Working from home: When is it too much of a good thing?

Abstract

Based on an integration of meta‐theoretical perspectives on the “too much of a good thing” effect with psychological demands and resources theories, we propose and test nonlinear relations between the percentage of time people work from home and a variety of important work‐related outcomes (i.e., professional isolation, work from home satisfaction, work from home self‐efficacy, work performance, job satisfaction). Then, also based on resource theories, we explore whether and how certain work from home resources (i.e., previous experience working from home, appropriate technologies to facilitate working from home, dedicated workspaces) buffer these nonlinear relations. Data on working from home were provided by n = 994 employees in Germany across 32 monthly measurement waves between April 2020 and December 2022. Our results support the general idea that the percentage of time people work from home has nonlinear associations with a variety of important work‐related outcomes. However, only in a few cases (i.e., work performance, job satisfaction) do these relations take the form of inverse U‐shapes that would be indicative of “too much of a good thing.” Our exploratory analysis suggests that, in several cases, work from home resources can buffer these nonlinear associations. These findings have implications for the continued development of meta‐theoretical perspectives on “too much of a good thing” and for employees' and organizations' attempts to make working from home a positive and productive experience.

Full text

WORKING FROM HOME
1
Working from Home: When is it Too Much of a Good Thing?
Cort W. Rudolph1 and Hannes Zacher2
1Department of Psychology, Wayne State University
2Wilhelm Wundt Institute of Psychology, Leipzig University
This is a pre-print version of an in-press accepted manuscript. Please cite as:
Rudolph, C.W. & Zacher, H. (2024, In Press). Working from home: When is it too much of a
good thing? Human Resources Development Quarterly.
Author Note
Cort W. Rudolph https://orcid.org/0000-0002-0536-9638
Hannes Zacher https://orcid.org/0000-0001-6336-2947
Ethics approval for this study was granted by the ethics advisory board of Leipzig
University (Protocol ID# 2019.06.27_eb_17, Study title: “A longitudinal study on experience
and behavior at work”).
The study reported in this article was funded by Volkswagen Foundation (Az. 96 849-1,
“Work and Health in the Time of COVID-19: A Longitudinal Study”).
Correspondence concerning this article should be addressed to Cort W. Rudolph, Wayne
State University, Department of Psychology, 5057 Woodward Ave, 7th Floor, Detroit MI 48202
(USA). Email: cort.rudolph@wayne.edu

WORKING FROM HOME
2
Abstract
Based on an integration of meta-theoretical perspectives on the “too much of a good thing” effect
with psychological demands and resources theories, we propose and test nonlinear relations
between the percentage of time people work from home and a variety of important work-related
outcomes (i.e., professional isolation, work from home satisfaction, work from home self-
efficacy, work performance, job satisfaction). Then, also based on resource theories, we explore
whether and how certain work from home resources (i.e., previous experience working from
home, appropriate technologies to facilitate working from home, dedicated workspaces) buffer
these nonlinear relations. Data on working from home were provided by n = 994 employees in
Germany across thirty-two monthly measurement waves between April 2020 and December
2022. Our results support the general idea that the percentage of time people work from home
has nonlinear associations with a variety of important work-related outcomes. However, only in
a few cases (i.e., work performance, job satisfaction) do these relations take the form of inverse
U-shapes that would be indicative of “too much of a good thing.” Our exploratory analysis
suggests that, in several cases, work from home resources can buffer these nonlinear
associations. These findings have implications for the continued development of meta-theoretical
perspectives on “too much of a good thing” and for employees’ and organizations’ attempts to
make working from home a positive and productive experience.
Keywords: work from home; demands; resources; nonlinear; too much of a good thing

WORKING FROM HOME
3
Working from Home: When is it Too Much of a Good Thing?
Working from home (i.e., a form of “telecommuting” sometimes also referred to as
“remote work”) is an important type of flexible work arrangement (Lewis, 2003; Mattis, 1990),
with distinct implications for human resources development (HRD; e.g., Carnevale & Hatak,
2020). For example, in many modern organizations, employees’ ability and motivation to work
from home is a relevant consideration in recruitment, selection, and training and development.
Regarding the latter, employees need to develop the knowledge, skills, and attitudes that enable
them to work effectively from home, while maintaining their wellbeing. Although the COVID-
19 pandemic (March 2020 to early 2023) catalyzed a focus on working from home in both
research and practice (e.g., Kniffin et al., 2021; Rudolph et al., 2021), it is by no means of new
phenomenon. Indeed, research has focused on related practices (e.g., telecommuting) and its
implications for (virtual) HRD for some time (Bennett, 2010). As noted by Bennett and
McWhorter (2021), “…many of the changes to the workplace were already underway due to
digital transformation, and the pandemic simply accelerated this transition, as well as brought a
few unique challenges that may eventually become extinct or remain embedded in civic
culture…” (pp. 6-7).
For the field of HRD, the increased prevalence of working from home over the last
several years has been noted as a double-edged sword (Kaiser et al., 2022; Park et al., 2023).
That is, different outcomes associated with working from home have not been universally
positive. On the one hand, working from home presents challenges to traditional “in person”
means of managing and developing workers. On the other hand, there are many potentially
unrealized opportunities associated with working from home (e.g., allowing employees
flexibility and choice in their work locations). Indeed, related topics of work flexibility and

WORKING FROM HOME
4
remote work have been noted as one of Forbes “Top 10 HR Trends” for 2023, making this topic
of primary concern to contemporary HRD practitioners (Meister, 2023). Accordingly, there is a
pressing need for research to consider how working from home affects a variety of work
outcomes, whether these relations are universally linear and positive, and whether there are
certain boundary conditions that affect the strength of these relations. Indeed, a closer
understanding of such associations is important for understanding potential targets for HRD
interventions.
This study is motivated by a longstanding question in the HRD literature regarding the
relative benefits and potential costs of working from home for employees. Indeed, citing earlier
writing by Stephens and Szajna (1998), Bennett (2010) points out that, when it comes to working
from home, “In some cases, such as with autonomy, a factor can represent both an advantage and
a disadvantage. For example, some people may thrive when they have greater autonomy,
whereas others may become aimless or feel abandoned” (p. 731). Given these general
observations, the core research questions that drive our study are, “when, and under which
conditions, is working from home ‘too much of a good thing’?” More specifically, borrowing
support from the theories reviewed below, we investigate how the extent of working from home
is related to important work-related outcomes, and which boundary conditions may shape these
associations.
The act of working from home (vs. working from more traditional workplaces) generally
has small, albeit consistent, linear relations with several important work-related outcomes (see
Shockley, 2014). However, evidence for nonlinear (e.g., quadratic or “U-shaped”) relations
between time spent working from home and work outcomes suggests that the amount of time
people work from home may be more important than whether they work from home (Allen et al.,

WORKING FROM HOME
5
2015; Shockley, 2014). For example, multiple studies have shown that the highest levels of job
and life satisfaction are experienced at moderate levels of working from home, amounting to
about two workdays per week (see Fonner & Roloff, 2010; Golden, 2006; Golden & Veiga,
2005; Virick et al., 2010). This finding is consistent with meta-analytic evidence regarding the
nonlinear influence of telecommuting on work-related outcomes (Gajendran & Harrison, 2007).
Following from such research, best practice recommendations developed approximately
ten years ago suggest that, to maximize its effectiveness, working from home should be limited
to approximately two workdays per week (i.e., about 2/5ths or 40% of a typical five-day
workweek; Shockley, 2014). However, this suggestion has not been thoroughly tested
empirically, at both the between- and within-person levels of analysis (i.e., in terms of
interindividual differences vs. intraindividual variability across time), and for multiple work-
related outcomes. Indeed, there is mixed support for the assumed positive relation between
working from home and work-related outcomes, and there is a lack of research on how to
harness the assumed benefits by optimizing the balance of working from home versus working
from more traditional workplaces (e.g., offices). Other notable limitations of research on working
from home and employee outcomes are that the evidence so far is either (a) meta-analytic
(Gajendran & Harrison, 2007), which is at a different conceptual level of analysis (i.e., study-
level vs. individual-level effects of the extent of working from home), (b) cross-sectional (e.g.,
Anakpo et al., 2023; Van der Lippe & Lippeńyi, 2020), which does not allow conclusions about
dynamic within-person relations, or (c) if studies are longitudinal, they are typically based off of
very few time points (e.g., Graham et al., 2023, consider three timepoints between October 2020
and November 2021; Park et al., 2023, consider two timepoints between April 2021 and
May/June 2021).

WORKING FROM HOME
6
Beyond these empirical rationales, there are strong theoretical reasons to consider
nonlinear relations between working from home and work-related outcomes. Indeed, the meta-
theoretical framework of “too much of a good thing” suggests that there exists a “tipping point”
(i.e., an inflection point) at which engaging in otherwise “good” behaviors can become “bad”
(Grant & Schwartz, 2011; Pierce & Aguinis, 2013; Warr, 1994). Such “too much of a good
thing” effects of working from home may be explained for each specific outcome investigated
here by drawing from psychological demands and resources theories, such as the job demands-
resources model (Bakker & Demerouti, 2014; Demerouti et al., 2001). These theories suggest
that work design has positive effects on employee motivation and wellbeing when work-related
resources outweigh or compensate for work demands (e.g., high autonomy and high workload,
respectively), and negative effects when work demands outweigh resources (e.g., high workload
and low autonomy, respectively).
Based on an integration of perspectives on the “too much of a good thing” effect and
psychological demands and resources theories, we investigate nonlinear relations between the
percentage of time people work from home and a variety of important work outcomes (i.e.,
feelings of professional isolation, work from home self-efficacy, work from home satisfaction,
work performance, job satisfaction; see appendix for definitions of these constructs). We
consider these outcomes because they represent commonly studied indicators of employee
motivation and wellbeing related to working from home (e.g., Gajendran & Harrison, 2007;
Golden & Veiga, 2005; Golden et al., 2008; Shockley, 2014). We assume that the percentage of
time people work from home will manifest nonlinear relations with such outcomes (see Figure 1
for a conceptual representation of the “too much of a good thing” effect).
We also consider, based on psychological resource theories and through exploratory

WORKING FROM HOME
7
moderator analyses, how working from home resources (i.e., previous experience working from
home, appropriate technologies to facilitate working from home, dedicated workspaces at home)
may act to buffer these nonlinear relations. Again, we chose these moderators based on
theorizing and research on working from home (Gajendran & Harrison, 2007; Shockley, 2014).
To accomplish the goals of this study, we consider data from a panel of n = 994 employees who
participated in a large-scale longitudinal study with thirty-two monthly measurement waves
across nearly three years (i.e., considering data on working from home between April 2020 and
December 2022) to disentangle both average, between-person relations between the amount of
working from home and work outcomes, and how within-person variability in the amount of
working from home over time relates to such outcomes.
This study contributes to our theoretical understanding of working from home in several
ways. First, we contribute to the meta-theoretical perspective on “too much of a good thing” by
integrating this theoretical lens with psychological demands and resources theories and
extending it to the study of working from home. So far, most research on this perspective has
focused on other domains of organizational behavior and beyond (e.g., leadership, personality;
Pierce & Aguinis, 2013), with fewer studies focusing on job design and related topics (e.g.,
Stiglbauer & Kovacs, 2018; Warr, 1994), such as flexible work arrangements, including working
from home (e.g., Golden, 2006; Golden & Veiga, 2005; Virick et al., 2010). Still, the duality of
both the benefits and challenges of working from home have been recognized as a broader,
macroeconomic phenomenon (Behrens et al., 2021).
Second, most research on working from home focuses on the between-person level of
analysis with cross-sectional (i.e., single time point) research designs (e.g., Oakman et al., 2022;
Weitzer et al., 2021). Thus, we contribute to the literature on working from home by

WORKING FROM HOME
8
simultaneously examining how differences between employees and variability within employees
regarding their percentage of time working from home relate to important work outcomes.
Finally, based on psychological resource theories, we consider actionable resources (i.e.,
experience, technologies, workspaces) that could be the focus for HRD interventions aimed at
improving the experience and effectiveness of working from home (e.g., Demerouti, 2023). Our
exploratory moderator analyses therefore focus on these resources as potential malleable factors
that can inform future HRD intervention work.
Overall, by integrating psychological demands and resource theories, this work represents
a novel extension of the meta-theoretical perspective on “too much of a good thing” beyond
typical domains of study and adds to the myriad ways that “too much of a good thing” can
manifest in work-relevant outcomes. Moreover, by adopting a longitudinal research design, we
extend theorizing on “too much of a good thing” to both the between- and within-person levels
of analysis. This is an important contribution to this meta-theoretical perspective, as we can
investigate and explore the importance of differentiating “too much” for people in general (i.e.,
between-person nonlinear effects of working from home) from “too much” for individual
employees (i.e., within-person nonlinear effects of working from home).
Recent Research on Working from Home
Given the near ubiquity of working from home during the COVID-19 pandemic, there is
no shortage of empirical research related to the topic in recent years. Indeed, a handful of
systematic reviews have summarized the influence of working from home during the pandemic,
suggesting that, in general, working from home may be a “double edged sword.” That is, there
are both positive and negative aspects of working from home with implications for worker
attitudes, motivation, productivity, and related outcomes. With respect to the present study,

WORKING FROM HOME
9
working from home is likely to be positively associated with both positively valanced (e.g., work
from home self-efficacy, work from home satisfaction, work performance, job satisfaction) and
certain negatively valanced (e.g., professional isolation) work outcomes. As we argue next, these
patterns are likely to be affected by the balance between specific work-related demands and
resources relevant to each outcome that are encountered when working from home.
A systematic review by Hackney et al. (2022) reports mixed support for working from
home on employee productivity across k = 37 reviewed studies, with observed linear relations
being either negative, null, or positive across these studies. Similarly, a systematic review by
Vleeshouwers et al. (2022) concluded that there was mixed evidence for relations between
working from home and aspects of the psychosocial work environment (e.g., professional
isolation; job satisfaction; k = 43). For example, research has found inconsistent linear relations
between working from home and job satisfaction, with some studies suggesting positive and
some null relations. Vleeshouwers et al. (2022) concluded that there was a pressing need for
higher-quality research on working from home to address these questions, including the need for
research to consider multi-wave longitudinal designs, include precise measures of time use and
location of work, and to use validated measures – each of which we address in the present study.
Taken together, this mixed support and evidence for negative, null, or positive relations across
studies may suggest that working from home does not have a simple linear and monotonic
relation with all outcomes, and that it may rather manifest in terms of nonlinear (i.e., inverse U-
shaped) relations. Such relations are consistent with the meta-theoretical principle of “too much
of a good thing,” discussed next.
Working from Home can be “Too Much of a Good Thing”
The idea of “too much of a good thing” reflects a meta-theoretical principle that certain

