields such as IT and other communication services, knowledge-intensive business services, education, publishing activities, telecommunications, finance and insurance. The number grew to 13.5% in 2021 [35].

In Romania, much lower figures were recorded compared to the European average (0.4% in 2018 and 0.8% in 2019), including during the COVID-19 pandemic. Some of the causes were: a job market dominated by underqualified and manual workforce, reduced numbers of employees in fields that require extensive use of ITC, limited digital infrastructure, traditional management styles, overcrowded homes, and lack of confidence in digital procedures to solve problems related to public institutions [36]. If we take into account the extremely low interest in IT&C training from companies throughout the year 2019—a JCR study [34] (p. 6) indicates only 6% companies offered training—we can shape a general idea of the conditions and the impact of transitioning to telework in the pandemic. Another study [4] (p. 48) whose purpose was to estimate the range of employment suited to telework reports only 27% of employers, a number which places our country, among others in Eastern Europe, in the lowest ranks. Other data collected by Eurofund in February/March 2021 highlights the preference of only 17.9% of employees to working from home daily, and direct refusal to telework of 33.3% [37].

While experimenting with telework during the pandemic, employees’ opinions concerning this form of work varied. Two years after the pandemic, results from a Pew research center study showed many North American employees were teleworking more from choice than necessity, and that their professional life had changed significantly [38]. The study also shows the benefits (i.e. improved work-life balance, easier work, greater ability to meet deadlines), but also the drawbacks (i.e. lack of real connection with colleagues). The same was the case in European Union countries including Romania (see

Table 1. Differences in teleworking preferences.

Work from Home Preference	EU 27	Romania
Jun/Jul 2020	13.3%	14.4%
Feb/Mar 2021	15.7%	17.9% 1

¹ Sources: [37,39].

), as shown by Eurofund data [37,39]. The study revealed that 59.9% of Romanian teleworkers were ‘satisfied’ or ‘very satisfied’ with the experience [39].

The changes in perception are not arbitrary. Critical times are characterized by fluidity and disillusionment [40] (p. 94), and the data collected from respondents under strong emotional impact generated by perceived risks, fear, stress and other negative emotions cannot accurately reflect future preferences and behaviors [41], which is why it is recommended that data be collected when respondents can offer more precise information. Gathering the data later, after stabilization of the situation, offers respondents the time needed to access more information and to process data logically, which can offer a more precise picture of the way in which certain changes affected them. It is to be noted that not only can the data collected be distorted, but also the cognitive processes of analysts and decision makers [42]. On the other hand, if too much time has passed since the events unfolded, it could have negative effects on the accuracy of the data collected. In the case of the lifting of COVID restrictions, the authors consider that data collection after 2–4 months is reasonable.

Shortly after the COVID-19 pandemic, numerous changes in people’s emotional condition were reported. While some research papers showed a relatively good level of adaptation to the situation [43], other ones described negative feelings generated by a perceived insecurity [44,45,46]. Extreme negative events, perceived as uncertainty or threat, can trigger negative emotions. In this context it is expected that new decisions will be met with resistance. Uncertainty and fear can also amplify current problems based on prejudice or create new ones. It is to be noted how new ideas were created through repeated messages [47]. However, as the situation stabilized, companies and employees were able to better understand the benefits and challenges brought by telework, and studies moved toward analyzing results concerning sustainability, stressing the necessity of issuing policies to decrease the main risks, and offering practical perspectives for teleworkers and managers [48,49,50].

