Show that you understand the current state of research on your topic.
The length of a research proposal can vary quite a bit. A bachelor’s or master’s thesis proposal can be just a few pages, while proposals for PhD dissertations or research funding are usually much longer and more detailed. Your supervisor can help you determine the best length for your work.
One trick to get started is to think of your proposal’s structure as a shorter version of your thesis or dissertation , only without the results , conclusion and discussion sections.
Download our research proposal template
Discover proofreading & editing
Writing a research proposal can be quite challenging, but a good starting point could be to look at some examples. We’ve included a few for you below.
Like your dissertation or thesis, the proposal will usually have a title page that includes:
The first part of your proposal is the initial pitch for your project. Make sure it succinctly explains what you want to do and why.
Your introduction should:
To guide your introduction , include information about:
As you get started, it’s important to demonstrate that you’re familiar with the most important research on your topic. A strong literature review shows your reader that your project has a solid foundation in existing knowledge or theory. It also shows that you’re not simply repeating what other people have already done or said, but rather using existing research as a jumping-off point for your own.
In this section, share exactly how your project will contribute to ongoing conversations in the field by:
Following the literature review, restate your main objectives . This brings the focus back to your own project. Next, your research design or methodology section will describe your overall approach, and the practical steps you will take to answer your research questions.
? or ? , , or research design? | |
, )? ? | |
, , , )? | |
? |
To finish your proposal on a strong note, explore the potential implications of your research for your field. Emphasize again what you aim to contribute and why it matters.
For example, your results might have implications for:
Last but not least, your research proposal must include correct citations for every source you have used, compiled in a reference list . To create citations quickly and easily, you can use our free APA citation generator .
Some institutions or funders require a detailed timeline of the project, asking you to forecast what you will do at each stage and how long it may take. While not always required, be sure to check the requirements of your project.
Here’s an example schedule to help you get started. You can also download a template at the button below.
Download our research schedule template
Research phase | Objectives | Deadline |
---|---|---|
1. Background research and literature review | 20th January | |
2. Research design planning | and data analysis methods | 13th February |
3. Data collection and preparation | with selected participants and code interviews | 24th March |
4. Data analysis | of interview transcripts | 22nd April |
5. Writing | 17th June | |
6. Revision | final work | 28th July |
If you are applying for research funding, chances are you will have to include a detailed budget. This shows your estimates of how much each part of your project will cost.
Make sure to check what type of costs the funding body will agree to cover. For each item, include:
To determine your budget, think about:
If you want to know more about the research process , methodology , research bias , or statistics , make sure to check out some of our other articles with explanations and examples.
Methodology
Statistics
Research bias
Once you’ve decided on your research objectives , you need to explain them in your paper, at the end of your problem statement .
Keep your research objectives clear and concise, and use appropriate verbs to accurately convey the work that you will carry out for each one.
I will compare …
A research aim is a broad statement indicating the general purpose of your research project. It should appear in your introduction at the end of your problem statement , before your research objectives.
Research objectives are more specific than your research aim. They indicate the specific ways you’ll address the overarching aim.
A PhD, which is short for philosophiae doctor (doctor of philosophy in Latin), is the highest university degree that can be obtained. In a PhD, students spend 3–5 years writing a dissertation , which aims to make a significant, original contribution to current knowledge.
A PhD is intended to prepare students for a career as a researcher, whether that be in academia, the public sector, or the private sector.
A master’s is a 1- or 2-year graduate degree that can prepare you for a variety of careers.
All master’s involve graduate-level coursework. Some are research-intensive and intend to prepare students for further study in a PhD; these usually require their students to write a master’s thesis . Others focus on professional training for a specific career.
Critical thinking refers to the ability to evaluate information and to be aware of biases or assumptions, including your own.
Like information literacy , it involves evaluating arguments, identifying and solving problems in an objective and systematic way, and clearly communicating your ideas.
The best way to remember the difference between a research plan and a research proposal is that they have fundamentally different audiences. A research plan helps you, the researcher, organize your thoughts. On the other hand, a dissertation proposal or research proposal aims to convince others (e.g., a supervisor, a funding body, or a dissertation committee) that your research topic is relevant and worthy of being conducted.
If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.
McCombes, S. & George, T. (2023, November 21). How to Write a Research Proposal | Examples & Templates. Scribbr. Retrieved June 10, 2024, from https://www.scribbr.com/research-process/research-proposal/
Other students also liked, how to write a problem statement | guide & examples, writing strong research questions | criteria & examples, how to write a literature review | guide, examples, & templates, "i thought ai proofreading was useless but..".
I've been using Scribbr for years now and I know it's a service that won't disappoint. It does a good job spotting mistakes”
© THE INTERCEPT
ALL RIGHTS RESERVED
The proposal, rejected by U.S. military research agency DARPA, describes the insertion of human-specific cleavage sites into SARS-related bat coronaviruses.
A grant proposal written by the U.S.-based nonprofit the EcoHealth Alliance and submitted in 2018 to the Defense Advanced Research Projects Agency, or DARPA, provides evidence that the group was working — or at least planning to work — on several risky areas of research. Among the scientific tasks the group described in its proposal, which was rejected by DARPA, was the creation of full-length infectious clones of bat SARS-related coronaviruses and the insertion of a tiny part of the virus known as a “proteolytic cleavage site” into bat coronaviruses. Of particular interest was a type of cleavage site able to interact with furin, an enzyme expressed in human cells.
The EcoHealth Alliance did not respond to inquiries about the document, despite having answered previous queries from The Intercept about the group’s government-funded coronavirus research. The group’s president, Peter Daszak, acknowledged the public discussion of an unfunded EcoHealth proposal in a tweet on Saturday. He did not dispute its authenticity.
Original reporting. fearless journalism. delivered to you..
Since the genetic code of the coronavirus that caused the pandemic was first sequenced, scientists have puzzled over the “furin cleavage site.” This strange feature on the spike protein of the virus had never been seen in SARS-related betacoronaviruses, the class to which SARS-CoV-2, the coronavirus that causes the respiratory illness Covid-19, belongs.
The furin cleavage site enables the virus to more efficiently bind to and release its genetic material into a human cell and is one of the reasons that the virus is so easily transmissible and harmful. But scientists are divided over how this particular site wound up in the virus, and the cleavage site became a major focus of the heated debate over the origins of the pandemic.