WORKING FROM HOME
10
seemingly positive phenomena may exhibit nonlinear relations with outcomes, such that there
are linear and positive relations to a point, which reach an inflection point, and then become
negative (i.e., they take the form of an “inverse U-shaped” relation; see Figure 1 for an example).
Research on “too much of a good thing” phenomena at work has focused on various topics,
including personality and job performance (e.g., Le et al., 2011), job characteristics (e.g.,
Langfred, 2004; Warr, 1994; Zhou, 2020), leadership (e.g., Antonakis et al., 2017; Harris &
Kacmar, 2006; Peterson, 1999), self-presentation (e.g., Baron, 1986), and social identification
(Avanzi et al., 2020).
Linear Effects of Working from Home. We assume that working from home is a
generally positive phenomenon (i.e., spending more time working from home benefits a variety
of work-related outcomes). Based on the job demands-resources model (Demerouti et al., 2001),
we argue that the act of working from home provides employees with several specific work-
related resources that, generally, outweigh work-related demands (e.g., workload). Indeed,
working from home to a greater extent should be generally positively related to both job
satisfaction and working from home satisfaction, because this mode of working satisfies
employees’ basic need for autonomy (Deci & Ryan, 2013) by enabling them to make more
decisions about where, when, and how to carry out their work tasks. Fulfilling one’s need for
autonomy and perceiving control over one’s work are key determinants of employee wellbeing,
including various forms of satisfaction (Terry & Jimmieson, 1999).
Furthermore, working from home to a greater extent should be generally positively
related to working from home self-efficacy, because it provides more opportunities for mastery
experiences regarding this flexible working arrangement, which satisfy employees’ basic need
for competence (Deci & Ryan, 2013). Working from home to a greater extent should further be

WORKING FROM HOME
11
generally positively related to different forms of work performance, which are directly
determined by the multiplicative function of employees’ knowledge, skills, and motivation
(Campbell & Wiernik, 2015). Although employees’ knowledge and skills relevant for
performance are the same when working from home or at traditional workplaces, greater need
fulfillment (i.e., needs for autonomy and competence) should lead to higher levels of work
performance.
We also recognize that working from home can be a “double-edged sword,” having
positive associations with both positively valanced and negatively valanced outcomes. We
assume that working from home to a greater extent will be generally positively associated with
professional isolation, because it provides fewer opportunities for employees to fulfill their basic
need for relatedness (Deci & Ryan, 2013) through personal interactions with their supervisor and
coworkers. Accordingly, we hypothesize the following:
Hypothesis 1: The percentage of time people work from home is positively and linearly
related to (a) professional isolation, (b) work from home satisfaction, (c) work from home self-
efficacy, (d) work performance, and (e) job satisfaction.
Nonlinear Effects of Working from Home. Consistent with the notion of “too much of
a good thing,” we also assume that the otherwise positive effect of working from home on work-
related outcomes has an inflection point, resulting in an overall (i.e., average) positive effect that
is qualified by in inverse U-shaped (i.e., a negative quadratic or second-order polynomial)
relation. Such a pattern is supported by multiple convening lines of research and theorizing on
the non-monotonic effects of working from home (e.g., Gajendran & Harrison, 2007; Golden,
2006; Vander Elst et al., 2017; Virick et al., 2010). Support for these arguments is especially
drawn from theory and research on the demands imposed by high levels of working from home

WORKING FROM HOME
12
which, consistent with the job demands-resources model (Bakker & Demerouti, 2014), are likely
to outweigh the work-related resources that go along with working from home. Complimentary
support for this assertion can be found in needs-based theoretical perspectives, such as self-
determination theory (e.g., Deci & Ryan, 2013) and related job design perspectives that outline
which factors help satisfy employees basic needs (e.g., job enrichment). These various lines of
support are next reviewed and, in turn, integrated into our predictions regarding the nonlinear
effects of working from home.
First, the act of working from home represents an exercise in autonomy which, to some
extent, is a reflection on one’s choice to enact flexibility and choice in one’s work location
(Sewell & Taskin, 2015). Both theory and research on “the paradox of choice” (e.g., Iyengar &
Lepper, 2000) suggests choice exhibits non-monotonic effects, in that having more choice is
“good” to a point (e.g., satisfaction increases) and then “bad” after (e.g., satisfaction decreases).
Similarly, both theoretical (Hobfoll, 1989, 2002; Warr, 1994) and empirical (e.g., Kubicek et al.,
2014; Stiglbauer & Kovacs, 2018) arguments support the idea that too much autonomy may,
paradoxically, be a bad thing.
Regarding research on the non-monotonic effects of job autonomy, a meta-analysis found
evidence for inverse U-shaped relations between autonomy and worker wellbeing (including job
satisfaction) and behavioral (including job performance) outcomes, suggesting that, beyond a
certain point, additional autonomy may be detrimental (Yam Mei Har, 2021). Similarly,
regarding research on the non-monotonic effects of working from home, meta-analytic evidence
suggests that telecommuting intensity moderates the otherwise positive relationship between
telecommuting and certain work outcomes (Gajendran & Harrison, 2007). Specifically,
Gajendran and Harrison (2007) find that telecommuting is negatively associated with certain job-

WORKING FROM HOME
13
related outcomes (e.g., work-family balance, coworker relationship quality) when engaged in for
more than two days per week. Similar findings are reported in primary empirical studies. For
example, Virick et al. (2010) report an inverse U-shaped relation between the extent of
telecommuting and job satisfaction. These findings are likewise consistent with research that has
applied job demands-resources theory (Bakker et al., 2023) to suggest that low levels of working
from home can be resourceful, whereas excessive working from home can be demanding (e.g.,
Galanti et al., 2021; Golden & Veiga, 2005; Song & Gao, 2020). Both lines of empirical
evidence suggest the potential for inverse U-shaped relations between working from home and
various work outcomes.
Second, based on the job demands-resources model, the ability to work from home, and
the associated job autonomy that this affords employees, can be viewed as a work-related
resource, as it can help individuals offset the detrimental effects of demands (Demerouti et al.,
2001). Moreover, working from home is an exercise in autonomy, as it reflects both choice and
agency. However, like other manifestations of afforded job autonomy, working from home may
present a paradox in that “too much” can be detrimental. For example, the vitamin model (Warr,
1987) proposes that different motivational job characteristics, including job autonomy, operate
like vitamins, in that their effectiveness for positively or negatively influencing work outcomes
depends on their overall levels. At low levels, any small incremental change in favorable job
characteristics (e.g., flexibility to work from home) are likely to have positive results. However,
beyond a certain level, no additional benefit is derived from adding “more” of such
characteristics to one’s job. This can result in either a tapering-off of the effectiveness (i.e., a
“constant effect”) or a detrimental negative pattern (i.e., “additional decrement”). Considering
the vitamin model in the present study, a shift from not working from home at all toward

WORKING FROM HOME
14
relatively lower levels of working from home (e.g., one or two days a week) likely benefits
certain work outcomes (e.g., job attitudes, work performance). However, at higher levels (e.g.,
three to five days a week), working from home may not show any additional benefit or may
become a detriment to one’s ability to perform certain aspects of their job.
Additionally, conservation of resources theory (Hobfoll, 1989, 2002) offers that people
act in various ways that protect their limited pool of available resources. The experience of
depleted resources is stressful and can further deplete people’s pool of available resources (i.e., a
“loss spiral,” Hobfoll et al., 2018). Having the flexibility and autonomy to work from home is
otherwise considered a resource. However, when experienced at particularly high levels, people
must dedicate other personal resources (e.g., personal time, energy) toward managing high levels
of autonomy at the expense of other job requirements. The re-allocation of personal resources,
otherwise reserved for non-work-related endeavors, may lead to higher levels of strain, which
subsequently further reduces one’s capacity to manage their autonomy at the expense of other
work outcomes.
Consistent with the argument that working from home reflects, in part, employees’
choices to enact flexibility, we expect that working from home operates in a similar paradoxical
manner. That is, working from home for some percentage of one’s time is likely to be a positive
experience up to a specific point, after which point it becomes a negative experience, because it
adds to one’s overall workload (Schwartz, 2005; Warr, 2002, 2016). Working from home
presents people with a great deal of flexibility and autonomy in the way that they approach
various work tasks. This may result in people making sub-optimal choices about how they spend
their time, leading them to work off-hours or to put off otherwise important tasks, with
consequences for work-related outcomes (Schwartz, 2005). Moreover, even if employees’ work

WORKING FROM HOME
15
demands are met by the flexibility afforded when working from home, the high-level of
autonomy that comes with this option may result in sub-optimally high levels of accountability
for the outcomes of one’s work, which, according to activation theory, can be a stressful
experience (e.g., Gardner, 1986; Gardner & Cummings, 1988). Such experiences are likely
manifested in common work outcomes, including job satisfaction and work performance, but
also in motivational and wellbeing outcomes that reflect one’s experience working from home,
including work from home satisfaction and self-efficacy and experiences of professional
isolation.
Spending some percentage of time working from home is likely to satisfy individual’s
basic needs for autonomy, competence, and relatedness up to a certain point, but can also deprive
these needs thereafter (i.e., at higher levels of working from home). From a self-determination
theory perspective, working from home can both help and harm the satisfaction of basic needs
for autonomy, competence, and relatedness (Gagné et al., 2022). For example, working from
home to a moderate amount affords workers with more flexible work schedules but, when
enacted to a greater extent, also may increase conflicts between work and non-work roles (i.e.,
stifling one’s need for autonomy). Working from home to a small to moderate amount likewise
affords workers with opportunities to master new challenges and may present unique learning
opportunities, but hassles and constraints associated with communications technology and
information overload are also possible when working from home to a greater extent (i.e., stifling
one’s need for competence).
Finally, working from home to a small or moderate extent likely increases employee’s
experiences of professional isolation, because they recognize that they are missing out on typical
day-to-day social interactions (e.g., ad hoc conversations; lunch with colleagues) on days when

WORKING FROM HOME
16
they are not working from the office (i.e., stifling one’s need for relatedness). However, at higher
levels of working from home, people are more likely to have developed strategies and routines to
ensure such interactions are still possible, for example, by using technologies that allow for
people to both connect virtually across different locations and to communicate face-to-face (i.e.,
benefiting one’s need for relatedness). Indeed, research suggests that professional isolation
experienced when working from home may be mitigated to the extent that employees proactively
initiate interdependence between themselves and their coworkers (Bell, 2020).
Considering opportunities and challenges associated with the enactment of basic needs,
we argue that small to moderate levels of working from home are especially characterized by
opportunities to satisfy employees’ needs for autonomy and competence. However, the
satisfaction of these basic needs is likely thwarted by higher levels of working from home. The
negative effects of very high levels of autonomy are well documented. For example, Baltes et al.
(2002) found that (low) high levels of job autonomy are (positively) negatively associated with
job satisfaction, suggesting an inverse U-shaped association. Similarly, research on job
enrichment finds nonlinear relations between certain elements of enriched work environments
(which help satisfy needs for autonomy and competence; Fernet et al., 2013) and work outcomes.
For example, Champoux (1992) found an inverse U-shaped relation between job scope and
various dimensions of work satisfaction. Given its affordance of autonomy, working from home
can be viewed through a job enrichment lens, in that it offers employees a higher degree of
responsibility and variety in the way that they approach their work.
At relatively lower levels, an increase in the extent of working from home should fulfill
employees’ needs for autonomy and competence (Deci & Ryan, 2013) and in turn, lead to higher
working from home satisfaction and job satisfaction, as well as working from home self-efficacy,

WORKING FROM HOME
17
respectively. Moreover, work performance should be increased due to the motivational effects of
the satisfaction of multiple basic needs (Gagné & Deci, 2005). Additionally, at relatively lower
levels, a greater extent of working from home should initially lead to an increase in professional
isolation because employees’ may struggle to meet their basic need for relatedness when
working remotely (Deci & Ryan, 2013). Indeed, research posits otherwise beneficial relations
between flexibility and autonomy and these outcomes, in that higher levels of work flexibility
and autonomy are associated with higher levels of job and work-related satisfaction (e.g., Baltes
et al., 2002; Dierdorff & Jensen, 2018) and performance (e.g., Dierdorff & Jensen, 2018), as well
as higher levels of self-efficacy (e.g., Kubicek et al., 2014). Likewise, research points to positive
relations between working from home and the experience of professional isolation (e.g., Spilker
& Breaugh, 2021).
At relatively higher levels (i.e., after an inflection point), an increase in the extent of
working from home may result in additional work and personal demands that thwart the positive
effects of employees’ fulfillment of autonomy and competence needs. Specifically, employees
may be more likely to feel overwhelmed by high levels of autonomy and flexibility and in turn,
experience decreases in working from home satisfaction and job satisfaction. These additional
demands at relatively high levels of working from home likely also reduce their opportunities to
experience mastery and a sense of competence, which should reduce their working from home
self-efficacy. Moreover, at relatively high levels of working from home, work performance
should also be reduced due to the additional work demands outweighing the positive
motivational effects. Additionally, at relatively higher levels, a greater extent of working from
home should result in a decrease in professional isolation. As suggested, this may be due to the
development of means and routines that allow such employees to maintain social interactions,

WORKING FROM HOME
18
bolstering their need for relatedness. It could also be that such employees realign the way they
approach their work (i.e., they “craft” different aspects of their jobs) to allow their needs to
match their work arrangements. In doing so, they may prioritize managing their core work tasks,
especially given additional work and personal demands, while de-prioritizing other aspects of
their work that are more closely tied to need for relatedness (see Biron et al., 2023 for a dynamic
model linking need satisfaction to work outcomes for teleworkers).
In summary, there is both a strong theoretical rationale and ample empirical evidence to
suggest that working from home will have a generally positive influence on a variety of work
outcomes, but that such positive effects can become “too much of a good thing” at a certain
point. Thus, we posit that the positive linear relations described in Hypothesis 1 are qualified by
a nonlinear (i.e. inverse U-shaped) association. Accordingly, we hypothesize:
Hypothesis 2: The positive linear relation between the percentage of time people work
from home and (a) professional isolation, (b) work from home satisfaction, (c) work from home
self-efficacy, (d) work performance, and (e) job satisfaction is qualified by negative, nonlinear
(i.e., quadratic or second order polynomial) relations.
We consider tests of our two hypotheses at both the between- and within-person levels of
analysis, and likewise explore whether these assumed nonlinear patterns are conditional upon
work from home resources, both of which are discussed next.
Between- and Within-Person Effects of Working from Home
One distinct advantage of the longitudinal nature of the data we consider here is the
ability to differentiate between-person from within-person linear and nonlinear relations between
the percentage of time people work from home and the outcomes we consider. Such relations are
largely untested, and there is a lack of clear theoretical guidance to aid in forming differential

WORKING FROM HOME
19
hypotheses about them. For example, it could be that the relations are isomorphic, that is, that the
effects are similar at both levels of analysis (Zacher & Rudolph, 2020). Alternatively, it could be
that different functional forms of the relation between working from home and the outcomes we
consider emerge at different levels of analysis (e.g., linear vs. nonlinear; U-shaped vs. inverse U-
shaped). As suggested, considering both within-person and between-person relations is an
important contribution to the “too much of a good thing” meta-theoretical perspective, because
we can differentiate “too much” for people in general (i.e., between-person nonlinear effects of
working from home) from “too much” for individual employees (i.e., within-person nonlinear
effects of working from home).
Despite the lack of clear theory to guide us here, there is some emerging thought on this
matter from qualitative research on the concept of “autonomy calibration,” which suggests that
people essentially tune their levels of autonomy to an optimal level (Prengler et al., 2021). This
perspective hints at the idea that differential effects of working from home might be expected at
the between- versus the within-person level of analysis. The autonomy calibration phenomenon
is analogous to the existence of an interindividual “set point” for job autonomy, with deviations
from this point being to some extent disruptive (for similar arguments about wellbeing, see
Diener et al., 2009). Given the level of autonomy it affords individuals, working from home
could represent such a dynamic phenomenon, especially at the within-person level of analysis.
Indeed, people likely hold clear expectations regarding their (i.e., average) pattern of working
from home during any given week, and deviations from that calibrated point result in suboptimal
levels of important work outcomes (i.e., much like homeostatic perspectives on dynamic
equilibrium, see Heady & Wearing, 1989). Still, theory in this space is not yet well formalized to
the extent that it is possible to make clear predictions here. Accordingly, we explore the form of