Therefore, working from home has become a viable and increasingly accepted solution with a tendency to grow and opportunities for sustainable development. Although telework is currently not covering the desired proportion of employees due to the gap between companies’ expectations and real possibilities of digitalization [51], efforts are being made and the international strategies are being designed (i.e., the European model towards a digitalized economy and society [52]). Under these circumstances, it is very important to prepare organizations for successful growth in teleworking, through accepting aims related to sustainability and awareness of the way in which certain factors related to work performance can help the process. For this reason, the purpose of this research to analyze the perception of employees on the factors (i.e., motivation, stress, coordination, self-organization) that affected their performance when teleworking. The impact of these factors, identified from the specialized literature, is specified in the following section and in Section 3. This research brings together a set of variables that have been mentioned fragmentarily in other studies, which also allows the analysis of the relationships between them. In addition, the interest of this research is to analyze how these factors reflect the perception of Romanian employees regarding telework, because there are specific aspects that describe the situation in Romania. Thus, this study covers some gaps in the literature regarding a consistent set of variables that influence performance within telework, to specify the relationships between them, and to generate some recommendations for management to capitalize on these results.

2.2. Telework and Work Performance

With regards to the concept of telework we need to point out that: ‘telework’ and a host of other terms, such as ‘homeworking’, ‘telehomeworking’, ‘telecommuting’, ‘remote working’, ‘virtual work’, ‘electronic homeworking’, ‘distributed work’, ‘e-Work’, ‘home-anchored work’, ‘flexible work’, ‘flexplace’, and ‘distance work’ have been used interchangeably [18,53]. Consequently, there is some variation in how the concept is defined (see examples in

Table 2. Definitions of telework and similar concepts.

Definitions of Telework	Authors
telework as an activity which “includes all work-related substitutions of telecommunications and related information technologies for travel”	Jack Nilles, 1973, Apud Collins, 2005 [54]
“we define telework as paid work from home, a satellite office, a telework center or any other work station outside of the main office for at least one day per workweek”	Illegems, Verbeke, 2004, pp. 319–320 [55]
“telework or telecommuting describes work undertaken away from traditional offices by means of technology”	Böll et al., 2014, p. 1 [56]
“telecommuting is a work practice that involves members of an organization substituting a portion of their typical work hours (ranging from a few hours per week to nearly full-time) to work away from a central workplace—typically principally from home—using technology to interact with others as needed to conduct work tasks”	Allen et al., 2015, p. 44 [53]
“telework (remote work, telecommuting, flexible work, virtual work) is a work practice that involves working away from the corporate office for a portion of the work week, typically from home, and using technology as needed to conduct work”	Golden, 2021, p. 1 [28]
“the form of work organization through which the employee, regularly and voluntarily, fulfills his attributions specific to the position, occupation, or profession he holds, in another place than the work organized by the employer, at least one day a month, using information and communication technology”	Law No. 81/2018 of 30 March 2018, Romania Article no. 2/a [57]

), but essentially the meaning is the same as it refers to working outside the organization using facilities offered by new information and communication technologies.

Many studies have shown increased teleworker performance following increased working time away from the office, lack of interruption from traditional work environments, flexibility in organizing working time, the possibility for better work-life balance, and reduced absenteeism [8,9,10].

However, there are also studies stressing that certain factors can negatively impact individual performance, some of which include social and professional isolation, monitoring/supervising issues [8,11], enhanced gender inequalities [12] (p. 802), perceived stress [13], techno stress due to the home-work conflict and professional overload [14] (p. 441), distraction from professional obligations [15], feedback issues, work overload, and self-management issues [16] (p. 126).

The transition to telework brought about transformations in the contents and manner of work, and we can speak about adjustments and job reconfigurations, but also adjustments to the jobs’ motivational features. It is known that a job configuration can have a motivational effect [58] and studies have shown that such reconfigurations can influence employees’ internal motivation [59], and consequently their work performance [60].

Telework is extending to new professions and categories of staff [1], but results on the impact of telework on employees’ perceived performance vary. There are positive and negative features, sometimes without a clear-cut line between them, and the results are contradictory. Of course, more knowledge is required for organizational management and for managerial decisions that would bring about desired behaviors and correct flaws. This validates the importance of this research, the hypotheses of which are taken from the specialized literature and presented in the following section.