Many who believe that the virus that caused the pandemic emerged from a laboratory have pointed out that it is unlikely that the particular sequence of amino acids that make up the furin cleavage site would have occurred naturally.
Adherents of the idea that SARS-CoV-2 emerged from a natural spillover from animal hosts have argued that it could have evolved naturally from an as-yet undiscovered virus. Further, they argued, scientists were unlikely to have engineered the feature.
“There is no logical reason why an engineered virus would utilize such a suboptimal furin cleavage site, which would entail such an unusual and needlessly complex feat of genetic engineering,” 23 scientists wrote earlier this month in an article in the journal Cell. “There is no evidence of prior research at the [Wuhan Institute of Virology] involving the artificial insertion of complete furin cleavage sites into coronaviruses.”
But the proposal describes the process of looking for novel furin cleavage sites in bat coronaviruses the scientists had sampled and inserting them into the spikes of SARS-related viruses in the laboratory.
“We will introduce appropriate human-specific cleavage sites and evaluate growth potential in [a type of mammalian cell commonly used in microbiology] and HAE cultures,” referring to cells found in the lining of the human airway, the proposal states.
The new proposal, which also described a plan to mass vaccinate bats in caves, does not provide conclusive evidence that the virus that caused the pandemic emerged from a lab. And virus experts remain sharply divided over its origins. But several scientists who work with coronaviruses told The Intercept that they felt that the proposal shifted the terrain of the debate.
“Some kind of threshold has been crossed,” said Alina Chan, a Boston-based scientist and co-author of the upcoming book “Viral: The Search for the Origin of Covid-19.” Chan has been vocal about the need to thoroughly investigate the possibility that SARS-CoV-2 emerged from a lab while remaining open to both possible theories of its development. For Chan, the revelation from the proposal was the description of the insertion of a novel furin cleavage site into bat coronaviruses — something people previously speculated, but had no evidence, may have happened.
“Let’s look at the big picture: A novel SARS coronavirus emerges in Wuhan with a novel cleavage site in it. We now have evidence that, in early 2018, they had pitched inserting novel cleavage sites into novel SARS-related viruses in their lab,” said Chan. “This definitely tips the scales for me. And I think it should do that for many other scientists too.”
Richard Ebright, a molecular biologist at Rutgers University who has espoused the possibility that SARS-CoV-2 may have originated in a lab, agreed. “The relevance of this is that SARS Cov-2, the pandemic virus, is the only virus in its entire genus of SARS-related coronaviruses that contains a fully functional cleavage site at the S1, S2 junction,” said Ebright, referring to the place where two subunits of the spike protein meet. “And here is a proposal from the beginning of 2018, proposing explicitly to engineer that sequence at that position in chimeric lab-generated coronaviruses.”
“A possible transmission chain is now logically consistent — which it was not before I read the proposal.”
Martin Wikelski, a director at the Max Planck Institute of Animal Behavior in Germany, whose work tracking bats and other animals was referenced in the grant application without his knowledge, also said it made him more open to the idea that the pandemic may have its roots in a lab. “The information in the proposal certainly changes my thoughts about a possible origin of SARS-CoV-2,” Wikelski told The Intercept. “In fact, a possible transmission chain is now logically consistent — which it was not before I read the proposal.”
But others insisted that the research posed little or no threat and pointed out that the proposal called for most of the genetic engineering work to be done in North Carolina rather than China. “Given that the work wasn’t funded and wasn’t proposed to take place in Wuhan anyway it’s hard to assess any bearing on the origin of SARS-CoV-2,” Stephen Goldstein, a scientist who studies the evolution of viral genes at the University of Utah, and an author of the recent Cell article, wrote in an email to The Intercept.
Other scientists contacted by The Intercept noted that there is published evidence that the Wuhan Institute of Virology was already engaged in some of the genetic engineering work described in the proposal and that viruses designed in North Carolina could easily be used in China. “The mail is filled with little envelopes with plasmid dried on to filter paper that scientists routinely send each other,” said Jack Nunberg, director of the Montana Biotechnology Center at the University of Montana.
Vincent Racaniello, a professor of microbiology and immunology at Columbia University, was adamant that the proposal did not change his opinion that the pandemic was caused by a natural spillover from animals to humans. “There are zero data to support a lab origin ‘notion,’” Racaniello wrote in an email. He said he believed that the research being proposed had the potential to fall in the category of gain-of-function research of concern, as did an experiment that was detailed in another grant proposal recently obtained by The Intercept. The government funds such research, in which scientists intentionally make viruses more pathogenic or transmissible in order to study them, only in a narrow range of circumstances . And DARPA rejected the proposal at least in part because of concerns that it involved such research.
While Racaniello acknowledged that the research in the DARPA proposal entailed some danger, he said “the benefits far, far outweigh the risk.” He also said the fact that the viruses described in the proposal were not known pathogens mitigated the concern. “This is not SARS,” he said, referring to SARS-CoV-1, the virus that caused a 2003 outbreak. “It’s SARS-related.”
But SARS-CoV-2 is not a direct descendant of that virus — it’s a relative.
In fact, the viruses described in the grant proposal, which was first posted online by the research group DRASTIC , were not known pathogens. And the authors of the grant proposal make the case that because the scientists would be using SARS-related bat viruses, as opposed to the SARS virus that was known to infect humans, the research was exempt from “gain-of-function concerns.” But according to several scientists interviewed by The Intercept, the viruses presented a threat nevertheless.
“The work describes generating full-length bat SARS-related coronaviruses that are thought to pose a risk of human spillover. And that’s the type of work that people could plausibly postulate could have led to a lab-associated origin of SARS-CoV-2,” said Jesse Bloom, a professor at Fred Hutchinson Cancer Research Center and director of the Bloom Lab, which studies the evolution of viruses. Bloom pointed out that the scientists acknowledge the risk to humans in their proposal. “It’s an explicit goal of the grant to identify the bat SARS-related coronaviruses that they think pose the highest risk.”