WORKING FROM HOME
20
between-person and within-person linear and nonlinear relations rather than make explicit
hypotheses about them.
Work from Home Resources as Potential Buffers
According to psychological resource theories (e.g., Gorgievski et al., 2011; Hobfoll,
1989, 2002; Hobfoll et al., 2018), both intangible (e.g., experience) and tangible (e.g.,
equipment, physical space) resources can benefit people at work (e.g., offset negative
experiences; benefit positive experiences). Job demands-resources theory (Bakker & Demerouti,
2014) and the related work-home resources model (ten Brummelhuis & Bakker, 2012) both offer
that various job and personal resources work together to offset the detrimental experience of
work demands, and ultimately benefit motivational and wellbeing processes (e.g., engagement)
and work performance. Moreover, extending our line of theorizing presented above, conservation
of resources theory (Hobfoll, 1989, 2002; Hobfoll et al., 2018) suggests that personal resources
can offset the experience of demands and their influence on work-related outcomes.
Presently, we consider whether three specific working from home resources buffer the
assumed nonlinear relations between percentage of time working from home and work-related
outcomes: (a) experience working from home before COVID-19, (b) possessing the necessary
equipment and technologies to work from home, and (c) having a dedicated workspace in one’s
home. According to the work-home resources model (ten Brummelhuis & Bakker, 2012), which
is based on conservation of resources theory, experience is a key “constructive” personal
resource, whereas possessing necessary equipment and technologies and having dedicated
workspaces represent contextual resources categorized by objects or conditions (see also
Hobfoll, 1989).
Research suggests that resources are generally important for managing the potential ill-

WORKING FROM HOME
21
effects of working from home on a variety of work-related outcomes (Frodermann et al., 2021;
Shockley et al., 2020). Such findings are complimented by research conducted both during and
prior to the pandemic, which suggests that specific resources (e.g., prior experience working
from home; Hayes et al., 2021; a suitable workspace, Nakrošienė et al., 2019; proper equipment,
Niebuhr et al., 2022) are important for maintaining productivity and related outcomes (e.g.,
wellbeing) while working from home. Moreover, beyond resource theories, the consideration of
work from home resources can be justified theoretically based on self-determination theory (e.g.,
Ryan & Deci, 2009), as such resources can help meet people’s need for autonomy, competence,
and relatedness when those are not exclusively satisfied by the mere act of working from home
(e.g., Biron et al., 2023; Gagné et al., 2022). Thus, we explore whether working from home
resources buffer (i.e., weaken the strength of) the nonlinear relations of working from home and
the work outcomes investigated here.
Method
Openness and Transparency
Data, R code to replicate the analyses reported below, and complete results are available
in our online supplemental materials: https://osf.io/h6q7j/. The data used in this paper were
collected as part of a larger, ongoing longitudinal study with 35 measurement waves across three
years (December 2019 to December 2022). Twelve articles based on the same dataset have been
published (see Table S1 in the online supplemental materials for a data transparency matrix). Ten
of these articles have completely different research questions and include totally different
substantive variables than the current study. The current study makes use of job performance and
job satisfaction data from time [T] T3 – T4 and T6 – T35 that has also been partially reported in
two other papers (Weiss et al., 2022; Zacher et al., 2021). However, the current study contributes

WORKING FROM HOME
22
above and beyond these studies by considering job performance data at 24 and job satisfaction
data at 25 additional measurement waves. No study based on this dataset has focused on working
from home.
Participants and Procedure
This study was approved by the ethics advisory board of Leipzig University (approval
number: 2019.06.27_eb_17). A sample of n = 994 working adults from Germany were drawn
from a larger 35-wave longitudinal panel study collected between December of 2019 and
December of 2022. Participants who work from home were not explicitly sampled as part of this
study, however importantly, during the COVID-19 pandemic, most “non system relevant”
employees were afforded the option to work from home. Respondents were broadly sampled
from different organizations and represent a wide-variety of industries ranging from the
extraction of raw materials (i.e., primary sector) and manufacturing (i.e., secondary sector), to
service industries which exist to facilitate the transport, distribution and sale of goods produced
in the secondary sector (i.e., tertiary sector). Table 1 provides additional demographic descriptive
statistics for the n = 994 respondents. Importantly, in our sample of German employees,
participants purposefully chose to work from home—even during times of high infection rates
during the COVID-19 pandemic. Indeed, according to German law, all participants could choose
whether to work from home, voluntarily (see Brion & Westhues, 2020). To ensure data quality,
an ISO 26362 certified professional panel company was commissioned to recruit participants
from a nationally representative online panel in Germany. To be included, participants had to be
at least 18 years old, be working full-time, and report that they could work at least some
percentage of their time from home across the timeframe considered here.
Owing to the complex design and sampling strategy of this study, data from thirty-two

WORKING FROM HOME
23
different monthly waves were considered in our analysis, as described next. In some cases (i.e.,
for demographics and working from home resources) available data from multiple waves was
considered to account for (a) the addition of two refresher samples added in July 2020 and
October 2020 and (b) observed missingness on key moderators (e.g., for any given respondent, if
data on working from home resources was missing at Tx, data from Tx+1 was used). Specifically,
the waves of data across time considered here include T1 and T22 (i.e., December 2019 and
November 2021; combined demographics from the primary sample and two refresher samples),
T3, T4, and T6 (i.e., April, May, and July of 2020; combined reports of the availability of
working from home resources), and T3 – T4 and T6 – T35 (April 2020 – May 2020 and July
2020 – December 2022; reports of percentage of time working from home per week and work-
related outcomes). Of note, working from home data were not collected at T2 (March 2020) and
T5 (June 2020), however at T1 (i.e., December 2019), the average percentage of time working
from home pre-pandemic was M = 15.11% (SD = 25.21%).
As is typically the case in longitudinal data collections, some missingness was observed
across the sample of n = 994 respondents considered in our focal analyses, especially on
exogenous covariates and moderators that are considered as part of our supplementary analyses.
To use “all available data” in these analyses (see Newman, 2009), the sample sizes across
various iterations of the models presented here vary to some extent (see Tables 1, 2, 3, and 4)
from n = 554 to n = 994. Table 1 compares demographic and substantive variables for complete
and incomplete responders on relevant baseline demographic and substantive variables.
Descriptive statistics for observed sample sizes for each wave of the study are available in our
online supplemental materials. Of note, the mixed effects modeling strategy we adopt here is
well-geared for handling missing data on endogenous (i.e., outcome) variables considered here,

WORKING FROM HOME
24
under the assumption that data are “missing at random” (MAR; Enders, 2022; Singer & Willett,
2003). Evidence in support of the assumption of MAR can be considered if a given variable’s
missingness is unrelated to the variable itself, however missingness may be associated with other
variables in the dataset (i.e., missingness on one variable is likely associated with missingness on
another; see Beaujean, 2014; Enders, 2022). Regarding the former, we dummy coded
missingness at each timepoint (i.e., 0 = missing; 1 = complete) for each variable and considered
correlations between missingness on each variable and the variable itself. The average
correlations between missingness and associated variables ranged from rxy = -0.06 to 0.01,
suggesting no appreciable pattern of missingness that would deviate from MAR (e.g., missing
not at random; MNAR). Moreover, to address the possibility of differential attrition over time,
we followed the advice of Goodman and Blum (1996) and tested a binary logistic regression
model predicting attrition status (i.e., complete vs. incomplete) from the demographic and
substantive variables considered here. This model predicted less than 8% (R2 = .079) of the
variance in attrition over time (Tjur, 2009), suggesting that the possibility of systematic selection
is unlikely. Complete accountings of these analyses are available in our online supplemental
materials. Additionally, for the sake of space, a complete accounting of descriptive statistics and
intercorrelations among study variables across all thirty-two time points as well as summaries of
within- and between-person correlations can be found in our online supplemental materials.
Measures
Participants were asked to respond to a series of measures, reflecting on their experience
at work across the past four weeks (i.e., corresponding to the one-month time lag between survey
administrations) when considering their responses.
Percentage of Time Working from Home. A single item developed by Golden and

WORKING FROM HOME
25
colleagues (see Golden & Eddleston, 2020; Golden et al., 2008) was used to collect the
percentage of time working from home at T3 – T4 and T6 – T35. Participants were asked to
report the percentage of their time they spent working from home (i.e., 0% to 100%),
representing the proportion of an average workweek spent working at home, away from the
office, across the past four weeks. The average test-retest reliability for this measure was rxy =
.715 (SD = .096).
Professional Isolation. Perceived professional isolation was measured using four items
adapted from the scale by Golden et al. (2008; see also Galanti et al., 2021). The items were (a)
“I felt left out on work-related activities and meetings,” (b) “I missed out on work-related
opportunities,” (c) “I missed face-to-face contact with coworkers,” and (d) “I missed informal
interaction with others at work” (5-point scale ranging from 1 = never to 5 = very often).
Reliabilities for this measure were acceptable at both the within- (⍺ = .793, ⍵ = .798, VARave =
.526) and between-person (⍺ = .898, ⍵ = .862, VARave = .625) levels of analysis.#
Work from Home Self-Efficacy. Consistent with guidance from the literature regarding
the assessment of domain-specific self-efficacy (e.g., Bandura, 2006; Raghuram et al., 2003), we
measured work from home self-efficacy with a single item developed for this study, “I am
convinced that I can successfully complete my work tasks when I am working from home” (5-
point scale ranging from 1 = not at all to 5 = very much). The average test-retest reliability for
this measure was rxy = .660 (SD = .068).
Work from Home Satisfaction. Based on research on single items assessing job
satisfaction (e.g., Fisher et al., 2016; Matthews et al., 2022; Wanous et al., 1997), we collected
work from home satisfaction with a single item, “All in all, how satisfied were you with working
from home?” (5-point scale ranging from 1 = very dissatisfied to 5 = very satisfied). The average

WORKING FROM HOME
26
test-retest reliability for this measure was rxy = .609 (SD = .080).
Work Performance. Task proficiency, adaptivity, and proactivity were assessed by self-
report using three items each from Griffin et al. (2007). Example items are “I carried out the core
parts of my job well” (task proficiency), “I adapted well to changes in my core tasks”
(adaptivity), and “I initiated better ways of doing my core tasks” (proactivity). The 7-point
response scale ranged from 1 = never to 7 = always. Reliabilities for all three dimensions were
acceptable at both the within- (task: ⍺ = .828, ⍵ = . 828, VARave = .616; adaptive: ⍺ = .747, ⍵ =
.757, VARave = .511; proactive: ⍺ = .845, ⍵ = .846, VARave = .647) and between-person (task: ⍺
= .938, ⍵ = . 947, VARave = .857; adaptive: ⍺ = .896, ⍵ = .902, VARave = .755; proactive: ⍺ =
.981, ⍵ = .981, VARave = .944) levels of analysis.
Job Satisfaction. Job satisfaction was assessed with a widely-used single item (“All in
all, how satisfied were you with your work?”; Fisher et al., 2016; Matthews et al., 2022; Wanous
et al., 1997). The scale was anchored from 1 = very dissatisfied to 7 = very satisfied. Research
has demonstrated that single items of job satisfaction represent reliable and valid
operationalizations of this rather homogeneous construct (Dolbier et al., 2005; Wanous et al.,
1997). The average test-retest reliability for this measure was rxy = .600 (SD = .088).
Work from Home Resources. Three different work from home resources were collected
with single items, and were used in our supplemental exploratory analysis of moderators. First,
previous experience working from home was assessed with the question: “Before the COVID-19
pandemic, did you generally have the opportunity to work from home?” (0 = “no,” 1 = “yes”).
Technologies to facilitate working from home were measured with the question “Do you have
the necessary technical equipment at home to work from home (e.g., computer, internet)?” (0 =
“no,” 1 = “yes”). Finally, dedicated workspace was measured with the question, “If you worked

WORKING FROM HOME
27
from home in the past weeks, where did you work?” (0 = “dedicated workspace,” 1 = “other;”
e.g., in a location in a different room or different places at home, depending on demands).
Covariates. Several personal and work-related demographic variables were collected
with single items and were used in a supplemental covariate sensitivity analysis. Specifically, in
such supplemental analyses, we controlled for chronological age (i.e., continuous), sex (0 =
“male,” 1 = “female”), educational attainment (0 = “lower secondary school,” to 3 =
“college/university or technical college”), industry (0 = “primary & secondary sector,” 1 =
“tertiary sector”), contract hours (i.e., continuous), and whether or not respondents had at-home
childcare responsibilities (i.e., 0 = “no,” 1 = “yes”). We also control for whether data were
collected during a period of national lockdown associated with the COVID-19 pandemic (i.e., 0
= “typical period,” 1 = “lockdown period”). These covariates were specifically chosen because
research has variously shown that experiences of working from home can vary based on such
demographic and employment-related features (e.g., Craig & Churchill, 2021).
Analysis
To account for the nesting of data within-person and over time, which leads to non-
independence of observations, we adopted a mixed effects linear approach modeling using the
`lmer` function from the {lme4} package for R. The main predictor in these models, percentage
of time working from home, was person mean centered into orthogonal between- and within-
person components (Bolger & Laurenceau, 2013). Between-person percentage of time working
from home represents each respondent’s average, which was grand mean centered (i.e., for any
given respondent, scores below zero indicate working from home less than the average of the
entire sample, whereas scores above zero indicate working from home more than the average of
the entire sample). In contrast, within-person percentage of time working from home represents

WORKING FROM HOME
28
deviations around each respondent’s average (i.e., for any given respondent, scores below zero
indicate working from home less than their own personal average amount of time working from
home, whereas scores above zero indicate working from home more than their own personal
average amount of time working from home).
To test our central hypotheses, we specified mixed effects models where each outcome
variable was simultaneously regressed onto linear and quadratic forms of between- and within-
person percent of time working from home, which were entered into the model as fixed effects.
Each model also controlled for the timing of the survey to account for exogenous time-varying
sources of variance (e.g., pandemic lockdowns; Rudolph & Zacher, 2021). Linear and quadratic
terms were specified in these models as second-degree orthogonal power polynomials using the
`poly` function in the R {stats} package (Kennedy & Gentle, 1980). Consistent with best
practices for the modeling of statistical control variables (Becker, 2005; Becker et al., 2016;
Breaugh, 2006), these focal models were supplemented with a covariate sensitivity analysis and
an exploratory analysis of working from home resources as moderators.
Results
Descriptive statistics for our sample can be found in Table 1. Intercorrelations among
study variables at the between- and within-person levels of analysis and additional multilevel
descriptive statistics are available in our online supplemental materials. As a first step in our
analysis, we considered a two-level confirmatory factor analysis to differentiate the fit of the four
multi-item scales (i.e., professional isolation; task, proactive, and adaptive work performance) to
the data at both the between- and within-person levels of analysis. This model fit the data
acceptably [c2(118) = 8,823.130, p < 0.001, RMSEA = 0.075, CFI = 0.897, SRMRwithin = 0.058,
SRMRbetween = 0.138] and gave us grounds to differentiate these outcomes in our statistical