3. Conceptual Model and Hypotheses

The review of specialized literature allowed the authors to highlight some relations of interest regarding the factors determining employee performance in teleworking. These relations are at the core of the conceptual model developed in the present research. We highlighted four factors that impact employees’ job performance in teleworking conditions (dependent variable), both directly as well as following their interrelation:

• Employees’ motivation;
• Employees’ dependence on coordination by a superior;
• Employees’ self-organizing ability;
• Stress perceived by an employee.

Analysis of the specialized literature revealed the following relationships comprising these factors: a positive impact of motivation on performance [8,9,10,31], a negative influence of employees’ dependence on manager coordination on performance [11,16], a negative impact of stress on performance [13,14], a positive relationship between employees’ dependence on manager support and the level of stress [13,61,62], and a positive relationship between self-organization and motivation [63,64,65].

In addition to these, age [66] and gender [67] are factors that affect existing relationships. Although their influence is intuitive, it is also based on literature and needs to be reconfirmed.

Starting from the factors and relationships identified in specialized literature, the following research hypotheses were formulated:

Hypothesis 1 (H1).

Motivation has a positive impact on employees’ job performance in teleworking conditions.

Hypothesis 2 (H2).

Employees’ dependence on coordination by the manager negatively affects their performance during telework.

Hypothesis 3 (H3).

Perceived stress has a negative impact on employees’ job performance in teleworking conditions.

Hypothesis 4 (H4).

Employees’ dependence on coordination by the manager positively correlates with the level of stress perceived during telework.

Hypothesis 5 (H5).

The ability of the employees to self-organize their activity is positively correlated with the level of perceived motivation.

In addition, the model will check for age and gender influences by taking into account their interrelation with motivation (for age and gender), dependence on coordination (gender), self-organizing ability (age) and stress (gender).

4. Summary of Research Design

Research Instruments, Data Selection and Analysis

In order to collect the data, the authors developed an online questionnaire and targeted employees who teleworked during the pandemic in Romania, specifically in Iasi and its surroundings. The data was collected in June 2022, about three months after COVID-19 restrictions were lifted (March 2022). Thus, the possibility that stress and negative emotions affect the accuracy of answers was reduced [41], an aspect that was detailed in Section 2. A total of 228 questionnaires were completed. After eliminating outliers identified by applying Mahalanobis Distance, 219 answers remained, which were subjected to subsequent analysis. The authors consider the sample size to be satisfactory, given that a typical value for SEM is around 200 cases; the researchers recommend a minimum of 10 cases per parameter for the results to be considered reliable [68] (p. 12). Some sources, such as Wolf et al. [69] and Sideridis et al. [70], identify even smaller samples as sufficient.

Identification questions included information about age, gender, education, field of work, level of seniority on the job, etc. They were followed by questions regarding the elements of the econometric model. First, there were question that referred to the evolution of job performance in teleworking conditions: “How was your professional performance during telework?” with response options presented in a 5-point Likert scale, from (1) Much weaker to (5) Much better. Next were questions that targeted four constructs (latent variables): motivation, coordination dependence, self-organization, and stress. The response options were also organized in a 5-point Likert scale, from (1) Strongly disagree to (5) Strongly agree.

The analysis began with descriptive statistics and normality test, especially the multivariate normality which represents the vulnerability of the Maximum Likelihood procedure, commonly applied for Structural Equation Modeling (SEM) [71]. The next stage was the data reliability test utilizing Chronbach’s Alpha and Convergent Reliability (CR) and data validity test using Variance Extract (Ave). Then, Confirmatory Factor Analysis was used to test the model fitness, and Structural Equation Modeling (SEM) was used to examine the relationships between variables and to test the hypotheses. Structural Equation Modeling was chosen due to the fact that it allows building a model on theoretical constructs (factors) that cannot be observed directly (latent variables), but can be measured indirectly through observable variables [68,71]. Other important advantages are represented by the fact that SEM allows the evaluation of the entire model, which brings a higher-level perspective compared to other techniques [68], but also the possibility to represent the structural relationships pictorially, which allows a clearer conceptualization of the theory [71]. A possible disadvantage of SEM is the fact that it requires a larger sample compared to other methods [68], but we are confident that we managed to satisfy the requirements in this respect.