Stuart Newman, a professor of cell biology who directs the developmental biology laboratory at New York Medical College, also said the fact that the viruses weren’t known to be dangerous didn’t preclude the possibility that they might become so. “That’s really disingenuous,” Newman said of the argument. “The people that are claiming natural emergence say that it begins with a bat virus that evolved to be compatible with humans. If you use that logic, then this virus could be a threat because it could also make that transition.” Newman, a longtime critic of gain-of-function research and founder of the Council for Responsible Genetics, said that the proposal confirmed some of his worst fears. “This is not like slightly stepping over the line,” said Newman. “This is doing everything that people say is going to cause a pandemic if you do it.”
While the grant proposal does not provide the smoking gun that SARS-CoV-2 escaped from a lab, for some scientists it adds to the evidence that it might have. “Whether that particular study did or didn’t [lead to the pandemic], it certainly could have,” said Nunberg, of Montana Biotechnology Center. “Once you make an unnatural virus, you’re basically setting it up in an unstable evolutionary place. The virus is going to undergo a whole bunch of changes to try and cope with its imperfections. So who knows what will come of it.” The risks of such research are profound and irreversible, he said. “You can’t call back the virus once you release it into the environment.”
DARPA, a division of the Department of Defense, said regulations prevented it from confirming that it had reviewed the proposal. “Since EcoHealth Alliance may or may not be the direct source of the material in question, and we are precluded by Federal Acquisition Regulations from divulging bidders or any associated proposal details, we recommend that you reach out to them to confirm the document’s authenticity,” a DARPA spokesperson wrote in an email to The Intercept. The British Daily Telegraph reported that it had confirmed the document’s legitimacy with a former member of the Trump administration.
The Telegraph story erroneously reported that the scientists proposed to inoculate bats with live viruses. In fact, they hoped to inoculate them with chimeric S proteins, which were proposed to be developed through a subcontract in the grant in Ralph Baric’s lab at the University of North Carolina at Chapel Hill, not in Wuhan. Baric did not respond to The Intercept’s request for comment.
Many questions remain about the proposal, including whether any of the research described in it was completed. Even without the DARPA funding, there were many other potential ways to pay for the experiments. And scientists interviewed for this article agreed that often researchers do some of the science they describe in proposals before or after they submit them.
“This was a highly funded group of researchers that wouldn’t let one rejection halt their work,” said Chan, the “Viral” author.
Perhaps the most troubling question about the proposal is why, within the small group of scientists who have been searching for information that could shed light on the origins of the pandemic, there has apparently been so little awareness of the planned work until now. Peter Daszak and Linfa Wang, two of the researchers who submitted the proposal, did not previously acknowledge it.
Daszak, the EcoHealth Alliance president, has actively sought to quash interest in the idea that the novel coronavirus originated in a lab. In February 2020, as the pandemic began to grip major cities in the U.S., he began organizing scientists to write an open letter that was published in the Lancet addressing the origins of the virus. “The rapid, open, and transparent sharing of data on this outbreak is now being threatened by rumours and misinformation around its origins,” read the statement signed by Daszak and 26 co-authors. “We stand together to strongly condemn conspiracy theories suggesting that COVID-19 does not have a natural origin.”
Daszak directed and gathered signatures for the letter, all the while suggesting that he and his collaborators on the proposed DARPA project, Baric and Wang, distance themselves from the effort.
“I spoke with Linfa [Wang] last night about the statement we sent round. He thinks, and I agree with him, that you, me and him should not sign this statement, so it has some distance from us and therefore doesn’t work in a counterproductive way,” Daszak wrote to Baric in February 2020, just weeks before it appeared in the journal, according to an email surfaced a year later by public health investigative research group U.S. Right to Know. “We’ll then put it out in a way that doesn’t link it back to our collaboration so we maximize an independent voice.” Ultimately, Daszak did sign the letter.
“I also think this is a good decision,” Baric replied. “Otherwise it looks self-serving and we lose impact.”
Baric and Wang — a professor in the emerging infectious diseases program at Duke-NUS Medical School, Singapore — did not respond to inquiries from The Intercept about their decision not to sign the letter in the Lancet.
Daszak was also a member of the joint team the World Health Organization sent to China in February 2020 to investigate the origins of the pandemic, which concluded that it was “extremely unlikely” that the virus had been released from a laboratory. (In March, WHO called for further investigation of the origins of the virus and stated that “all hypotheses remain open.”)
“I find it really disappointing that one of the members of the joint WHO-China team, which is essentially the group of scientists that were tasked as representatives of both the scientific community and the World Health Organization of investigating this, are actually on this proposal, knew that this line of research was at least under consideration, and didn’t mention it all,” said Bloom, of Fred Hutch. “Whatever information that relates to help people think about this just needs to be made transparently available and explained.”
Correction: September 24, 2021
A previous version of this article stated incorrectly that the EcoHealth Alliance proposal had been featured on Sky News Australia.
Correction: September 23, 2021, 3:30 p.m.
A previous version of this article stated incorrectly that Linfa Wang was a member of the WHO-China team.
Additional credits:.
Additional Reporting: Mara Hvistendahl
Latest stories.
Israel’s War on Gaza
Shrouq Aila
I was on the road to the camp when Israel launched its raid — and saw the aftermath in a nearby hospital.
Deconstructed Podcast
Deconstructed
“One side or the other is going to win,” Alito told a person he thought was a right-wing activist.
Shawn Musgrave
The federal judge hearing a human rights case disputed allegations he might be impartial but recused himself out of an “abundance of caution.”
An art and a science: 5 tips to mastering freelance proposal contracts.
Business Signing a Contract Buy - sell house.
Unless you’re a nepo baby, if you want to be a freelancer, you gotta write proposals. Proposals get freelancers jobs, period. Fortunately, I’ve got some tips for you on how to write the perfect one. This is for everyone (even the nepo babies) who want to make a career as an independent contractor.
The first step in writing an effective proposal is understanding your client's needs. This involves meticulous research to grasp not only the project requirements but also the client's goals, challenges, and company culture. Successful freelancers know that a proposal is more than a bid; it's a solution. By presenting your services as the answer to their problems, you position yourself as not just a contractor, but a partner.
A well-structured proposal can be the difference between winning and losing a contract. In fact, it’s kinda...what the contract hangs on. So, be sure to include an intro, a description of the problem, a proposed solution (you should be able to solve it for the client, that’s...also the point), key milestones, a timeline, and the price.