WORKING FROM HOME
29
models, reported next.
Primary Hypothesis Tests
Table 2 summarizes the models specified to test our central hypotheses. We initially
noted that there were significant (p < .05), positive effects of survey timing in five of the seven
models, suggesting that there were over time increases in working from home satisfaction and
decreases in professional isolation and work performance across the thirty-two waves we
considered here.
Tests of Hypothesis 1. In partial support of Hypothesis 1, significant, positive, linear
effects of between- and within-person percent of time working from home were observed for all
outcomes, except for the between-person effect on proactive performance and the within-person
effect on job satisfaction (see Table 2; n.b., we do not elaborate on these findings fully here for
the sake of space and because such effects are qualified in the presence of nonlinear terms). This
suggests that, in general and at both levels of analysis, working from home for a greater
percentage of time is associated with higher levels of (a) professional isolation, (b) work from
home satisfaction, (c) work from home self-efficacy, and (d) task and adaptive work
performance, as well as proactive performance at the within-person level of analysis. Likewise,
working from home for a greater percentage of time is associated with higher levels of (e) job
satisfaction at the between-person level of analysis.
Tests of Hypothesis 2. Tests of nonlinear relations qualify these findings regarding linear
effects, but the form of such nonlinear effects cannot be ascertained from the parameter estimates
alone. Specifically, observed significance of the quadratic terms suggests nonlinearity in the
relation between percent of time working from home and the outcomes considered here, but it
does not speak to the form of the nonlinearity. For example, a negative quadratic effect might

WORKING FROM HOME
30
suggest an inverse U-shaped relationship, but it could also suggest a generally positive
relationship that increases at an increasingly-decreasing rate (i.e., a relationship that “tapers off”
at higher levels, but that otherwise suggests that the highest predicted level of an outcome occurs
at the highest levels of the predictor; a “diminishing return,” in other words). Similar, albeit
opposite, interpretations could be made for a positive quadratic effect. Thus, to better understand
these effects and bolster our hypothesis tests, we plotted predicted slopes for each quadratic term
in the models described next (see Figure 2).
Support for Hypothesis 2 was generally mixed. Consistent with our predictions, we
observed a significant (p < .05) nonlinear (i.e., negative quadratic) effect of between-person
percent of time working from home for (a) professional isolation (ɣ = -26.535, SE = 2.069, see
Figure 2, panel “A”). However, in contrast to Hypothesis 2, significant (p < .05) nonlinear (i.e.,
negative quadratic) effects of between-person percent of time working from home were observed
for (b) work from home satisfaction (ɣ = -7.859, SE = 3.392, see Figure 2, panel “C”) and (c)
work from home self-efficacy (ɣ = -10.937, SE = 3.355, see Figure 2, panel “E”), but the form of
these effects is not consistent with “too much of a good thing.” Also contrary to our hypothesis, a
significant nonlinear, but positive, quadratic effect was observed for (d) task performance (ɣ =
16.321, SE = 3.613, see Figure 2, panel “G”).
Consistent with Hypothesis 2, significant nonlinear (i.e., negative quadratic) effects of
within-person percent of time working from home were observed for (d) work performance (i.e.,
task: ɣ = -5.847, SE = .790; adaptive: ɣ = -4.813, SE = .932; proactive: ɣ = -4.116, SE = 1.042, ,
see Figure 2, panels “H,” “J,” and “L”) and (e) job satisfaction (ɣ = -2.899, SE = .985, see Figure
2, panel “N”).
To follow up these findings, we additionally considered estimates of marginal effects

WORKING FROM HOME
31
across levels of the percentage of time working from home, which allow us to examine the form
of nonlinear relationships more clearly (e.g., Simonsohn, 2018). To qualify as an inverse U-
shaped quadratic effect, marginal effects should be most strongly positive when the percentage
of time working from home is low, become less strongly positive as this value increases, and
eventually “tip” to negative (see Figure 1). Indeed, we find this general pattern of marginal
effects in several instances here (see Table 5 and Figure 2). Specifically, within-person
percentage of time working from home had an inverse U-shaped relation with work performance
(i.e., task, adaptive, and proactive performance) and job satisfaction (see Figure 2, panels “H,”
“J,” “L,” & “N”). A complete summary of the marginal effects for these outcomes are
summarized in Table 5. To better contextualize the form of these relations against one-another,
Figure 2 plots all nonlinear effects from each model tested, regardless of statistical significance.
Sensitivity Analysis
We considered additional tests of covariate sensitivity and for sensitivity to the
systematic effects of time. Specifically, regarding covariates, we re-specified our focal models
additionally modeling control variables (i.e., chronological age, sex, educational attainment,
industry, contract hours, at-home childcare responsibilities, lockdown periods). Table 3
summarizes these models. In short, the substantive conclusions from our focal analyses were
upheld with the additional inclusion of these demographic control variables, bolstering our
confidence in the reported findings (Becker, 2005). Additionally, regarding sensitivity to the
systematic effects of time, we re-specified our focal models additionally considering interactions
between working from home variables (i.e., between- and within-person) and time. In summary,
although there were some statistically-significant differential time-graded patterns associated
with working from home, the resulting trajectories across time, taken together, by-and-large

WORKING FROM HOME
32
align with our focal results that control for unconditional, exogenous effects of time. A complete
accounting of this analysis is available in our online supplemental materials.
Exploratory Analysis of Moderators
To address our research question regarding working from home resources as “buffers” of
nonlinear between- and within-person relations of percent time working from home and the
outcomes considered here, we considered models with such resources as moderators. To
maximize our sample size when testing these conditional effects (i.e., given observed
missingness on certain exogenous covariates), we specified these models without the previously-
mentioned covariates. However, summaries of sensitivity analyses with covariates included are
available in our online supplementary materials. We initially observe several statistically
significant (p < .05) and notable main effects of working from home resources in these models
(see Table 4). For example, experience with working from home before the pandemic was
associated with higher levels of work from home self-efficacy (ɣ = .205, SE = .067), having
appropriate technology and equipment to work from home was associated with lower
professional isolation (ɣ = -.326, SE = .089) and higher work from home satisfaction (ɣ = .374,
SE = .095), work from home self-efficacy (ɣ = .502, SE = .096), and task performance (ɣ = .490,
SE = .106). Lastly, having a dedicated workspace at home was associated with higher levels of
work from home satisfaction (ɣ = .187, SE = .054), work from home self-efficacy (ɣ = .146, SE =
.054), adaptive performance (ɣ = .225, SE = .065) and proactive performance (ɣ = .372, SE =
.078), and job satisfaction (ɣ = .265, SE = .076).
Several statistically significant (p < .05) interactions between percentage of time working
from home and these resources were likewise observed here, and are summarized in Table 4,
depicted graphically in Figure 3, and described in terms of conditional marginal effects in Tables

WORKING FROM HOME
33
6 and 7 (n.b., given the exploratory nature of these analysis, we do not elaborate fully on each
observed interaction here for the sake of space). For example, we found that having proper
equipment to work from home moderated the nonlinear relation between between-person
percentage of time working from home and professional isolation (ɣ = 38.339, SE = 11.408). The
predicted slopes associated with this interaction (see Figure 3, panel “B”) suggest a stronger
inverse U-shaped relation for people without such equipment, whereas there is a weaker, albeit
still inverse U-shaped relation observed for people who do possess such equipment. The
marginal effects associated with this interaction corroborate this (see Table 6).
As another example, we found that having a dedicated workspace at home moderated the
nonlinear relation between within-person percentage of time working from home and task
performance (ɣ = 5.467, SE = 1.826). The predicted slopes associated with this interaction (see
Figure 3, panel “I”) suggest a stronger inverse U-shaped relation for people without such a
dedicated workspace, and a weaker, nearly linear relation observed for people who do have a
dedicated workspace. As before, this interpretation is corroborated by the marginal effects
reported in Table 7.
Discussion
It has been suggested that there is an “ideal” amount of time for employees to experience
the benefits of working from home. Likewise, the question of whether such an ideal amount
exists has been broadly discussed as organizations try to strike a balance between “moving back
to the office” after strict pandemic restrictions and employees desire for flexible working
options. Despite this, there has been mixed evidence in the literature regarding the (assumed)
positive influence of working from home. This study was conducted to address the possibility
that working from home for a relatively large proportion of time may represent “too much of a

WORKING FROM HOME
34
good thing” for a variety of work-related outcomes. We sought to demonstrate whether inverse
U-shaped relations exist, such that working from home may present a doubled-edged sword. The
results of our study are mixed with respect to this hypothesis but are nonetheless interesting and
informative of the broader idea of the “ideal” amount of time to work from home.
In general, we found evidence for nonlinear effects of the percentage of time working
from home, however in only a few cases did we find evidence of inverse U-shaped effects that
would be indicative of “too much of a good thing.” Indeed, the results presented here generally
challenge research (e.g., Gajendran & Harrison, 2007) and practical recommendations (e.g.,
Shockley, 2014) regarding the optimal amount of time to work from home to maximize its
purported benefits. It would appear from these results that there is not one single percentage of
time that is “just right” to maximize the benefits of working from home.
For example, at the between-person level of analysis, working from home 11.63% more
than the sample average (40.20%), thus approximately 2.5 days per week, seems to maximize
professional isolation (see Figure 2, Panel “A”), whereas at the within-person level of analysis,
working from home approximately 20% to 25% more that one’s average seems to benefit work
performance. Moreover, at least in one case in our primary analysis (i.e., work performance; see
Figure 2, Panels “H,” “J,” and “L”) and variously in our exploratory analyses (see Figure 3), we
find that working from home more than one’s usual amount can, in some instances and for
individuals with different work from home resources, be detrimental to certain work outcomes.
For example, people with experience working from home may experience lower levels of task
performance when they work from home more than they do on average (see Figure 3, Panel “J”
and Table 7). This observation is, to some extent, consistent with the notion of “autonomy
calibration” (Prengler et al., 2021) and suggests that there may be intraindividual set points for

WORKING FROM HOME
35
maximizing the benefits of working from home, with positive or negative deviations therefrom
being associated with lower levels of work performance and job satisfaction.
Theoretical and Practical Implications
Our findings have several implications for theory development and for future studies on
working from home. First, we advance meta-theoretical perspectives on the “too much of a good
thing” effect by exploring the boundary conditions under which such an effect manifests when
working from home. Our findings suggest that working from home for a higher percentage of
time per week can, under certain conditions, indeed constitute “too much of a good thing.”
However, our findings suggest that whether this is the case depends on the type of outcome
considered (e.g., professional isolation vs. work from home satisfaction), the level of analysis
(i.e., between- vs. within-person), and individuals’ resource status (e.g., availability of adequate
technical equipment to facilitate working from home). Based on our findings, the “too much of a
good thing” perspective, and psychological demands and resources theories (e.g., Demerouti et
al., 2001), theory development efforts should focus on building an integrative explanatory
framework on the boundary conditions that make working from home and related flexible work
arrangements most effective, demarcating when it is “too little of a good thing,” “just right,” and
“too much of a good thing” (e.g., a “goldilocks” experience; Martin & Keyes, 2015).
Second, we contribute to the literature on flexible work arrangements and the “too much
of a good thing” phenomena by considering differential effects of the amount of time working
from home on various outcomes at the between-person and the within-person levels of analysis.
To our knowledge, no studies of “too much of a good thing” have differentiated between- from
within-person relations, making this study a novel extension of and contribution to research in
this area. Future theoretical work could extend this approach by not only considering outcomes,

WORKING FROM HOME
36
but also predictors of working from home at different conceptual levels. This approach might
include abilities, motivational factors (e.g., personality, attitudes), and non-work characteristics
(e.g., family demands and resources) as individual difference predictors, as well as work and
nonwork events as predictors at the within-person level that represent contextual opportunities
and constraints to working from home. Furthermore, this approach could be extended by
considering factors at “higher” levels, including the team (e.g., proportion of team members
working from home, leader support) and organizational levels of analysis (e.g., climates for
flexible work, flexible human resource management policies and practice).
Third, our exploratory moderator analyses suggest that specific work from home
resources, such as experience, equipment, and location, may influence the shape of the “too
much of a good thing” effect of working from home. Future theory development could advance
understanding of work from home resources by proposing a systematic taxonomy of such
person-related and contextual resources. For example, conservation of resources theory (Hobfoll,
1989) defines resources as things that individuals value, and distinguishes between resources as
objects, conditions, personal characteristics, and energies. Regarding working from home,
several such resource categories can be distinguished. For example, “object” resources are
physical entities that are valued, such as equipment, a dedicated workspace, but also less obvious
elements of one’s physical space, such as having a large living space that affords necessary
privacy. “Condition” resources include social circumstances that help employees obtain other
resources, such as being in a stable relationship, having tenure, and secure employment.
“Personal characteristics” as resources for working from home include attributes or skills that
enable employees to deal with negative emotions and help them attain goals, such as
conscientiousness, emotional stability, or social skills. Finally, “energy resources” for working

WORKING FROM HOME
37
from home may help obtain other resources but can also be valued in-and-of-themselves, such as
vitality, a sense of learning, and physical health.
This study has several implications for HRD practice and for the study of “virtual” HRD
(e.g., Bennett, 2014; McWhorter, 2010). To this end, it has been argued that working from home
is an important lens through which virtual HRD can be understood (Bennett, 2010). Indeed, our
findings provide some evidence on when the amount of working from home may be “too much
of a good thing” for certain outcomes and, thus, should be reduced. Although research has
suggested that working from home for approximately 40% of the work week (i.e., two days) is
optimal for job satisfaction (e.g., Gajendran & Harrison, 2007; Golden, 2006), our study does not
necessarily support this cut-off for other important work outcomes. Our findings regarding
within-person effects suggest that the influence of percentage of time working from home
reflects a largely individualized experience, and that it may benefit employees and supervisors to
actively negotiate and discuss work from home arrangements to the benefit of work performance
and job satisfaction. This suggests that both supervisors and their employees need to be aware of
both the positive and negative sides of working from home (e.g., see Table 1 in Bennett, 2010)
and maintain active lines of communication to discuss employee’s experiences of working from
home, which can be facilitated through virtual HRD interventions. As suggested by Bennett
(2010), “The field of HRD must ensure that VHRD [virtual HRD] does not add to workplace
stress but rather allows for greater confluence of work and learning, free of some of the typical
frustrations of telecommuting” (p. 732).
Moreover, our findings suggest that certain actionable work from home resources (i.e.,
experience, technologies, workspace) could be the focus of HRD interventions to improve the
experience and effectiveness of working from home, depending on its amount (see also