For SEM, the authors first used the Maximum Likelihood procedure, applied in most cases [71]. Given the premise of this procedure, related to the normality of data and the fact that the Likert scale used for data collection involves collecting ordinal categorical data (which, by definition, cannot represent a normal distribution), the authors subsequently applied the process of Bayesian estimation using the Markov Chain Monte Carlo algorithm and compared the results with those obtained by Maximum Likelihood. The analyses were performed in SPSS and Amos v.20.

5. Results

5.1. Descriptive Statistics

As stated in the Section 4, 219 questionnaires were analyzed, and the biodata of the sample group is presented in

Table 3. Identification data for the respondents.

Gender	Frequency	Percent	Education	Frequency	Percent
Masculine	58	26.5	High school	19	8.7
Feminine	161	73.5	Other	5	2.3
Age	Frequency	Percent	Bachelor	84	38.4
Under 25 years	111	50.7	Master	101	46.1
26–35 years	50	22.8	PhD	10	4.6
36–45 years	33	15.1	Company size	Frequency	Percent
46–55 years	24	11	Between 1–9 employees	34	15.5
Over 56 years	1	0.5	Between 10–49 employees	60	27.4
Field of work	Frequency	Percent	Between 50–99 employees	29	13.2
Agriculture	2	0.9	Between 100–249 employees	19	8.7
Industry	3	1.4	Over 250 employees	77	35.2
Services	59	26.9	Seniority on position	Frequency	Percent
Retail sales	19	8.7	Under 1 year	68	31.1
Hospitality industry	7	3.2	1–3 years	78	35.6
IT	41	18.7	3–5 years	27	12.3
Education	31	14.6	5–10 years	12	5.5
Other	56	25.6	Over 10 years	34	15.5

. The questionnaire was completed mostly by women (73.5%), with men representing only about a quarter of respondents (26.5%). The majority of the study participants were relatively young, with 111 respondents under 25 years old (50.6%), 50 respondents aged between 26 and 35 (22.8%), and 33 respondents aged between 36 and 45 (15.1%). Most respondents had completed some higher education: 101 participants with a Master’s degree (46.1%), 84 with a Bachelor’s degree (38.4%), and the remaining 19 were high school graduates (8.7%). Respondents worked in large and small organizations: 77 people worked for companies with over 250 employees (35.2%), 60 people worked in small companies (10 to 49 employees) (27.4%), and 34 people came from micro-enterprises with fewer than 10 employees (15.5%). The organizations mainly represented the service sector (59 respondents, 26.9%), 41 people worked in the IT sector (18.7%), 31 employees came from the education sector (14.6%), while 19 worked in retail sales (8.7%). There were respondents from the hospitality, industry, and agriculture sectors, and also 56 people from other fields (25.6%). Most of the participants had a low seniority in their current positions; 78 of them had been in the current position between 1 and 3 years (35.6%) and 68 people (31.1%) for less than 1 year, followed by respondents with over 10 years (15.5%), between 3 and 5 years (12.3%), and between 5 and 10 years in the current position (5.5%).

Identification data reveal some limitations of the study, represented by the disproportionate structure of the sample: the majority of respondents are young people (under 25) (50.6%), identify as female (73.5%), have completed some level of higher education (84.5%), and had low seniority in their current position (66.7% with three years or less).

5.2. Evaluation of the Model Fitness

Testing the model fitness started with a reliability check of the data with the use of Chronbach’s Alpha and Convergent Reliability (CR), and a validity check using Average Variance Extract (Ave). The results of the estimates confirmed the reliability of the data (

Table 4. Results of the model’s reliability and validity analysis.