The tone of your proposal should reflect both your personal brand and the client's culture. But to be on the safe side, err more towards mirroring the client. As my mother always says, never put too much of yourself into it. If their brand is formal and professional, mirror that in your writing. If they're more laid-back or creative, a more conversational tone could be fine. But put in the work to get to know them.
Platforms like Bonsai, Proposify, Trello, Asana, and FreshBooks allow freelancers to organize themselves, as well as to create and track proposals. They even come with analytics, which can be helpful (or hurtful, depending on how well your business is going). I don’t need data to tell you that automating something makes it faster, though, so give these tools a try.
Samsung just confirmed galaxy ring release date, feature list and u.s. sales, wwe raw results winners and grades as bron breakker destroys ricochet, ask for feedback.
After submitting a proposal, follow up within a week if you haven’t heard back. This feels annoying (because it is), but annoying people get contracts. If your proposal is rejected, see if they’ll tell you why. They might not, but if they do, it’s a chance to improve your proposal-writing skill for next time. The art of proposal writing is not static; continuously seek out resources to improve your skills—either from potential clients, other freelancers, or professional organizations. You got this.
One Community. Many Voices. Create a free account to share your thoughts.
Our community is about connecting people through open and thoughtful conversations. We want our readers to share their views and exchange ideas and facts in a safe space.
In order to do so, please follow the posting rules in our site's Terms of Service. We've summarized some of those key rules below. Simply put, keep it civil.
Your post will be rejected if we notice that it seems to contain:
User accounts will be blocked if we notice or believe that users are engaged in:
So, how can you be a power user?
Thanks for reading our community guidelines. Please read the full list of posting rules found in our site's Terms of Service.
Purdue Online Writing Lab Purdue OWL® College of Liberal Arts
This page is brought to you by the OWL at Purdue University. When printing this page, you must include the entire legal notice.
Copyright ©1995-2018 by The Writing Lab & The OWL at Purdue and Purdue University. All rights reserved. This material may not be published, reproduced, broadcast, rewritten, or redistributed without permission. Use of this site constitutes acceptance of our terms and conditions of fair use.
The Online Writing Lab (the Purdue OWL) at Purdue University houses writing resources and instructional material, and we provide these as a free service at Purdue. Students, members of the community, and users worldwide will find information to assist with many writing projects. Teachers and trainers may use this material for in-class and out-of-class instruction.
The On-Campus and Online versions of Purdue OWL assist clients in their development as writers—no matter what their skill level—with on-campus consultations, online participation, and community engagement. The Purdue OWL serves the Purdue West Lafayette and Indianapolis campuses and coordinates with local literacy initiatives. The Purdue OWL offers global support through online reference materials and services.
Facebook twitter.
Search for:
Candid learning offers information and resources that are specifically designed to meet the needs of grantseekers., build the skills you need to do good..
Sign up for live training from experts
Engage diverse donors with culturally informed marketing strategies, introduction to finding grants.
On-demand training you can take in your own time
Tell your story: reach donors using your free candid profile, crafting compelling grant proposals, sign up for our newsletter, proposal writing.
Build the essential knowledge and skills to craft effective grant proposals.
Fine-tune prospecting skills and learn how to divers ify your sources of funding to thrive.
Get the skills to strategize, build infrastructure , and scale for greater impact .
Your nonprofit’s success depends on the growth of the individuals working for it–including you. Start maximizing your human resources.
Get the details on starting a nonprofit and how to determine if you’re ready .
Earn a Seal of Transparency and unlock benefits that can help you reach your fundraising goals.
Strengthen your grantmaking practices by tapping the practical wisdom of a diverse group of your experienced funding peers.
View all topics
Professional development for large groups.
Give your teams access to the training they need to craft grant winning proposals and fundraise effectively. Our experts can put together a training package to suit your organization’s goals.
An official website of the United States government
The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.
The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.
Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .
Jingjing qu.
1 Shanghai AI Lab, China
2 School of Business and Administration, Northeastern University, Shenyang China
Despite being a worldwide disaster, the COVID‐19 pandemic has also provided an opportunity for renewed discussion about the way we work. By contextualizing in the early periods of China's ending of lockdown policy on COVID‐19, this paper offers evidence to respond to an essential discussion in the field of working from home (WFH): In terms of job performance, can WFH replace working from the office (WFO)? The present study compares job performance in terms of quality and productivity between WFH and WFO from 861 Chinese respondents using entropy balance matching, a quasi‐experimental methodology. Results reveal that WFH enhances job performance in terms of job quality but lowers it in terms of job productivity. In addition, the present study aims to capture and empirically measure the variations in fundamental job characteristics in terms of job control and job demand between WFH and WFO by applying the job demand control support model. More specifically, we find that job control items, such as ‘talking right’ and ‘work rate’, and job demand items, such as ‘a long time of intense concentration’ and ‘hecticness of the job’, are vital factors that contribute to how these differences exert influence on employees' performance in the context of the pandemic.
The outbreak of the COVID‐19 pandemic boosted an unprecedentedly massive and rapid shift of people's work routines (Bartram and Cooke 2022 ; Yan et al. 2021 ). To a large extent, millions of employees around the world have been forced to resort to remote work (Bouziri et al. 2020 ; Hurley and Popescu 2021 ; Rogers 2021 ; Woods and Miklencicova 2021 ), which leads to the most significant social experiment of ‘working from home (WFH)’ emerging in human history (Zhang, Yu and Marin 2021 ). According to a report in LinkedIn, as Asia‐Pacific responded to the crisis, organizations in China, Australia, India and Singapore, quickly adapted to support a remote workforce. WFH differs considerably from working from office (WFO) in terms of job attributes and work environment. WFO is characterized by a relatively high degree of formalization and a fixed working routine, including place, time, and task arrangements (Palumbo 2020 ). Information and communications technology (ICT) was widely adopted with regard to work and organizational management (Balica 2019 ; Kassick 2019 ; Nemțeanu, Dabija and Stanca 2021 ; Olsen 2019 ). WFH is characterized by the freedom from constraints associated with working in a formal and fixed workplace due to progress in ICT (Nakrošienė, Bučiūnienė and Goštautaitė 2019 ).