WORKING FROM HOME
38
Demerouti, 2023). Organizations would be well advised to adopt an HRD approach that gives
employees opportunities to gain experience working from home, but also provide them with
appropriate technologies to do so. Organizations may also consider helping employees set up
dedicated workspaces to optimize the benefits of working from home (see also Rudolph &
Zacher, 2021). For example, organizations might consider affording employees with a one-time
or recurring (e.g., yearly) stipend to help them establish comfortable work from home spaces.
For the field of virtual HRD, this study suggests the need to carefully consider how
formal and informal learning at home may be encouraged through virtual interventions (e.g., to
improve expertise, and performance, to encourage community-building). In doing so,
organizations must be particularly attuned to issues of equity as they related to a broader ecology
of virtual HRD. For example, as argued by Bennett and Bierema (2010), ensuring the visibility
of employees who work from home plays an important role in ensuring equitable career
development for all employees, regardless of their work arrangements. Moreover, considering
the demands potentially imposed by working from home, careful attention must be paid to the
time related aspects of such interventions (e.g., timing, time commitment) so as not to
overburden employees (i.e., thereby making a “good thing” a “bad thing”). Moreover, our
findings regarding the buffering influence of work from home equipment suggests that affording
employees with appropriate technologies may help optimize their experiences of working from
home. This view is consistent with systems-based theoretical models that position the affordance
of technology as a key resource that drives organizational effectiveness (e.g., Rummler &
Brache, 1995).
Limitations and Future Research
Despite the strengths of this work (e.g., a strong theoretical and empirical grounding, the

WORKING FROM HOME
39
adoption of a longitudinal research design), this study has several limitations that could be
addressed in future research. First, all variables were measured using self-reports. To avoid the
potential problem of common method bias, we followed recommendations by methodologists
(Podsakoff et al., 2003) and collected relatively unambiguous measures with different response
options across multiple time intervals. Nevertheless, future research could also collect objective
data (e.g., company data on who worked “where and when”) and other-reports on working from
home (e.g., from supervisors, coworkers, spouses) to corroborate self-reports of time spent
working from home.
Second, many of our constructs were measured using single items. On the one hand, such
economic and practical measures can be justified, especially in longitudinal surveys, when they
refer to rather homogenous constructs (e.g., the amount of and experience with work from home;
Fisher et al., 2016; Matthews et al., 2022). On the other hand, single item measures may be
criticized for not capturing complex constructs well (i.e., potentially exhibiting low content
validity), for fewer points of discrimination (i.e., lower sensitivity), and for lacking an estimate
of internal-consistency reliability (see McIver & Carmines, 1981). However, to the last critique,
we note acceptable levels of test-retest reliability across such items reported here. Still, future
research should consider including multi-item measures of these constructs, to the extent that it is
possible (i.e., practical and reasonable, given other study design characteristics).
Third, we considered only one relatively broad indicator of the amount of working from
home in this study, that is, the percentage of time in an average week that peopled worked from
home (measured retrospectively each month). It may be possible to use more fine-grained
assessments of the duration of working from home in future research, such as collecting data
every day on the number of hours that people worked from home (including after hours; e.g.,

WORKING FROM HOME
40
Kühner et al., 2023). Still, our approach of directly assessing percentage of time working from
home has analytic advantages over related approaches (e.g., the computed ratio of time working
from home to time worked total), which represent challenging constructions of composite
variables (see Cohen et al., 2014; section 2.10.5). Furthermore, we included relatively limited
sets of work outcomes and moderators in this study. Future research could consider additional
outcomes, such as work engagement and exhaustion, as well as additional work from home
resources based on theoretical considerations regarding different resource categories (i.e.,
objects, conditions, personal characteristics, and energies; see Hobfoll, 1989). Moreover, we
acknowledge the potential for reciprocal relations between working from home and several of
our assumed outcomes. For example, regarding work from home self-efficacy, the frequency of
working from home might potentially be both a predictor and an outcome of this form of self-
efficacy (see Sitzmann & Yeo, 2013). Although not the focus of our investigation, future
research would benefit from considering such reinforcing relations.
Finally, this study was conducted during a time when many people were given the option
to work from home, and may not generalize to other employees/contexts, especially beyond the
relatively restrictive conditions imposed to curtail the spread of COVID-19 (Zacher & Rudolph,
2022). Still, it is important to note that our findings are consistent with other research conducted
prior to the COVID-19 pandemic that likewise speaks to nonlinear influences of working from
home on work outcomes (e.g., Gajendran & Harrison, 2007; Virick et al., 2010).
Conclusion
Based on an integration of the meta-theoretical perspective of “too much of a good thing”
and psychological demands and resources theories, we examined nonlinear relations between the
percentage of time people work from home and several important work-related outcomes. At the

WORKING FROM HOME
41
between-person level of analysis, we found evidence for a “too much of a good thing” effect for
professional isolation. At the within-person level of analysis, we found more robust evidence for
this effect for work performance (i.e., task, adaptive, proactive performance) and job satisfaction.
Furthermore, based on psychological resource theories, we explored work from home resources
(i.e., previous experience working from home, appropriate technologies to facilitate working
from home, dedicated workspaces) as moderators. Results of these exploratory moderator
analyses were likewise mixed, suggesting for example that the “too much of a good thing” effect
may occur in terms of reduced job satisfaction among employees with less experience working
from home and who work from home more or less than they typically do. While these
phenomena are clearly complex, our hope is that this study inspires researchers to further
investigate the “double edged sword” that working from home presents to employees, and that
our conclusions inform the continued development of meta-theoretical perspectives regarding
“too much of a good thing.”

WORKING FROM HOME
42
References
Allen, T. D., Golden, T. D., & Shockley, K. M. (2015). How effective is telecommuting?
Assessing the status of our scientific findings. Psychological Science in the Public
Interest, 16(2), 40–68. https://doi.org/10.1177/1529100615593273
Anakpo, G., Nqwayibana, Z., & Mishi, S. (2023). The Impact of Work-from-Home on Employee
Performance and Productivity: A Systematic Review. Sustainability, 15(5), 4529.
https://doi.org/10.3390/su15054529
Antonakis, J., House, R. J., & Simonton, D. K. (2017). Can super smart leaders suffer from too
much of a good thing? The curvilinear effect of intelligence on perceived leadership
behavior. Journal of Applied Psychology, 102(7), 1003–1021.
https://doi.org/10.1037/apl0000221
Avanzi, L., Savadori, L., Fraccaroli, F., Ciampa, V., & van Dick, R. (2020). Too-much-of-a-
good-thing? The curvilinear relation between identification, overcommitment, and
employee well-being. Current Psychology, 1-11. https://doi.org/10.1007/s12144-020-
00655-x
Bakker, A. B., & Demerouti, E. (2014). Job demands-resources theory. In P. Y. Chen, & C. L.
Cooper (Eds.), Wellbeing: A complete reference guide. Work and wellbeing (pp. 37–64).
Wiley-Blackwell.
Bakker, A. B., Demerouti, E., & Sanz-Vergel, A. (2023). Job demands–resources theory: Ten
years later. Annual Review of Organizational Psychology and Organizational
Behavior, 10(1), 25–53. https://doi.org/10.1146/annurev-orgpsych-120920-053933
Baltes, B. B., Bauer, C. C., Bajdo, L. M., & Parker, C. P. (2002). The use of multitrait-
multimethod data for detecting nonlinear relationships: The case of psychological climate

WORKING FROM HOME
43
and job satisfaction. Journal of Business and Psychology, 17, 3-
17. https://doi.org/10.1023/A:1016231816394
Bandura, A. (2006). Guide to construction of self-efficacy scales. In F. Pajares & T. Urdan
(Eds.), Self-efficacy Beliefs of Adolescents (Vol. 5, pp. 307–337). Greenwich, CT:
Information Age.
Baron, R. A. (1986). Self‐presentation in job interviews: When there can be “too much of a good
thing” Journal of Applied Social Psychology, 16(1), 16-28.
https://doi.org/10.1111/j.1559-1816.1986.tb02275.x
Beaujean, A. A. (2014). Latent variable modeling using R: A step-by-step
guide. Routledge/Taylor & Francis Group. https://doi.org/10.4324/9781315869780
Becker, T. E. (2005). Potential problems in the statistical control of variables in organizational
research: A qualitative analysis with recommendations. Organizational Research
Methods, 8(3), 274-289. https://doi.org/10.1177/1094428105278021
Becker, T. E., Atinc, G., Breaugh, J. A., Carlson, K. D., Edwards, J. R., & Spector, P. E. (2016).
Statistical control in correlational studies: 10 essential recommendations for
organizational researchers. Journal of Organizational Behavior, 37(2), 157-167.
https://doi.org/10.1002/job.2053
Behrens, K., Kichko, S., & Thisse, J.-F. (2021). Working from home: Too much of a good
thing? [SSRN Scholarly Paper]. https://doi.org/10.2139/ssrn.3768910
Bell, C. J. (2020). Feeling remote: Factors influencing isolation in remote workers (Publication
no. 28002722). [Doctoral dissertation, Middle Tennessee State University].
Bennett, E. E. (2010). The coming paradigm shift: Synthesis and future directions for virtual
HRD. Advances in Developing Human Resources, 12(6), 728-741.

WORKING FROM HOME
44
https://doi.org/10.1177/1523422310394796
Bennett, E. E. (2014). Introducing new perspectives on virtual human resource
development. Advances in Developing Human Resources, 16(3), 263-280.
https://doi.org/10.1177/152342231453209
Bennett, E. E., & Bierema L. L. (2010). The ecology of virtual HRD. Advances in Developing
Human Resources, 12, 632-647. https://doi.org/10.1177/15234223103947
Bennett, E. E., & McWhorter, R. R. (2021).Virtual HRD’s role in crisis and the post COVID-19
professional life world: Accelerating skills for digital transformation. Advances in
Developing Human Resources, 23(1), 5–25.
https://journals.sagepub.com/doi/pdf/10.1177/ 1523422320973288
Biron, M., Casper, W. J., & Raghuram, S. (2023). Crafting telework: A process model of need
satisfaction to foster telework outcomes. Personnel Review, 52(3). 671-686.
https://doi.org/10.1108/PR-04-2021-0259
Bolger, N., & Laurenceau, J. P. (2013). Intensive longitudinal methods: An introduction to diary
and experience sampling research. Guilford press.
Breaugh, J. A. (2006). Rethinking the control of nuisance variables in theory testing. Journal of
Business and Psychology, 20, 429-443. https://doi.org/10.1007/s10869-005-9009-y
Brion, A., & Westhues, J. (2020). Working from home in Germany: What you need to know.
Qvive. Accessed from: https://www.qivive.com/en/knowledge/publications/working-
home-germany-what-you-need-know
Campbell, J. P., & Wiernik, B. M. (2015). The modeling and assessment of work performance.
Annual Review of Organizational Psychology and Organizational Behavior, 2(1), 47-74.
https://doi.org/10.1146/annurev-orgpsych-032414-111427

WORKING FROM HOME
45
Carnevale, J. B., & Hatak, I. (2020). Employee adjustment and well-being in the era of COVID-
19: Implications for human resource management. Journal of Business Research, 116,
183-187. https://doi.org/10.1016/j.jbusres.2020.05.037
Champoux, J. E. (1992). A multivariate analysis of curvilinear relationships among job scope,
work context satisfactions, and affective outcomes. Human Relations, 45(1), 87–111.
https://doi.org/10.1177/001872679204500105
Cohen, P., West, S. G., & Aiken, L. S. (2014). Applied multiple regression/correlation analysis
for the behavioral sciences. Psychology Press.
Craig, L., & Churchill, B. (2021). Working and caring at home: Gender differences in the effects
of COVID-19 on paid and unpaid labor in Australia. Feminist Economics, 27(1-2), 310-
326. https://doi.org/10.1080/13545701.2020.1831039
Deci, E. L., & Ryan, R. M. (2013). Intrinsic Motivation and Self-Determination in Human
Behavior. Springer.
Demerouti, E. (2023). Effective employee strategies for remote working: An online self-training
intervention. Journal of Vocational Behavior, https://doi.org/10.1016/j.jvb.2023.103857.
Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-
resources model of burnout. Journal of Applied Psychology, 86(3), 499–512.
https://doi.org/10.1037/0021-9010.86.3.499
Diener, E., Lucas, R. E., & Scollon, C. N. (2009). Beyond the hedonic treadmill: Revising the
adaptation theory of well-being. In E. Diener (Ed.), The Science of Well-Being: The
Collected Works of Ed Diener (pp. 103–118). Springer Netherlands.
https://doi.org/10.1007/978-90-481-2350-6_5
Dierdorff, E. C., & Jensen, J. M. (2018). Crafting in context: Exploring when job crafting is

WORKING FROM HOME
46
dysfunctional for performance effectiveness. Journal of Applied Psychology, 103(5),
463–477. https://doi.org/10.1037/apl0000295
Dolbier, C. L., Webster, J. A., McCalister, K. T., Mallon, M. W., & Steinhardt, M. A. (2005).
Reliability and validity of a single-item measure of job satisfaction. American Journal of
Health Promotion, 19(3), 194–198. https://doi.org/10.4278/0890-1171-19.3.194
Enders, C. K. (2022). Applied missing data analysis. Guilford Publications.
Fernet, C., Austin, S., Trépanier, S. G., & Dussault, M. (2013). How do job characteristics
contribute to burnout? Exploring the distinct mediating roles of perceived autonomy,
competence, and relatedness. European Journal of Work and Organizational Psychology,
22(2), 123-137. https://doi.org/10.1080/1359432X.2011.632161
Fisher, G. G., Matthews, R. A., & Gibbons, A. M. (2016). Developing and investigating the use
of single-item measures in organizational research. Journal of Occupational Health
Psychology, 21(1), 3–23. https://doi.org/10.1037/a0039139
Fonner, K. L., & Roloff, M. E. (2010). Why teleworkers are more satisfied with their jobs than
are office-based workers: When less contact is beneficial. Journal of Applied
Communication Research, 38(4), 336-361.
https://doi.org/10.1080/00909882.2010.513998
Frodermann, C., Grunau, P., Haas, G.-C., & Müller, D. (2021). Homeoffice in Zeiten von
Corona: Nutzung, Hindernisse und Zukunftswünsche [Working from home in times of
the Covid-19 pandemic: Extent, challenges and potentials for the future]. IAB-
Kurzbericht. https://ideas.repec.org//p/iab/iabkbe/202105.html
Gagné, M., & Deci, E. L. (2005). Self‐determination theory and work motivation. Journal of
Organizational Behavior, 26(4), 331-362. https://doi.org/10.1002/job.322