Construct	Item	Loading	Cronbach’s Alpha	CR	AVE
Motivation	Motiv1—The novelty of telework has pushed me to be more productive	0.715	0.820	0.824	0.612
	Motiv2—I think that the prospects of my future career are better when I telework	0.729
	Motiv3—I think that telework ensures a better work-life balance, which positively influences the quality and quantity of the work performed	0.890
Coordination dependence (CoordDepend)	CoorD1—The chief’s support is important to perform the tasks well	0.488	0.631	0.712	0.577
	CoorD2—It is important that the boss give me clear indications about what I have to do to have good performance during telework	0.957
Self-organization (SelfOrganize)	SOrg1—The time I wasted in traffic was better used during telework	0.774	0.902	0.907	0.765
	SOrg2—The comfort of home helped me/helps me work better and more efficiently	0.948
	SOrg3—I organize my work time better when I’m home	0.893
Stress	Stre1—Fear that I will lose my job made me work more during telework	0.875	0.688	0.713	0.563
	Stre2—Fear of the virus affected/affects my work	0.600

). Even though Cronbach’s Alpha for Coordination dependence (CoordDepend) and Stress (Stress) is a little under 0.7 (a unanimously accepted limit) which reflects moderate reliability [72], the CR value is over 0.7 for both constructs, demonstrating an acceptable reliability [73]. Overall, the CR values for the analyzed constructs are between 0.712 and 0.907. The discriminatory validation of the model constructs was made by using Average Variance Extracted (Ave and demonstrated results between 0.563 and 0.765, with values over 0.5 [74] considered acceptable. The loadings of the factors, representing the variant of the respective factor explained by the latent variable, are presented in the same table. All values are over 0.4, which is the minimum limit accepted for this sample group size [75].

Regarding the fitness of the model, the results of the Confirmatory Factor Analysis demonstrate that the model is solid. For example, the ratio between the minimum discrepancy function and the degrees of freedom (Cmin/DF) is 2.026, a favorable situation, because the values of Cmin/DF under 3 show a good fit of the model and a material evidence of a survey [76]. In addition, NFI (normed fit index) is 0.916, a very good level of model fitness, considering that the values of this indicator are between 0 and 1, 1 being the perfect fit. Another example is the result of RMSEA calculation (Root Mean Squared Error Approximation), the value of 0.069 demonstrating an acceptable level of model fitness [77]. In conclusion, the authors can proceed to Structural Equations Modeling (SEM) to test the research hypotheses.

5.3. Structural Equations Modelling

To begin, SEM was performed using Maximum Likelihood estimation, as this procedure is used in most situations when SEM is done for data collected using Likert scales [71], even if some of its premises are specific to continuous data. The results of the modeling are presented in Figure 1.

All the relationships presented, except the one related to how Stress influences Performance, were confirmed to be significant. The results of hypotheses tested are presented in

Table 5. Results of the evaluation of hypotheses.

Hypothesis	Relationship	Estimate	Acceptance or Rejection of the Hypothesis
1	Motivation → Performance	0.612 *** 1	Accept (Major impact)
2	CoordDepend → Performance	−0.205 **	Accept (Minor-Medium impact)
3	Stress → Performance	−0.021	Reject (Non-significant)
4	CoordDepend ↔ Stress	0.490 ***	Accept (Moderate relationship)
5	Motivation ↔ SelfOrganize	0.960 ***	Accept (Very strong relationship)

¹ Notes. *** p < 0.001; ** p < 0.02.

. The range of the impact is established according to [78], as follows: the value of β below 0.10 signifies a “minor” impact, the values around 0.30 mean a “medium” influence, and the value over 0.50 shows a “major” impact. The regression coefficient for the relationship between Motivation and Performance (β = 0.612, p < 0.001) demonstrates that the motivation of the employees had a major and significant positive impact on their performance in the context of telework, which allows us to accept H1. The H2 hypothesis is also accepted due to the significant negative impact of employees’ dependence on coordination from the manager on their performance during telework, even if this impact is minor to medium (β = −0.205, p < 0.02). H3 was rejected because even if the (negative) nature of the impact of Stress on Performance meets the expectations, the result is very low, so statistical significance has not been demonstrated.