Long before the COVID‐19 pandemic, WFH had already been suggested as a modern human resource policy for organizations, and it has resulted in a definite trend firmly entrenched in society (Illegems, Verbeke and S’Jegers 2001 ; Stanek and Mokhtarian 1998 ). It enables employees to be more productive by avoiding long commutes, skirting office politics, having fewer office distractions, and giving more chance to develop a better work–life balance (Hopkins and McKay 2019 ; Nakrošienė, Bučiūnienė and Goštautaitė 2019 ). Simultaneously, a stream of scholars have argued that WFH is not an alternative working routine and may even lead to poor employee performance (Fonner and Roloff 2010 ). Thus, a key question in the field has been raised: Can WFH replace WFO? Around this question, the debate has become fierce alongside the development of ICT and globalization. Nevertheless, past research has not yet reached a consensus, which constitutes a significant gap in the current knowledge.
Thus, drawing on the above research gaps, the present research is designed as a comparison study contextualized in the ongoing COVID‐19 pandemic. On the basis of the job demand–control–support (JDCS) model, a well‐documented theory that elucidates the effects of fundamental job characteristics (Johnson and Hall 1988 ), and combined with entropy balance matching (Watson and Elliot 2016 ), the present study investigates the difference between WFH group and other working cohorts in terms of job characteristics and its effects on job performance. More specifically, based on the JDCS model, we propose the mediation effect of job demand and job control and the moderation effect of employers' anti‐epidemic policy as the social support on the relationship between job demand/job control and employee job performance.
The contributions of this study are as follows. First, we shed new light on the mixed effects of WFH on job performance. We find that WFH can increase job quality but reduce job productivity. Second, underpinned by the JDCS framework, the present paper empirically tests the differences of job characteristics between WFH and other working routines regarding job demand, job control and social supports, and its direct and indirect effects on employees' satisfaction on performance. In this case, the present paper extends the JDCS model from the field of classical work routine to understand WFH. Furthermore, we employ the entropy balancing method to alleviate the methodological concerns with selection bias in the previous literature. Doing so allows for examining the causal effect of WFH on job characteristics and job performance to support the random hypothesis in comparison quasi‐experiment research.
The remainder of this paper is organized as follows. The next section presents the literature review, followed by a discussion of the hypothesis development. Further sections present the methods and results, respectively. The final section presents a discussion and implications, followed by future scope and conclusion.
WFH is a working arrangement in which employees fulfill the essential responsibilities that their job entails while remaining at home using ICT (International Labor Organization 2020 , 5). Although a slight difference exists among terms such as WFH, teleworking, telecommuting and remote working, these concepts are largely interchangeable. WFH is considered home‐based teleworking, because teleworking may include various locations away from the primary worksite or the employers' premises (such as mobile working). Telecommuting refers to substituting telecommunications for commuter travel. Some differences exist between the terms teleworking and telecommuting, mainly because teleworking is broader and may not always be a substitute for commuting, but they are relatively minor. The basic difference between telework and remote work is that a teleworker uses personal electronic devices in addition to working physically remotely from a place other than an office or company premises, whereas remote work does not require visits to the main workplace or the use of electronic personal devices; and compared with WFH, remote work has the flexibility to work anywhere rather than being limited to the home. In addition, WFH may imply a long‐term contract, and individuals may have an emotional relationship with the organization; however, in remote work, this is not easy to achieve (Tønnessena, Dhira and Flåten 2021 ).
This paper aims to illustrate whether WFH can replace the classical working routine. A comparison study between WFH and other working routines seems to be a promising way to solve this question. However, we should consider two significant challenges of conducting a comparison study on WFH and other working routines. First, a ubiquitous theoretical framework is critical for providing solid support to capture fundamental job characteristics of diverse working routines. Only by doing so can we compare the difference between WFH and the other cohorts at the datum line. Second, we need to conquer the self‐selection bias. Most employees considering the possibility of WFH as the alternative way are familiar with applying ICT applications (e.g. email and online meeting apps) and necessary equipment (e.g. laptop and smartphone). In addition, employees' meta‐cognitive knowledge – their understanding of their capacity to cope with various situations under WFH ways (e.g. interruption caused by children and communication with line manager) – may play a similar self‐selective role. On the basis of these self‐selective factors, individuals evaluate the advantages and disadvantages of WFH and make decisions (Williams, McDonald and Cathcart 2017 ). Not controlling for this nonrandom self‐selection implies that observed job performance may reflect individuals' superior knowledge, capacity, or equipment rather than the actual effect of WFH. However, it is difficult to isolate the effects of job characteristics of WFH and the influence of individual heterogeneity explicitly associated with WFH. Thus, this paper adopts the JDCS model to investigate the effect of WFH on employees' job performance.
In the last 20 years, inconsistent findings have been found on the effect of WFH on employees' performance, especially in terms of work efficiency, turnover intention, goal completion, work motivation and job satisfaction (Gajendran and Harrison 2007 ; Golden 2006 ). On the one hand, some studies have found that WFH leads to high job performance (Bloom et al. 2015 ; Campo, Avolio and Carlier 2021 ; Choukir et al. 2022 ; Ipsen et al. 2021 ; Liu, Wan and Fan 2021 ). On the other hand, studies have found that WFH may lead to employees' lack of supervision, miscommunication, and less organizational commitment (Madell 2021 ). These disadvantages can create uncertainty that affects job satisfaction and consequently lead to lowering performance among employees, as gauged by companies' key performance indicators (Pepitone 2013 ). Some scholars have argued that WFH is negatively related to employees' job performance (Mustajab et al. 2020 ; Van Der Lippe and Lippényi 2020 ). Raišienė et al. ( 2020 ) suggested an investigation of the influence of WFH on job performance based on a contingency view, which depends on employees' gender, age, education, work experience, and telework experience. Table 1 summarizes the related literature.