WORKING FROM HOME
47
Gagné, M., Parker, S. K., Griffin, M. A., Dunlop, P. D., Knight, C., Klonek, F. E., & Parent-
Rocheleau, X. (2022). Understanding and shaping the future of work with self-
determination theory. Nature Reviews Psychology, 1(7), 378–392.
https://doi.org/10.1038/s44159-022-00056-w
Gajendran, R. S., & Harrison, D. A. (2007). The good, the bad, and the unknown about
telecommuting: Meta-analysis of psychological mediators and individual
consequences. Journal of Applied Psychology, 92(6), 1524–
1541. https://doi.org/10.1037/0021-9010.92.6.1524
Galanti, T., Guidetti, G., Mazzei, E., Zappalà, S., & Toscano, F. (2021). Work from home during
the COVID-19 outbreak: The impact on employees’ remote work productivity,
engagement, and stress. Journal of Occupational and Environmental Medicine, 63(7),
e426. https://doi.org/10.1097/JOM.0000000000002236
Gardner, D. G. (1986). Activation theory and task design: An empirical test of several new
predictions. Journal of Applied Psychology, 71(3), 411-418.
https://doi.org/10.1037/0021-9010.71.3.411
Gardner, D. G., & Cummings, L. L. (1988). Activation theory and task design: Review and
reconceptualization. In B. M. Staw & L. L. Cummings (Eds.), Research in
Organizational Behavior (Vol. 10). JAI Press Inc.
Golden, T. D. (2006). The role of relationships in understanding telecommuter satisfaction.
Journal of Organizational Behavior, 27(3), 319-340. https://doi.org/10.1002/job.369
Golden, T. D., & Eddleston, K. A. (2020). Is there a price telecommuters pay? Examining the
relationship between telecommuting and objective career success. Journal of Vocational
Behavior, 116, 103348. https://doi.org/10.1016/j.jvb.2019.103348

WORKING FROM HOME
48
Golden, T. D., & Veiga, J. F. (2005). The impact of extent of telecommuting on job satisfaction:
Resolving inconsistent findings. Journal of Management, 31(2), 301-318.
https://doi.org/10.1177/0149206304271768
Golden, T. D., Veiga, J. F., & Dino, R. N. (2008). The impact of professional isolation on
teleworker job performance and turnover intentions: Does time spent teleworking,
interacting face-to-face, or having access to communication-enhancing technology
matter? Journal of Applied Psychology, 93(6), 1412–1421.
https://doi.org/10.1037/a0012722
Goodman, J. S., & Blum, T. C. (1996). Assessing the non-random sampling effects of subject
attrition in longitudinal research. Journal of Management, 22(4), 627-652.
https://doi.org/10.1016/S0149-2063(96)90027-6
Gorgievski, M. J., Halbesleben, J. R., & Bakker, A. B. (2011). Expanding the boundaries of
psychological resource theories. Journal of Occupational and Organizational
Psychology, 84(1), 1-7. https://doi.org/10.1111/j.2044-8325.2010.02015.x
Graham, M., Lambert, K. A., Weale, V., Stuckey, R., & Oakman, J. (2023). Working from home
during the COVID 19 pandemic: a longitudinal examination of employees’ sense of
community and social support and impacts on self-rated health. BMC Public
Health, 23(1), 11. https://doi.org/10.1186/s12889-022-14904-0
Grant, A. M., & Schwartz, B. (2011). Too much of a good thing: The challenge and opportunity
of the inverted U. Perspectives on Psychological Science, 6(1), 61–76.
https://doi.org/10.1177/1745691610393523
Griffin, M. A., Neal, A., & Parker, S. K. (2007). A new model of work role performance:
Positive behavior in uncertain and interdependent contexts. Academy of Management

WORKING FROM HOME
49
Journal, 50(2), 327–347. https://doi.org/10.5465/AMJ.2007.24634438
Hackney, A., Yung, M., Somasundram, K. G., Nowrouzi-Kia, B., Oakman, J., & Yazdani, A.
(2022). Working in the digital economy: A systematic review of the impact of work from
home arrangements on personal and organizational performance and productivity. PLOS
ONE, 17(10), e0274728. https://doi.org/10.1371/journal.pone.0274728
Harris, K. J., & Kacmar, K. M. (2006). Too much of a good thing: The curvilinear effect of
leader-member exchange on stress. The Journal of Social Psychology, 146(1), 65-84.
https://doi.org/10.3200/SOCP.146.1.65-84
Hayes, S. W., Priestley, J. L., Moore, B. A., & Ray, H. E. (2021). Perceived stress, work-related
burnout, and working from home before and during covid-19: An examination of workers
in the United States. SAGE Open, 11(4), 215824402110581.
https://doi.org/10.1177/21582440211058193
Heady, B., & Wearing, A. (1989). Personality, life event and subjective well-being: Toward a
dynamic equilibrium model. Journal of Personality and Social Psychology, 57, 731-739.
http://dx.doi.org/10.1037/0022-3514.57.4.731
Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress.
American Psychologist, 44, 513–524. https://doi.org/10.1037/0003-066X.44.3.513
Hobfoll, S. E. (2002). Social and psychological resources and adaptation. Review of General
Psychology, 6, 307–324. https://doi.org/10.1037/1089-2680 .6.4.307
Hobfoll, S. E., Halbesleben, J., Neveu, J. P., & Westman, M. (2018). Conservation of resources
in the organizational context: The reality of resources and their consequences. Annual
Review of Organizational Psychology and Organizational Behavior, 5, 103-128.
https://doi.org/10.1146/annurev-orgpsych-032117-104640

WORKING FROM HOME
50
Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too much
of a good thing? Journal of Personality and Social Psychology, 79(6), 995–
1006. https://doi.org/10.1037/0022-3514.79.6.995
Kaiser, S., Suess, S., Cohen, R., Mikkelsen, E. N., & Pedersen, A. R. (2022). Working from
home: Findings and prospects for further research. German Journal of Human Resource
Management, 36(3), 205-212. https://doi.org/10.1177/23970022221106973
Kennedy, W. J., & Gentle, J. E. (1980). Statistical Computing. Routledge.
Kniffin, K. M., Narayanan, J., Anseel, F., Antonakis, J., Ashford, S. P., Bakker, A. B.,
Bamberger, P., Bapuji, H., Bhave, D. P., Choi, V. K., Creary, S. J., Demerouti, E., Flynn,
F. J., Gelfand, M. J., Greer, L. L., Johns, G., Kesebir, S., Klein, P. G., Lee, S. Y., . . .
Vugt, M. v. (2021). COVID-19 and the workplace: Implications, issues, and insights for
future research and action. American Psychologist, 76(1), 63–77.
https://doi.org/10.1037/amp0000716
Kubicek, B., Korunka, C., & Tement, S. (2014). Too much job control? Two studies on
curvilinear relations between job control and eldercare workers’ well-being. International
Journal of Nursing Studies, 51(12), 1644-1653.
https://doi.org/10.1016/j.ijnurstu.2014.05.005
Kühner, C., Rudolph, C. W., Derks, D., & Zacher, H. (2023). Technology-assisted supplemental
work: A meta-analysis. Journal of Vocational Behavior.
https://doi.org/10.1016/j.jvb.2023.103861
Langfred, C. W. (2004). Too much of a good thing? Negative effects of high trust and individual
autonomy in self-managing teams. Academy of Management Journal, 47(3), 385-399.
https://doi.org/10.5465/20159588

WORKING FROM HOME
51
Le, H., Oh, I.-S., Robbins, S. B., Ilies, R., Holland, E., & Westrick, P. (2011). Too much of a
good thing: Curvilinear relationships between personality traits and job
performance. Journal of Applied Psychology, 96(1), 113–
133. https://doi.org/10.1037/a0021016
Lewis, S. (2003). Flexible working arrangements: Implementation, outcomes, and management.
International Review of Industrial and Organizational Psychology 2003, 18, 1-28.
https://doi.org/10.1002/0470013346.ch1
Martin, C. C., & Keyes, C. L. M. (2015). Investigating the goldilocks hypothesis: The non-linear
impact of positive trait change on well-being. PLOS ONE, 10(7), e0131316.
https://doi.org/10.1371/journal.pone.0131316
Matthews, R. A., Pineault, L., & Hong, Y. H. (2022). Normalizing the use of single-item
measures: Validation of the single-item compendium for organizational psychology.
Journal of Business and Psychology, 37(4), 639-673. https://doi.org/10.1007/s10869-022-
09813-3
Mattis, M. C. (1990). New forms of flexible work arrangements for managers and professionals:
Myths and realities. Human Resource Planning, 13(2), 133-146
McIver, J., & Carmines, E. G. (1981). Unidimensional scaling (Vol. 24). SAGE.
McWhorter, R. R. (2010). Exploring the emergence of virtual human resource
development. Advances in Developing Human Resources, 12(6), 623-631.
https://doi.org/10.1177/1523422310395367
Meister, J. (2023). Top Ten HR Trends For The 2023 Workplace. Forbes. From:
https://www.forbes.com/sites/jeannemeister/2023/01/10/top-ten-hr-trends-for-the-2023-
workplace/?sh=339aacec5933

WORKING FROM HOME
52
Nakrošienė, A., Bučiūnienė, I., & Goštautaitė, B. (2019). Working from home: Characteristics
and outcomes of telework. International Journal of Manpower, 40(1), 87–101.
https://doi.org/10.1108/IJM-07-2017-0172
Newman, D. A. (2009). Missing data techniques and low response rates: The role of systematic
nonresponse parameters. In C. E. Lance & R. J. Vandenberg (Eds.), Statistical and
methodological myths and urban legends: Doctrine, verity, and fable in the organizational
and social sciences (pp. 7–36). New York, NY: Routledge.
Niebuhr, F., Borle, P., Börner-Zobel, F., & Voelter-Mahlknecht, S. (2022). Healthy and happy
working from home? Effects of working from home on employee health and job
satisfaction. International Journal of Environmental Research and Public Health, 19(3),
1122. https://doi.org/10.3390/ijerph19031122
Oakman, J., Kinsman, N., Lambert, K., Stuckey, R., Graham, M., & Weale, V. (2022). Working
from home in Australia during the COVID-19 pandemic: Cross-sectional results from the
Employees Working From Home (Ewfh) study. BMJ Open, 12(4), e052733.
https://doi.org/10.1136/bmjopen-2021-052733
Park, J. W., Park, S., & Cho, Y. J. (2023). More isn’t always better: Exploring the curvilinear
effects of telework. International Public Management Journal, 26, 744-763,
https://doi.org/10.1080/10967494.2023.2214133
Peterson, R. S. (1999). Can you have too much of a good thing? The limits of voice for
improving satisfaction with leaders. Personality and Social Psychology Bulletin, 25(3),
313–324. https://doi.org/10.1177/0146167299025003004
Pierce, J. R., & Aguinis, H. (2013). The too-much-of-a-good-thing effect in management.
Journal of Management, 39(2), 313–338. https://doi.org/10.1177/0149206311410060

WORKING FROM HOME
53
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi.org/10.1037/0021-
9010.88.5.879
Prengler, M., Klotz, A., & Murphy, C. B. (2021). A grounded model of autonomy calibration in
location-independent work arrangements. Academy of Management
Proceedings, 2021(1), 10281. https://doi.org/10.5465/AMBPP.2021.152
Raghuram, S., Wiesenfeld, B., & Garud, R. (2003). Technology enabled work: The role of self-
efficacy in determining telecommuter adjustment and structuring behavior. Journal of
Vocational Behavior, 63(2), 180-198. https://doi.org/10.1016/S0001-8791(03)00040-X
Rudolph, C. W., Allan, B., Clark, M., Hertel, G., Hirschi, A., Kunze, F., Shockley, K., Shoss,
M., Sonnentag, S., & Zacher, H. (2021). Pandemics: Implications for research and
practice in industrial and organizational psychology. Industrial and Organizational
Psychology, 14(1-2), 1-35. https://doi.org/10.1017/iop.2020.48
Rudolph, C. W., & Zacher, H. (2021). Employee wellbeing in the face of a pandemic:
organizational and managerial responses to COVID-19. APA Division 14: Society for
Industrial & Organizational Psychology Whitepaper. Available from:
https://www.siop.org/
Rummler, G. A., & Brache, A. P. (1995). Improving Performance: How to Manage the White
Space on the Organizational Chart. Jossey-Bass.
Ryan, R. M., & Deci, E. L. (2009). The darker and brighter sides of human existence: Basic
psychological needs as a unifying concept. Psychological Inquiry, 11, 319-338.
https://doi.org/10.1207/S15327965PLI1104_03

WORKING FROM HOME
54
Schwartz, B. (2005). The paradox of choice: Why more is less. Harper-Collins.
Sewell, G., & Taskin, L. (2015). Out of sight, out of mind in a new world of work? Autonomy,
control, and spatiotemporal scaling in telework. Organization Studies, 36(11), 1507-
1529. https://doi.org/10.1177/0170840615593587
Shockley, K. M. (2014). Telecommuting (Society for Industrial and Organizational Psychology
White Paper Series).
https://www.siop.org/Portals/84/docs/White%20Papers/ScientificAffairs/telecommuting.
pdf
Shockley, K. M., Allen, T. D., Dodd, H., & Waiwood, A. M. (2020). Rapid transition to remote
work during COVID-19: A study of predictors of employee well-being and productivity.
https://iwillugaresearch.wixsite.com/website/publications
Simonsohn, U. (2018). Two lines: A valid alternative to the invalid testing of u-shaped
relationships with quadratic regressions. Advances in Methods and Practices in
Psychological Science, 1(4), 538–555. https://doi.org/10.1177/2515245918805755
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and
event occurrence. Oxford University Press.
Sitzmann, T., & Yeo, G. (2013). A meta‐analytic investigation of the within‐person self‐efficacy
domain: Is self‐efficacy a product of past performance or a driver of future performance?
Personnel Psychology, 66(3), 531-568. https://doi.org/10.1111/peps.12035
Song, Y., & Gao, J. (2020). Does telework stress employees out? A study on working at home
and subjective well-being for wage/salary workers. Journal of Happiness Studies, 21(7),
2649–2668. https://doi.org/10.1007/s10902-019-00196-6
Spilker, M. A., & Breaugh, J. A. (2021). Potential ways to predict and manage telecommuters’

WORKING FROM HOME
55
feelings of professional isolation. Journal of Vocational Behavior, 131, 103646.
https://doi.org/10.1016/j.jvb.2021.103646
Stephens, G. K., & Szajna, B. (1998). Perceptions and expectations of why people choose a
telecommuting work style. In the 31st Annual Hawaii International Conference on System
Science. Kohala Coast, Hawaii: HICSS.
Stiglbauer, B., & Kovacs, C. (2018). The more, the better? Curvilinear effects of job autonomy
on well-being from vitamin model and PE-fit theory perspectives. Journal of
Occupational Health Psychology, 23(4), 520–536. https://doi.org/10.1037/ocp0000107
ten Brummelhuis, L. L., & Bakker, A. B. (2012). A resource perspective on the work–home
interface: The work–home resources model. American Psychologist, 67(7), 545–
556. https://doi.org/10.1037/a0027974
Terry, D. J., & Jimmieson, N. L. (1999). Work control and employee well-being: A decade
review. In C. L. Cooper & I. T. Robertson (Eds.), International Review of Industrial and
Organizational Psychology. Wiley.
Tjur, T. (2009). Coefficients of determination in logistic regression models - A new proposal:
The coefficient of discrimination. The American Statistician, 63(4), 366-372.
https://doi.org/10.1198/tast.2009.08210
Vander Elst, T., Verhoogen, R., Sercu, M., Van den Broeck, A., Baillien, E., & Godderis, L.
(2017). Not extent of telecommuting, but job characteristics as proximal predictors of
work-related well-being. Journal of Occupational and Environmental Medicine, 59(10),
e180–e186. https://doi.org/10.1097/JOM.0000000000001132
Van der Lippe, T., & Lippeńyi, Z. (2020). Co‐workers working from home and individual and
team performance. New Technology, Work and Employment, 35(1), 60-79.