Hypotheses H4 and H5 were also validated, as the relationships match the expectations and their significance is demonstrated. The employees’ dependence on coordination by the manager has a positive, moderately intense, but significant correlation with the level of Stress perceived during telework (r = 0.490, p < 0.001). At the same time, the ability of the employees to Self-organize their activity has a very strong and significant positive connection with the perceived level of Motivation (r = 0.960, p < 0.001): the employees with a better ability to organize their own activity will have a stronger motivation for that activity, and vice versa.

Also, all relationships between the control variables and the latent variables have been confirmed as statistically significant, even if the intensity of the correlation was not strong (Figure 1). Particularly, the age of the employees is negatively correlated with both their motivation (r= −0.283) and their ability to self-organize (r = −0.274), the relationship being a weak one. As for Gender, there are significant interconnections, but very weak ones, between Gender and Motivation (r = −0.074), on the one hand, Coordination dependence (r = 0.192) and Stress (r = 0.181), on the other.

Therefore, four out of five hypotheses were accepted, one being rejected due to the lack of statistical significance of the relationship. Even though the normality criteria presented in [71] were satisfied and there were no impediments in applying the Maximum Likelihood procedure, due to the fact that a Likert scale was used for collecting the data and in order to bring additional arguments to the model, the authors also performed SEM through the Bayesian estimation process. It was realized using the Markov Chain Monte Carlo algorithm and the results were compared with those obtained by Maximum Likelihood. The results, generated as Polygon and Trace diagnostic charts, demonstrate that Amos successfully identified the important characteristics of posterior distribution for all the elements and that in all distributions the convergence was achieved quickly, which is a clear indicator that the model was correctly featured [71].

In order to confirm the conclusions made based on the analysis of the diagnostic charts,

Table 6. Comparison of Unstandardized Parameter Estimates: Maximum Likelihood versus Bayesian estimation.

Parameter	Estimation Approach
	ML	Bayesian
Motiv1 ← Motivation	0.853	0.860
Motiv2 ← Motivation	1.000	1.000
Motiv3 ← Motivation	1.296	1.300
CoorD1 ← CoordDepend	0.430	0.428
CoorD2 ← CoordDepend	1.000	1.000
SOrg1 ← SelfOrganizing	0.717	0.716
SOrg2 ← SelfOrganizing	1.000	1.000
SOrg3 ← SelfOrganizing	1.016	1.017
Stre1 ← Stress	1.510	1.532
Stre2 ← Stress	1.000	1.000
Performance ← Motivation	0.723	0.726
Performance ← CoordDepend	−0.170	−0.170
Performance ← Stress	−0.021	−0.026

presents the comparison of unstandardized regression coefficient values calculated with the two procedures: Maximum Likelihood and the Bayesian estimation.

The estimates are very close for all the parameters and in agreement with the expectations set by the analysis of the charts. This is an additional argument for the validity of the proposed model [71].

6. Discussion

The objective of this study was to identify and investigate the influencing factors of self-perceived telework performance under telework conditions. The results of the research carried out by developing a model through SEM demonstrated the existence of significant relationships between variables, the results being consistent with those of the studies within the literature [8,9,10,11,13,16,31,43,61,62,63,64,65,79], even if one of the hypotheses was rejected.

With regards to the first hypothesis, the major positive impact that employees’ motivation has on their perceived performance in the context of telework has been demonstrated. Hence, the way employees perceived the advantages of telework motivated them and led to increased performance. Similar results were reported in other studies [8,9,10,31].

Another important result is shown by the negative influence that employees’ dependence on manager’s support has on their performance during telework, even if the intensity of the relation is minor to medium. Several studies highlighted the fact that the performance of teleworking employees was influenced by support and supervision from superiors [11,16]. The authors believe that, in the context of the pandemic, the rapid shift to telework justifies this perception, but for those familiar with or even accustomed to this form of work and for those with a high level of autonomy, the problems related to supervision did not manifest themselves so strongly, hence the minor to medium impact resulting from the current study.