Summary of related literature
Author | Objective | Methodology | Results/Findings | Association between WFH and performance |
---|---|---|---|---|
Bloom et al. ( ) | To investigate whether WFH works | Experiment | WFH led to a 13% performance increase | Positive |
Choukir et al. ( ) | To investigate the effects of WFH on job performance | Survey, SEM | WFH positively affects employees’ job performance | Positive |
Liu, Wan and Fan ( ) | To investigate the relationship between WFH and job performance | Survey, regression | WFH can improve job performance through job crafting | Positive |
Ipsen et al. ( ) | To investigate people’s experiences of WFH during the pandemic and to identify the main factors of advantages and disadvantages of WFH | Survey, descriptive statistics, exploratory factor analyses, ‐test, ANOVA | WFH can improve work efficiency | Positive |
Campo, Avolio and Carlier ( ) | To investigate the relationship among telework, job performance, work–life balance and family supportive supervisor behavior in the context of COVID‐19 | Survey, partial least squares structural equation modelling (PLS‐SEM) | WFH is positively correlated with job performance | Positive |
Van Der Lippe and Lippényi ( ) | To investigate the influence of co‐workers WFH on individual and team performance | Survey, SEM | WFH negatively impacted employee performance. Moreover, team performance is worse when more co‐workers are working from home | Negative |
Mustajab et al. ( ) | To investigate the impacts of working from home on employee productivity | Survey, qualitative method with an exploratory approach | WFH is responsible for the decline in employee productivity | Negative |
Raišienė et al. ( ) | To investigate the efficiency of WFH | Survey, correlation analysis | There are differences in the evaluation of factors affecting work efficiency and qualities required from a remote worker, depending on gender, age, education, work experience, and experience of telework | Contingency |
Which one is better influence of wfh on job performance.
The JDCS model provides a sound theoretical basis for the influence of WFH on job performance. It originated from the job demand–control (JDC) model, which explains how job characteristics alter employees' stress, performance and satisfaction (Karasek and Theorell 1990 ). The JDC model posits two fundamental characteristics of an occupation: job demand and job control. Job demand is defined initially as ‘physical consumptions and psychological tensions involved in accomplishing the workload’, which negatively relate to workplace well‐being and relevant performance (Karasek and Theorell 1990 , 291). Job control (originally decision latitude) is the extent to which an employee has the authority to decide and utilize skills concerning the job and exert a positive effect on workplace outcomes. The JDCS model compounds the prominence of environmental factors on the overall well‐being within the workplace (Baka 2020 ). Thus, social support was integrated into the JDC model (named JDCS model) as a further fundamental characteristic of the work environment, implicating its synergistic effect on reducing stress and promoting well‐being in the working environment (Johnson and Hall 1988 ).
Given the inconsistent findings on the relationship between WFH and job performance, we further investigate the effect of WFH on job performance based on the JDCS model. The COVID‐19 pandemic has made WFH a sudden reality, as the ILO defined WFH in the context of the COVID‐19 pandemic as a temporary and alternative home‐based teleworking arrangement (ILO 2020 ). Waizenegger et al. ( 2020 ) articulated the differences between remote e‐working before and during the COVID‐19 pandemic.
Given the two mechanisms of JDCS, we further investigate the effect of WFC on job performance separately from the perspective of job demand and job control. On the one hand, WFH may lead to high job control, which benefits job performance, because not all job functions and tasks can be done outside the employers' premises or the specified workplace (Waizenegger et al. 2020 ). WFH is not practical or feasible or cannot be deployed quickly in some jobs and tasks (Williams, McDonald and Cathcart 2017 ). Accordingly, employees can arrange their time and energy with adequate job autonomy when they are WFH. They can deal with tasks under the best working status and promote work productivity and quality. On the other hand, WFH may lead to high job demand, which decreases job performance. Job demands are typically operationalized in terms of quantitative aspects, such as workload and time pressure (Hopkins and McKay 2019 ; Karasek and Theorell 1990 ). The boundary between working and leisure times becomes ambiguous when employees are WFH. Employees are usually pushed to work for longer hours and face high job demand, which is harmful to work productivity and quality. Therefore, assessing the influence of WFH on employees' feeling of their work completion is vaguer and more complicated compared with WFO, which leads us to propose our first hypothesis as a set of two alternatives:
Employees who are WFH are more satisfied with their job performance (i.e. job quality and job productivity).
Employees who are WFH are less satisfied with their job performance (i.e. job quality and job productivity).
On the basis of the JDCS model (Karasek and Theorell 1990 ), we tend to examine the differences of job fundamental characteristics and the moderating effect of social support on job performance between WFH and other working routines. WFH may increase job demand due to its possibility of pushing individuals to work for longer hours and increase the intensity of individuals. It will lead to a high investment of personal resources and bring adverse effects afterward.
First, WFH acquires more personal energy and time to invest in dealing with ‘communication via technology’, and employees may need to learn and equip with knowledge accordingly, including terms of using WFH tools and methods of collaboration (Yang et al. 2021 ). Moreover, employees may face the risks of technology fatigue or crash, which may result in negative psychological effects of misinformation and putting off work accomplishments (Khan 2021 ). Second, when employees need to continue to work beyond the regular working hours, they will inevitably face continuous additional work pressure, which makes them unable to relax and recover physically and mentally. Accordingly, more personal time and resources are demanded to invest in the job (Xie et al. 2018 ). Ayyagari, Grover and Purvis ( 2011 ) believed that WFH forms in such a convenient manner where employees may be required to stay on call for quarantine for a long time. WFH may influence employees' everyday life and lead to a perception of higher expectations for their working hours and intensity by their company and work loading. Ter Hoeven, van Zoonen and Fonner ( 2016 ) also verified this and reported that WFH may cost extra job demands from employees, including financial assets, energy, time and psychological capital. If those demands are too high, they may further make a series of workplace deviation behaviors, such as time‐encroached behaviors, to alleviate their loss of personal resources (Vayre 2021 ), consequently reducing their job performance.
The relationship between WFH and job performance is mediated by job demand.
We further reason that the relationship between WFH and job performance is mediated by job control. The most prominent advantage of WFH is regarded as flexibly anytime and anywhere, which can significantly enhance employees' sense of job control and autonomy (Richardson and Thompson 2012 ). Mazmanian, Orlikowski and Yates ( 2013 ) found that employees who complete work tasks through WFH would have increased perceived work control and work flexibility. WFH can also enhance job autonomy in respect of task arrangement, work manner and task order (Mazmanian, Orlikowski and Yates 2013 ). Studies have also verified that WFH will promote employees' benefits in the field of the family via a more flexible and adaptable arrangement (Dockery and Bawa 2018 ). As a result, it can balance their work and family duties concerning the different daily situations and perform well (Tønnessena, Dhira and Flåten 2021 ).