WORKING FROM HOME
56
https://doi.org/10.1111/ntwe.12153
Virick, M., DaSilva, N., & Arrington, K. (2010). Moderators of the curvilinear relation between
extent of telecommuting and job and life satisfaction: The role of performance outcome
orientation and worker type. Human Relations, 63(1), 137–154.
https://doi.org/10.1177/0018726709349198
Vleeshouwers, J., Fløvik, L., Christensen, J. O., Johannessen, H. A., Bakke Finne, L., Mohr, B.,
Jørgensen, I. L., & Lunde, L.-K. (2022). The relationship between telework from home
and the psychosocial work environment: A systematic review. International Archives of
Occupational and Environmental Health, 95(10), 2025–2051.
https://doi.org/10.1007/s00420-022-01901-4
Wanous, J. P., Reichers, A. E., & Hudy, M. J. (1997). Overall job satisfaction: How good are
single-item measures? Journal of Applied Psychology, 82(2), 247–
252. https://doi.org/10.1037/0021-9010.82.2.247
Warr, P. (1987). Work, unemployment, and mental health. Oxford University Press.
Warr, P. (1994). A conceptual framework for the study of work and mental health. Work &
Stress, 8(2), 84-97. https://doi.org/10.1080/02678379408259982
Warr, P. (2002). The study of well-being, behaviour and attitudes. In P. Warr (Ed.), Psychology
at Work (5th ed.). Penguin.
Warr, P. (2016). Happiness and mental health: A framework of vitamins in the environment and
mental processes in the person. In J. C. Quick & C. L. Cooper (Eds.), Handbook of stress
and health: A guide to research and practice. Wiley.
Weiss, D., Weiss, M., Rudolph, C. W., & Zacher, H. (2022). Tough times at the top:
Occupational status predicts changes in job satisfaction in times of crisis. Journal of

WORKING FROM HOME
57
Vocational Behavior, 139(1), 103804. https://doi.org/10.1016/j.jvb.2022.103804
Weitzer, J., Papantoniou, K., Seidel, S., Klösch, G., Caniglia, G., Laubichler, M., Bertau, M.,
Birmann, B. M., Jäger, C. C., Zenk, L., Steiner, G., & Schernhammer, E. (2021).
Working from home, quality of life, and perceived productivity during the first 50-day
COVID-19 mitigation measures in Austria: A cross-sectional study. International
Archives of Occupational and Environmental Health, 94(8), 1823–1837.
https://doi.org/10.1007/s00420-021-01692-0
Yam Mei Har, C. (2021). Examining the Curvilinear Effects of Autonomy at Work (Doctoral
dissertation, Curtin University).
Zacher, H., & Rudolph, C. W. (2020). How a dynamic way of thinking can challenge existing
knowledge in organizational behavior. In Y. Griep & S. D. Hansen (Eds.), Handbook on
the temporal dynamics of organizational behavior (Vol. 1, pp. 8-25). Cheltenham, UK:
Edward Elgar. https://doi.org/10.4337/9781788974387.00009
Zacher, H., Rudolph, C. W., & Posch, M. (2021). Individual differences and changes in self-
reported work performance during the early stages of the COVID-19 pandemic.
Zeitschrift für Arbeits- und Organisationspsychologie, 65(4). https://doi.org/10.026/0932-
4089/a000365
Zacher, H., & Rudolph, C. W. (2022). Researching employee experiences and behavior in times
of crisis: Theoretical and methodological considerations and implications for human
resource management. German Journal of Human Resource Management, 36(1), 6-31.
https://doi.org/10.1177/23970022211058812
Zhou, E. (2020). The “too-much-of-a-good-thing” effect of job autonomy and its explanation
mechanism. Psychology, 11(2), 299-313. https://doi.org/10.4236/psych.2020.112019

WORKING FROM HOME
58
Table 1. Descriptive Statistics For Complete, Incomplete, and Panel Responders
Complete
Incomplete
P-Value
Overall
(N=554)
(N=440)
(N=994)
Age (Years)
Mean (SD)
44.70 (10.60)
45.50 (10.70)
.234
45.10 (10.70)
Median [Min, Max]
45.00 [19.00, 71.00]
46.00 [19.00, 69.00]
45.000 [19.00, 71.00]
Missing
0 (0%)
1 (0.20%)
1 (0.1%)
Gender
Male
322 (58.10%)
244 (55.50%)
.460
566 (56.90%)
Female
232 (41.90%)
195 (44.30%)
427 (43.00%)
Missing
0 (0%)
1 (0.20%)
1 (0.10%)
Education
Lower Secondary School
21 (3.80%)
18 (4.10%)
.363
39 (3.90%)
Intermediate Secondary School
164 (29.60%)
111 (25.20%)
275 (27.70%)
Upper Secondary School
95 (17.10%)
90 (20.50%)
185 (18.60%)
College/University or Technical College
274 (49.50%)
211 (48.00%)
485 (48.80%)
Missing
0 (0%)
10 (2.30%)
10 (1.00%)
Industry
Primary & Secondary Sector
58 (10.50%)
54 (12.30%)
.356
112 (11.30%)
Tertiary Sector
496 (89.50%)
376 (85.50%)
872 (87.70%)
Missing
0 (0%)
10 (2.30%)
10 (1.00%)
At Home Childcare
No
470 (84.80%)
312 (70.90%)
<0.001
782 (78.70%)
Yes
84 (15.20%)
16 (3.60%)
100 (10.10%)
Missing
0 (0%)
112 (25.50%)
112 (11.30%)
WFH Pre-COVID
No
316 (57.00%)
191 (43.40%)
.002
507 (51.00%)
Yes
238 (43.00%)
88 (20.00%)
326 (32.80%)
Missing
0 (0%)
161 (36.60%)
161 (16.20%)
WFH Equipment
No
56 (10.10%)
51 (11.60%)
.001
107 (10.80%)
Yes
498 (89.90%)
228 (51.80%)
726 (73.00%)
Missing
0 (0%)
161 (36.60%)
161 (16.20%)
WFH Location
Other
301 (54.30%)
180 (40.90%)
.006
481 (48.40%)
Own Room
253 (45.70%)
99 (22.50%)
352 (35.40%)
Missing
0 (0%)
161 (36.60%)
161 (16.20%)
Contract Hours
Mean (SD)
39.70 (2.76)
39.30 (2.41)
.027
39.50 (2.62)
Median [Min, Max]
40.00 [35.00, 60.00]
40.00 [35.00, 60.00]
40.00 [35.00, 60.00]

WORKING FROM HOME
59
Table 1 (continued). Descriptive Statistics For Complete, Incomplete, and Panel Responders
Complete
Incomplete
P-Value
Overall
(N=554)
(N=440)
(N=994)
T3 WFH Percent of Time
Mean (SD)
47.90 (39.40)
11.70 (30.70)
<0.001
45.10 (40.00)
Median [Min, Max]
50.00 [0, 100]
0 [0, 100]
50.00 [0, 100]
Missing
0 (0%)
393 (89.30%)
393 (39.50%)
T3 Professional Isolation
Mean (SD)
2.21 (0.90)
1.88 (0.82)
.010
2.18 (0.89)
Median [Min, Max]
2.25 [1.00, 5.00]
1.75 [1.00, 3.50]
2.00 [1.00, 5.00]
Missing
0 (0%)
393 (89.30%)
393 (39.50%)
T3 WFH Satisfaction
Mean (SD)
3.81 (0.99)
4.00 (1.00)
.636
3.81 (0.99)
Median [Min, Max]
4.00 [1.00, 5.00]
4.00 [3.00, 5.00]
4.00 [1.00, 5.00]
Missing
90 (16.20%)
433 (98.40%)
523 (52.60%)
T3 WFH Self Efficacy
Mean (SD)
3.92 (0.99)
3.63 (1.51)
.601
3.91 (1.00)
Median [Min, Max]
4.00 [1.00, 5.00]
4.00 [2.00, 5.00]
4.00 [1.00, 5.00]
Missing
84 (15.20%)
432 (98.20%)
516 (51.90%)
T3 Task Performance
Mean (SD)
5.72 (1.04)
5.94 (0.802)
.056
5.74 (1.02)
Median [Min, Max]
6.00 [1.00, 7.00]
6.00 [3.67, 7.00]
6.00 [1.00, 7.00]
Missing
0 (0%)
382 (86.80%)
382 (38.40%)
T3 Adaptive Performance
Mean (SD)
5.43 (1.06)
5.37 (1.04)
.688
5.43 (1.06)
Median [Min, Max]
5.67 [1.00, 7.00]
5.33 [2.33, 7.00]
5.67 [1.00, 7.00]
Missing
0 (0%)
382 (86.80%)
382 (38.40%)
T3 Proactive Performance
Mean (SD)
4.71 (1.27)
4.35 (1.47)
.082
4.67 (1.30)
Median [Min, Max]
4.67 [1.00, 7.00]
4.33 [1.00, 7.00]
4.67 [1.00, 7.00]
Missing
0 (0%)
382 (86.80%)
382 (38.40%)
T3 Job Satisfaction
Mean (SD)
5.13 (1.29)
5.07 (1.19)
.721
5.13 (1.28)
Median [Min, Max]
5.00 [1.00, 7.00]
5.00 [2.00, 7.00]
5.00 [1.00, 7.00]
Missing
0 (0%)
384 (87.30%)
384 (38.60%)

WORKING FROM HOME
60
Table 2. Summary of Primary Hypothesis Tests
Prof. Isolation
WFH Satisfaction
WFH Self-Efficacy
Task Perf
Adapt. Perf.
Proact. Perf.
Job Satisfaction
Predictors
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
(Intercept)
2.221
<0.001
3.641
<0.001
3.734
<0.001
5.722
<0.001
5.357
<0.001
4.573
<0.001
5.091
<0.001
Time
-.005
<0.001
.004
<0.001
.001
.062
-.007
<0.001
-.010
<0.001
-.006
<0.001
.001
.093
WFH% Between
9.189
.003
60.875
<0.001
72.886
<0.001
8.724
.018
9.784
.013
5.140
.283
12.986
.005
WFH% Between.2
-26.535
<0.001
-7.859
.018
-10.937
.001
16.321
<0.001
7.327
.058
-5.679
.228
8.336
.069
WFH% Within
10.368
<0.001
9.161
<0.001
8.231
<0.001
4.969
<0.001
4.830
<0.001
3.303
<0.001
-1.026
.228
WFH% Within2
.150
.836
.274
.759
1.175
.168
-5.847
<0.001
-4.813
<0.001
-4.116
<0.001
-2.899
.003
Random Effects
σ2
.390
.410
.370
.460
.640
.800
.720
τ00
0.46
0.48
0.50
0.65
0.72
1.09
1.04
ICC
.540
.540
.570
.580
.530
.580
.590
N
994
948
950
994
994
994
994
Observations
19,040
13,961
14,071
19,042
19,042
19,042
19,042
Marginal R2 / Conditional R2
0.056 / 0.568
0.159 / 0.615
0.213 / 0.666
0.024 / 0.593
0.014 / 0.536
0.003 / 0.578
0.008 / 0.596
Note. WFH% = percentage of time working from home (WFH). Prof. = professional, Perf. = perfomance, Adapt. = adaptive, Proact =
proactive. For the sake of space, only parameter estimates and p-values are reported here; complete results are available in our online
supplemental material: https://osf.io/h6q7j/

WORKING FROM HOME
61
Table 3. Covariate Sensivity Analysis
Prof. Isolation
WFH Satisfaction
WFH Self-Efficacy
Task Perf.
Adapt. Perf.
Proact. Perf.
Job Satisfaction
Predictors
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
(Intercept)
2.095
<0.001
3.504
<0.001
3.315
<0.001
5.538
<0.001
4.413
<0.001
2.425
<0.001
5.374
<0.001
Time
-.004
<0.001
.002
.002
.000
.745
-.006
<0.001
-.009
<0.001
-.006
<0.001
.000
.812
Age
-.017
<0.001
.009
.001
.008
.001
.019
<0.001
.008
.012
-.013
.001
.014
<0.001
Gender
-.143
.005
.145
.009
.211
<0.001
.234
<0.001
.266
<0.001
.066
.420
.050
.530
Education
.043
.108
-.037
.197
-.012
.693
.001
.974
.007
.850
.012
.787
-.041
.321
Industry
-.030
.699
.035
.677
-.012
.886
.078
.412
.159
.126
.119
.341
.109
.375
Contract Hours
.008
.421
-.006
.558
-.003
.802
-.007
.536
.012
.328
.048
.001
-.012
.423
Home Childcare
.188
.012
.009
.908
-.084
.285
-.222
.015
-.094
.346
.113
.345
-.027
.819
Lockdown Period
.036
.001
-.048
<0.001
-.025
.039
-.029
.013
-.018
.192
-.023
.139
-.033
.024
WFH Pre-COVID
-.110
.067
.083
.198
.153
.018
.102
.165
.110
.171
.121
.211
.171
.071
WFH Equipment
-.238
.002
.309
<0.001
.416
<0.001
.275
.003
.041
.682
-.149
.220
.017
.890
WFH Location
.112
.030
.201
<0.001
.170
.002
.022
.726
.233
.001
.393
<0.001
.226
.005
WFH % Between
15.282
<0.001
54.605
<0.001
62.366
<0.001
2.163
.654
2.294
.664
-2.388
.707
9.411
.129
WFH % Between2
-19.884
<0.001
-8.152
.028
-12.322
.001
12.246
.003
5.165
.256
1.405
.797
11.310
.034
WFH % Within
10.218
<0.001
8.760
<0.001
8.133
<0.001
5.449
<0.001
5.103
<0.001
3.324
.001
.236
.807
WFH % Within2
.073
.929
1.474
.139
1.455
.125
-4.775
<0.001
-4.350
<0.001
-3.541
.002
-2.214
.047
Random Effects
σ2
.400
.400
.370
.470
.630
.790
.740
τ00
.400
.430
.450
.610
.730
1.060
1.010
ICC
.500
.520
.550
.570
.540
.570
.580
N
739
701
703
739
739
739
739
Observations
15,332
11,065
11,155
15,334
15,334
15,334
15,334
Marginal R2 / Conditional R2
0.116 / 0.560
0.217 / 0.623
0.282 / 0.676
0.090 / 0.605
0.045 / 0.556
0.046 / 0.591
0.044 / 0.596
Note. WFH% = percentage of time working from home (WFH). Prof. = professional, Perf. = perfomance, Adapt. = adaptive, Proact =
proactive. For the sake of space, only parameter estimates and p-values are reported here; complete results are available in our online
supplemental material: https://osf.io/h6q7j/