Regarding the relationship between stress and perceived performance, no significant connection was identified, H3 being the only hypothesis that was rejected. Consequently, this contradicts the literature on which the hypothesis was based [13,14]. However, these results are compatible with those of other researchers who analyzed this topic in the context of the medical crisis [79] or who reported good emotional adaptation [43]. The authors believe that teleworkers felt protected in relation to possible negative consequences because this form of work was actually to their advantage.

The perceived level of stress, however, is linked to another significant relation: a correlation of moderate intensity with the employees’ dependence on coordination. In the context of telework, it was expected that such a connection would be identified. First of all, through various managerial practices, superiors have the ability to regulate the level of stress perceived by employees [13]. Arnold presents a series of studies demonstrating that, through leadership, managers succeed in reducing the level of stress at work [61]. In work-from-home circumstances, managers had limited opportunities to exert their influence on employees, including in the context of stress management, and employees that were more dependent on coordination from the supervisor began to perceive a higher level of stress. The same conclusion would be reached if one started from the understanding of the strong dependence on coordination as an expression of a reduced level of autonomy at work, the latter being associated with the reduction of stress at work [62].

Finally, it was demonstrated that employees who have the ability to organize their own activity also have a higher perceived level of motivation, and vice versa, the connection between the variables being very strong. In this case, the results of the analysis are also consistent with those of other research in the field. The positive connection between employees’ ability to work autonomously (including working from home) and a high level of motivation, particularly an intrinsic one, has been repeatedly demonstrated [63,64]. The same is true for the opposite relationship: more motivated employees tend to have a greater degree of autonomy at work [65].

7. Conclusions

7.1. Recommendations

Following the analysis of specialized literature and the results of this study, the authors suggest a series of recommendations for managerial practice:

• Capitalize on teleworking situations by better elaborating on everything related to the job description in order to increase motivation; employees are motivated by novelty, variety of tasks, autonomy, feedback, etc. [59].
• Know the aspirations of teleworkers, develop coherent strategies and transparent policies for the promotion and development of one’s career within telework.
• Develop policies and strategies in order to ensure a good work-life balance, starting from the investigation of factors that can have a negative influence in this respect.
• Develop practices and to offer tools that contribute to better communication between employees and superiors, between colleagues and collaborators.
• Train managers and employees for a better adaptation to telework conditions.
• Develop policies and strategies aimed at encouraging autonomous work and improving self-organization abilities.

7.2. Implications

This study brings contributions at a theoretical level by systematizing some important information from the specialized literature regarding the relationship between telework and performance on the one hand, and by conceptualizing the model and highlighting the conclusions on the other hand. Another theoretical contribution this study makes is advanced understanding of the domain results from the conceptual clarification of the notion of telework and the systematization of the factors determining performance in the context of telework. The practical value lies in the fact that the results can be capitalized on in the management area, as the variables included in the model are important for organizations that have to increasingly adapt to environments characterized by uncertainty. In addition, the identification of some relationships between variables brings a better understanding of how coherent strategies can be developed to support each other, leading to increasing their positive effects within the context of telework.

The results of this study can be capitalized upon especially in the development of strategies regarding Corporate Social Responsibility, considering that the implications of telework are manifested in several directions [80], such as the following: impact on employees (i.e., work-life balance, employee satisfaction, reduction in travel times, implications for career, etc.), impact on organizations (i.e., flexibility, productivity changes, cost changes, the response of customers, etc.), and impact on societal level (i.e., space savings, environmental outcomes, emission reductions, etc.).

7.3. Limitations and Directions for Future Research

The main limitation is geographical; the research was carried out in the northeastern region of Romania, so the results are affected by aspects related to cultural specificity, local managerial practices and other characteristics that were not specifically controlled for while building the model. For this reason, the extrapolation of the results of this research in other geographic areas should be done with caution. Identification data also reveal some of the limitations of the study, given the disproportionate structure of the sample: the majority of respondents are women, with higher education, young people under 25 years old, with low seniority in their current position. Another aspect is related to the fact that not all possible variables and relationships were taken into account, which represents both a limitation and an improvement idea for future research.