The relationship between WFH and job performance is mediated by job control.
Social support is characterized by helpful relations with supervisors and coworkers (Mayo et al. 2012 ). Previous evidence has argued that a lack of support from employers when applying WFH may lead to a series of problems and thus reduce job performance (Palumbo 2020 ). According to the JDCS model, social support often buffers the effects of job demands and job control on the work‐related outcomes of employees (Johnson and Hall 1988 ). We investigate the moderation effect of social support on the relationship between job demand/control and job performance.
First, WFH may lead to isolation among employees if they have fewer interactions with their coworkers, supervisors and managers. Second, employees may not get recognition and support when needed, which may lead to employees' dissatisfaction, as their social needs cannot be fulfilled by WFH (Marshall, Michaels and Mulki 2007 ). Another negative consequence is receiving less recognition for achievements because exhibiting their work achievements is more difficult when all communication is conducted electronically (Zhang 2016). The limitation exists because when employees are WFH, they usually submit their work when it is ready. However, their manager may not see the process involved in producing a deliverable; some employees may work overtime, but their work is only judged by the result, not by the difficulties they overcome. Thus, policies or strategies should be implemented to enhance employers' feeling of embeddedness, not only for the sake of job performance but also for their well‐being and sustainability of human resourcing of organizations.
Particularly, considering the context of the epidemic, support actions from employers aiming to be anti‐epidemic and protect employees will be essential to improve the positive consequences of WFH. Thus, the present paper takes employers' anti‐epidemic policy as prominent social support worthy of examining. Indeed, some Chinese companies coined proactive guidance and support for employees (Reeves et al. 2020 ). The support reportedly helped employees feel less stressed, experience more positive feelings toward their leader and their team, and created an atmosphere of trust and understanding that motivated them to apply themselves more fully to work (Xu and Thomas 2011 ). In this case, we suggest that a moderating effect of the employers' anti‐epidemic policy is significantly observed on the influence of WFH on job performance. Figure 1 shows the conceptual framework.
Social support moderates the relationship between job demand and job performance, such that the relationship is weaker when social support is high rather than low.
Social support moderates the relationship between job control and job performance, such that the relationship is stronger when social support is high rather than low.
The conceptual framework
Our sample was collected from China. It is the first region where the government applied a lockdown policy, which encouraged employers to organize their employees to WFH to mitigate the massive health crisis. Nevertheless, in March 2020, due to the sound control of COVID‐19 spread, after only a few months' lockdowns, Chinese citizens were able to return gradually to their normal work–life routine. As a result, some employees were WFH, and some of them returned to their normal work routine. Different from the previous research conducted in a limited number of industries or focusing on a particular occupation group, such a situation provides us a unique opportunity to design comparison research to understand early, initial reactions of a wide range of occupational groups and industries toward WFH and its social effect in the epidemic context.
Data were collected via an online survey, provided by a Chinese survey company called Wenjuanxing ( www.wjx.cn ), a platform providing functions equivalent to Amazon Mechanical Turk. Research on WFH confronts a widely noted difficulty in managing data face‐to‐face, especially during this particular epidemic term. Thus, we chose to issue and collect the questionnaire online.
We initially did a pilot survey on 1 March 2020, with 100 observations. Later, after adjustments to the questionnaire, we issued the formal study of 5 March 2020, a month after the earliest date for work resumption according to the Chinese government. Thus, some employees were returning to workplace (RTW), and some continued WFH after Chinese New Year. As mentioned before, this particular time allows us to do a comparison study that covers various types of occupation and organization to seek the differences between WFH and RTW when society is confronted with a significant public health emergency. After collecting data for two weeks, we gathered 1342 observations.
Furthermore, to alleviate the self‐selective bias caused by participants passively excluded from WFH due to lacking necessary conditions, we took the inclusion criteria that required the participants to be equipped with requirements of WFH, such as essential online tools and Internet access. We identified the qualified group by asking, ‘Do you think you have the qualified conditions to be working from home (e.g., possesses Internet access, laptop, smart phone, software, and apps)?’ Then, we selected those who answered yes. After cleansing invalid data, the final sample consisted of 861 individuals, among which 442 claimed that they were WFH, and 419 were RTW.
Our sample comprised participants who were portrayed as young and received a high‐level of education, who were aged around 31–35 on average. The participants were 44% male. The majority of the participants were qualified with undergraduate degree. Particularly, 9.98% of the participants were married without children, 58.65% were married with children, 30.89% were single without children, and 0.4% were single with children. Around half of the participants (50.41%) worked for private enterprises, 16.7% worked for state‐owned enterprises, 15.21% worked for foreign companies, and others worked in government or public institutions. The participants at management positions accounted for 41%. Those who had marketing duties accounted for 31%. Others had positions in R&D. The participants worked for 9.36 days on average after the Chinese New Year (also the deadline of the epidemic blockade), and 71% of them had experience of training or education while WFH. The participants were from 16 places in China, the largest portions were from Guangdong Province (13.43%), Shanghai (7.66%), Shandong (6.15%), and Jiangsu (6.15%).
Job performance was measured by two items adopted from a structured measurement coined by Viswesvaran, Ones and Schmidt’s ( 1996 ) measurement of job performance (overall job performance, productivity, and quality). We applied the two dimensions of job performance, namely, ‘productivity’ and ‘quality’, which were examined by self‐evaluation questions: 1) In terms of productivity, how do you evaluate the quantity or volume of work produced today (e.g. number of transactions completed)? 2) In terms of quality, how do you feel about how well the job was done today (You can consider several aspects of the quality of tasks completed, including lack of errors, accuracy to specifications, thoroughness, and amount of wastage)? The answers were measured using a Likert scale, from 1 (poor) to 5 (excellent). As a key self‐evaluation measurement of job performance, Viswesvaran, Ones and Schmidt’s ( 1996 ) instrument has been widely applied by following scholars in the fields of organizational behavior, psychology, and human resource management (Judge et al. 2001 ; Lee, Berry and Gonzalez‐Mulé 2019 ; Murphy 2020 ).
WFH was used here to identify the work status of respondents, with 1 representing WFH, and 0 representing WFO.
Job demand and job control were measured following Gonzalez‐Mulé and Cockburn ( 2017 ) work, which is a well‐documented instrument widely applied in research and referred to as the JDC model.