WORKING FROM HOME
62
Table 4. Exploratory Moderator Analysis
Prof. Isolation
WFH Satisfaction
WFH Self-Efficacy
Task Perf
Adapt. Perf.
Proact. Perf.
Job Satisfaction
Predictors
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
ɣ
p
(Intercept)
2.533
<0.001
3.190
<0.001
3.164
<0.001
5.231
<0.001
5.093
<0.001
4.559
<0.001
4.841
<0.001
Time
-.005
<0.001
.003
<0.001
.001
.496
-.006
<0.001
-.009
<0.001
-.006
<0.001
.000
.736
WFH Pre-COVID
-.143
.019
.170
.011
.205
.002
.052
.475
-.004
.957
.004
.964
.115
.208
WFH Equipment
-.326
<0.001
.374
<0.001
.502
<0.001
.490
<0.001
.169
.145
-.170
.223
.139
.299
WFH Location
.100
.044
.187
.001
.146
.007
.026
.665
.225
.001
.372
<0.001
.265
<0.001
WFH % Between
31.592
.024
42.414
.004
56.208
<0.001
-23.897
.156
-4.585
.802
10.446
.635
4.533
.831
WFH % Between2
-50.792
<0.001
3.289
.781
.097
.994
40.887
.002
24.266
.086
-.961
.955
49.680
.002
WFH % Within
11.444
<0.001
12.903
<0.001
8.610
.001
-6.717
.013
-3.379
.281
-2.678
.447
3.519
.299
WFH % Within2
1.285
.664
4.463
.187
4.658
.150
1.041
.745
-2.200
.554
1.750
.676
.780
.847
WFH Pre-COVID × WFH % Between
4.006
.644
-23.091
.016
-20.232
.034
-10.865
.296
-9.653
.393
.403
.976
-15.673
.231
WFH Pre-COVID × WFH % Between2
-16.054
.036
15.515
.063
11.348
.175
-13.174
.152
-22.750
.023
-20.985
.081
-20.568
.076
WFH Equipment × WFH % Between
-10.371
.469
21.449
.156
14.995
.330
36.141
.036
18.114
.332
-6.779
.763
17.601
.416
WFH Equipment × WFH % Between2
38.339
.001
-19.107
.124
-18.649
.141
-28.578
.037
-10.660
.474
7.465
.677
-34.196
.047
WFH Location × WFH % Between
-14.531
.036
-4.433
.555
-1.546
.838
4.060
.625
8.003
.375
3.625
.739
-1.206
.908
WFH Location × WFH % Between2
-1.474
.829
8.608
.238
8.920
.227
17.194
.036
8.118
.361
2.703
.801
15.645
.129
WFH Pre-COVID × WFH % Within
-4.044
.008
-1.662
.359
.275
.873
-.825
.618
-3.082
.109
-3.342
.123
-7.022
.001
WFH Pre-COVID × WFH % Within2
-4.357
.012
.054
.980
3.307
.100
-3.724
.049
-.825
.707
-.631
.798
6.877
.004
WFH Equipment × WFH % Within
-1.054
.684
-2.602
.377
2.134
.446
12.564
<0.001
10.696
.001
7.953
.030
.573
.871
WFH Equipment × WFH % Within2
.660
.833
-2.805
.439
-3.147
.364
-7.927
.020
-2.968
.452
-5.722
.198
-6.913
.106
WFH Location × WFH % Within
2.791
.051
-2.613
.123
-4.903
.002
1.086
.483
-.327
.856
-.711
.726
-3.748
.054
WFH Location × WFH % Within2
-2.163
.199
-2.440
.225
-3.224
.091
5.467
.003
1.666
.432
-.851
.722
-.375
.870
Random Effects
σ2
.400
.410
.370
.470
.630
.800
.740
τ00
0.43
0.45
0.47
0.63
0.73
1.08
1.00
ICC
.520
.520
.560
.570
.540
.570
.570
N
833
793
795
833
833
833
833
Observations
16,213
11,742
11,836
16,215
16,215
16,215
16,215
Marginal R2 / Conditional R2
0.085 / 0.560
0.190 / 0.615
0.253 / 0.670
0.057 / 0.599
0.031 / 0.553
0.024 / 0.585
0.039 / 0.591
Note. WFH% = percentage of time working from home. Prof. = professional, Perf. = perfomance, Adapt. = adaptive, Proact =
proactive. For the sake of space, only parameter estimates and p-values are reported here; complete results are available in our online
supplemental material: https://osf.io/h6q7j/

WORKING FROM HOME
63
Table 5. Marginal Effects for Focal Hypothesis Tests
Professional Isolation
WFH% Btwn.
ɣm
SE
CILow
CIHigh
Figure 2, Panel "A"
-40
.022
.002
.018
.027
-20
.014
.002
.011
.017
0
.005
.001
.004
.006
20
-.004
.001
-.005
-.002
40
-.012
.002
-.016
-.009
60
-.021
.003
-.027
-.016
Task Performance
WFH% Wthn.
ɣm
SE
CILow
CIHigh
Figure 2, Panel "H"
-100
.010
.001
.008
.013
-50
.006
.001
.005
.007
0
.002
.000
.001
.002
50
-.002
.001
-.004
-.001
100
-.007
.001
-.009
-.004
Adaptive Performance
WFH% Wthn.
ɣm
SE
CILow
CIHigh
Figure 2, Panel "J"
-100
.009
.001
.006
.012
-50
.005
.001
.004
.007
0
.002
.000
.001
.002
50
-.002
.001
-.003
.000
100
-.005
.001
-.008
-.003
Proactive Performance
WFH% Wthn.
ɣm
SE
CILow
CIHigh
Figure 2, Panel "L"
-100
.007
.002
.004
.010
-50
.004
.001
.003
.006
0
.001
.000
.001
.002
50
-.002
.001
-.003
.000
100
-.005
.002
-.008
-.002
Job Satisfaction
WFH% Wthn.
ɣm
SE
CILow
CIHigh
Figure 2, Panel "N"
-100
.004
.001
.001
.007
-50
.002
.001
.000
.003
0
.000
.000
-.001
.000
50
-.002
.001
-.004
-.001
100
-.005
.001
-.007
-.002
Note. WFH% = percentage of time working from home (WFH). Btwn. = between-person, Wthn.
= within-person. Labelled panels correspond to the Figure 2.

WORKING FROM HOME
64
Table 6. Marginal Effects for Between-Person Exploratory Moderator Tests
Professional Isolation
WFH % Btwn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "A"
-40
WFH Exp. = No
.0140
.0042
.0057
.0223
-20
.0099
.0025
.0050
.0148
0
.0058
.0013
.0034
.0083
20
.0017
.0020
-.0022
.0056
40
-.0024
.0037
-.0096
.0048
60
-.0065
.0055
-.0172
.0042
-40
WFH Exp. = Yes
.0273
.0059
.0158
.0388
-20
.0179
.0038
.0105
.0253
0
.0085
.0019
.0048
.0123
20
-.0009
.0017
-.0042
.0025
40
-.0102
.0035
-.0170
-.0035
60
-.0196
.0055
-.0305
-.0088
Professional Isolation
WFH % Btwn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "B"
-40
WFH Equip. = No
.0459
.0076
.0309
.0609
-20
.0292
.0045
.0204
.0379
0
.0125
.0026
.0074
.0175
20
-.0043
.0044
-.0128
.0043
40
-.0210
.0075
-.0357
-.0062
60
-.0377
.0109
-.0591
-.0162
-40
WFH Equip. = Yes
.0140
.0042
.0057
.0223
-20
.0099
.0025
.0050
.0148
0
.0058
.0013
.0034
.0083
20
.0017
.0020
-.0022
.0056
40
-.0024
.0037
-.0096
.0048
60
-.0065
.0055
-.0172
.0042
Task Performance
WFH % Btwn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "C"
-40
WFH Equip. = No
-.0366
.0092
-.0546
-.0186
-20
-.0232
.0054
-.0337
-.0127
0
-.0097
.0031
-.0158
-.0036
20
.0037
.0052
-.0065
.0140
40
.0172
.0090
-.0005
.0349
60
.0306
.0131
.0049
.0564
-40
WFH Equip. = Yes
-.0071
.0051
-.0170
.0029
-20
-.0030
.0030
-.0089
.0028
0
.0010
.0015
-.0019
.0040
20
.0051
.0024
.0004
.0098
40
.0091
.0044
.0006
.0177
60
.0132
.0065
.0004
.0260
Task Performance
WFH % Btwn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "D"
-40
WFH Loc. = Other
-.0071
.0051
-.0170
.0029
-20
-.0030
.0030
-.0089
.0028
0
.0010
.0015
-.0019
.0040
20
.0051
.0024
.0004
.0098
40
.0091
.0044
.0006
.0177
60
.0132
.0065
.0004
.0260
-40
WFH Loc. = Own Room
-.0196
.0064
-.0321
-.0071
-20
-.0099
.0039
-.0175
-.0023
0
-.0002
.0019
-.0039
.0036
20
.0095
.0026
.0044
.0147
40
.0193
.0049
.0096
.0289
60
.0290
.0075
.0142
.0437
Note. WFH% = percentage of time working from home (WFH). Exp. = experience, Equip. =
equipment, Loc. = location. Btwn. = between-person, Wthn. = within-person. Labelled panels
correspond to the Figure 3.

WORKING FROM HOME
65
Table 6 (continued). Marginal Effects for Between-Person Exploratory Moderator Tests
Job Satisfaction
WFH % Btwn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "E"
-40
WFH Equip. = No
-.0377
.0116
-.0604
-.0150
-20
-.0213
.0067
-.0344
-.0082
0
-.0050
.0039
-.0126
.0027
20
.0114
.0066
-.0016
.0244
40
.0277
.0113
.0055
.0499
60
.0441
.0165
.0118
.0763
-40
WFH Equip. = Yes
-.0075
.0064
-.0200
.0050
-20
-.0024
.0038
-.0098
.0049
0
.0027
.0019
-.0010
.0064
20
.0078
.0030
.0020
.0136
40
.0129
.0055
.0021
.0237
60
.0180
.0082
.0019
.0340
Note. WFH% = percentage of time working from home (WFH). Exp. = experience, Equip. =
equipment, Loc. = location. Btwn. = between-person, Wthn. = within-person. Labelled panels
correspond to the Figure 3.

WORKING FROM HOME
66
Table 7 Marginal Effects for Within-Person Exploratory Moderator Tests
Professional Isolation
WFH % Wthn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "F"
-100
WFH Exp. = No
.0008
.0018
-.0028
.0044
-50
.0023
.0010
.0004
.0042
0
.0037
.0004
.0030
.0044
50
.0051
.0010
.0033
.0070
100
.0066
.0018
.0030
.0101
-100
WFH Exp. = Yes
.0059
.0023
.0013
.0104
-50
.0041
.0012
.0017
.0065
0
.0023
.0005
.0013
.0033
50
.0005
.0013
-.0021
.0031
100
-.0012
.0024
-.0060
.0035
Task Performance
WFH % Wthn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "G"
-100
WFH Exp. = No
.0123
.0020
.0084
.0162
-50
.0072
.0011
.0052
.0093
0
.0022
.0004
.0014
.0030
50
-.0029
.0010
-.0049
-.0009
100
-.0080
.0020
-.0118
-.0041
-100
WFH Exp. = Yes
.0175
.0025
.0126
.0225
-50
.0097
.0013
.0072
.0123
0
.0019
.0006
.0009
.0030
50
-.0059
.0014
-.0087
-.0031
100
-.0137
.0026
-.0189
-.0085
Task Performance
WFH % Wthn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "H"
-100
WFH Equip. = No
-.0040
.0053
-.0143
.0064
-50
-.0032
.0030
-.0090
.0026
0
-.0024
.0010
-.0044
-.0005
50
-.0017
.0021
-.0057
.0024
100
-.0009
.0043
-.0093
.0076
-100
WFH Equip. = Yes
.0123
.0020
.0084
.0162
-50
.0072
.0011
.0052
.0093
0
.0022
.0004
.0014
.0030
50
-.0029
.0010
-.0049
-.0009
100
-.0080
.0020
-.0118
-.0041
Task Performance
WFH % Wthn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "I"
-100
WFH Loc. = Other
.0123
.0020
.0084
.0162
-50
.0072
.0011
.0052
.0093
0
.0022
.0004
.0014
.0030
50
-.0029
.0010
-.0049
-.0009
100
-.0080
.0020
-.0118
-.0041
-100
WFH Loc. = Own Room
.0046
.0027
-.0006
.0098
-50
.0035
.0014
.0007
.0064
0
.0025
.0005
.0015
.0035
50
.0015
.0013
-.0011
.0040
100
.0004
.0025
-.0045
.0053
Job Satisfaction
WFH % Wthn.
Moderator
ɣm
SE
CI Low
CI High
Figure 3, Panel "J"
-100
WFH Exp. = No
.0106
.0025
.0057
.0155
-50
.0061
.0013
.0034
.0087
0
.0015
.0005
.0006
.0025
50
-.0030
.0013
-.0055
-.0004
100
-.0075
.0025
-.0123
-.0027
-100
WFH Exp. = Yes
-.0022
.0032
-.0084
.0041
-50
-.0016
.0017
-.0049
.0016
0
-.0011
.0007
-.0024
.0003
50
-.0005
.0018
-.0041
.0030
100
.0000
.0033
-.0065
.0066
Note. WFH% = percentage of time working from home (WFH). Wthn. = within-person. (WFH).
Exp. = experience , Equip. = equipment, Loc. = location. Btwn. = between-person, Wthn. =
within-person. Labelled panels “F” to “J” correspond to the Figure 3.

WORKING FROM HOME
67
Figure 1. Conceptual Representation of a “Too Much of a Good Thing Effect.”
Note. Dashed lines (γm) represent marginal effects representing the curvlinear effect of
percentage of time working from home on work outcomes. Marginal effects are positive (+ γm)
when the percentage of time working from home is relatively low, zero (γm = 0) at moderate
levels, and negative when the percentage of time working from home is relatively high (− γm).

WORKING FROM HOME
68
Figure 2. Predicted Relations Between Percentage of Time Working From Home and Outcomes.
Note. Plotted points and corresponding percentages correspond to peaks or valleys of
statistically-significant nonlinear effects when the predicted marginal effect is zero.

WORKING FROM HOME
69
Figure 3. Predicted Conditional Relations Between Percentage of Time Working From Home
and Outcomes.
Note. Labelled panels “A” to “E” correspond to the marginal effects reported in Table 6; labelled
panels “F” to “J” correspond to the marginal effects reported in Table 7.

WORKING FROM HOME
70
Appendix. Definitions of Outcome Constructs
Construct
Definition
Key Citations
Professional Isolation
Feeling disconnected or detached from one's professional community.
Golden et al. (2008)
Galanti et al. (2021)
Work from Home Self-Efficacy
An individual's belief in their ability to effectively work from home.
Bandura (2006)
Raghuram et al. (2003)
Work from Home Satisfaction
An individual's level of contentment and fulfillment with their work
arrangements that allow them to work from home.
Wanous et al. (1997)
Fisher et al. (2016)
Work Performance
The extent to which an individual successfully carries out the tasks and
responsibilities required by their job through the enactment of of task-
focused, adaptive, and proactive behaviors.
Griffin et al. (2007)
Job Satisfaction
The level of contentment and fulfillment an individual experiences in
their job.
Dolbier et al. (2005)