Job demand was measured by eight questions (e.g. ‘To what extent do you agree that your job requires working very hard?’ ‘To what extent do you agree that your job requires working very fast?’). The answer was measured using a Likert scale, from 1 (completely disagree) to 5 (completely agree; Cronbach's alpha = 0.83).
Job control was measured by seven questions (e.g. ‘To what extent do you agree that your job allows you to make a lot of decisions on your own?’ ‘To what extent do you agree that you have a lot to say about what happens on your job?’). The answer was measured using a Likert scale, from 1 (completely disagree) to 5 (completely agree; Cronbach's alpha = 0.75).
Social support was measured by employees' satisfaction on employers' anti‐epidemic policy. The survey question was, ‘Overall, are you satisfied with your employers’ anti‐epidemic support (e.g. financial support, emotional support from line managers, anti‐epidemic knowledge guides, and clear guidelines of WFH)?’ The answer was a dummy one, 1 representing yes, and 0 indicating no.
First, we controlled for effective communication as a key factor that affects the quality of job performance, given that the majority of the literature has argued that ineffective communication is one of the greatest challenges of interpersonal collaborations mediated by ICTs in WHF (Wang et al. 2021 ). We controlled a set of communication factors in terms of ‘accurately delivered job content’ and ‘fully expressed the information’, among others. The answers were designed as a Likert scale, from 1 (completely disagree) to 5 (completely agree).
Furthermore, consistent with earlier studies, we controlled for difference of working hours, namely, the difference between daily working hours and today’s working hours, working experiences, normal daily working hours, daily number of colleagues they worked with, daily number of leaders they worked with, daily number of departments they worked with, daily commuting time, positions, age, gender, education, marital status, nature of employers, position levels, and days of starting work after the Chinese New Year. The definitions of variables are provided in Table A1 .
Definition of variables
Variables | Definition | Cronbach alpha |
---|---|---|
Condition qualified with WFH | Is measured by following question: ‘Do you think you own the qualified conditions to working from home? (e.g. able to access internet, have laptop, smart phone, necessary software and apps)’ Answer: Dummy, 1: yes; 0: no | n.a. |
Job performance – quality | Is measured by following question: ‘How do you feel about how well the job was done today? (You can consider several aspects of the quality of tasks completed including lack of errors, accuracy to specifications, thoroughness, and amount of wastage).’ Answer: A Likert Scale, 1 poor to 5 excellent | n.a. |
Job performance – productivity | Is measured by following question: ‘How do you evaluate the quantity or volume of work produced today? (e.g. number of transactions completed, extent of daily task completed)’ Answer: A Likert Scale, 1 poor to 5 excellent | n.a. |
WFH | Is measured by following question: ‘Do you work from home or return to workplace now?’ Answer: Dummy, 1: working from home; 0: working at workplace | |
Job control | Is measured by following 6 items: Con1: to what extent do you agree that your job allows you to make a lot of decisions on your own? Con2: to what extent do you agree that you have a lot of say about what happens on your job? Con3: to what extent do you agree that you can determine the order in which your work is to be done on your job? Con4: to what extent do you agree that you can determine when a task is to be done on your job? Con5: to what extent do you agree that you can determine your own work rate on your job? Con6: to what extent do you agree that you have very little freedom to decide how you do your work on the job? Answer: A Likert Scale, 1 completely disagree to 5 completely agree | .75 |
Job demand | Is measured by following 9 items: Dem1: to what extent do you agree that your job requires working very hard? Dem2: to what extent do you agree that your job requires working very fast? Dem3: to what extent do you agree that your job requires long periods of intense concentration? Dem4: to what extent do you agree that your job is very hectic? Dem5: to what extent do you agree that you have too much work to do everything well on your job? Dem6: to what extent do you agree that you are not asked to do an excessive amount of work at your job? (reverse scored) Dem7: to what extent do you agree that you have enough time to get the job done? (reverse scored) Dem8: to what extent do you agree that that you are free of conflicting demands that others make on your job? (reverse scored) Dem9: How frequently does your job require working under time pressure? Answer: A Likert Scale, 1 completely disagree to 5 completely agree | .77 |
Social support | Is measured by following question: ‘Overall, are you satisfied with your employer’s anti‐epidemic support? (e.g. financial support, emotional support from line managers, anti‐epidemic knowledge guides, clear guidelines of WFH)’ Answer: Dummy, 1: yes; 0: no | n.a. |
Effective communication | Is measured by following questions: Com1: to what extent do you agree that the inter‐personal communication related to your job can accurately delivery job content? Com2: to what extent do you agree that the inter‐personal communication related to your job fully express the information? Com3: to what extent do you agree that you are well acknowledged the process of the team project? Com4: to what extent do you agree that the inter‐personal communicating message is delivered in a positive way? Com5: to what extent do you agree that the inter‐personal communicating message is delivered in a negative way? Com6: recently, communication conflicts have quite often had a negative impact on completing my daily work. Com7: I feel the relationships with my colleagues are not as close asthey used to be. Answer: A Likert Scale, 1 completely disagree to 5 completely agree | .83 |
Daily working hours | Is measured by following question: ‘recently, how many hours have you needed to work daily?’ Answer: Numbers | n.a. |
Difference of working hours | Is calculated by: Daily working hours – Daily hours used to work Daily hours used to work is measured by following question: ‘how many hours did you need to work daily before lockdown?’ Answer: Numbers | n.a. |
Working experiences | Is measured by following question: ‘How many years since you got your first job’ Answer: years | n.a. |
Daily number of colleagues work with | Is measured by following question: ‘On average, how many colleagues do you need to communicate with on daily base?’ Answer: Numbers | n.a. |
Daily number of leaders work with | Is measured by following question: ‘On average, how many leaders do you need to report to on a daily basis?’ Answer: Numbers | n.a. |
Daily number of departments work with | Is measured by following question: ‘On average, how many departments do you need to communicate with on a daily basis?’ Answer: Numbers | n.a. |
Daily commuting time | Is measured by following question: ‘On average, how many hours did you spend commuting to the workplace?’ Answer: Numbers | n.a. |
Positions | Is measured by following question: ‘What is your position?’ Answer: 1: Management position, 2: R&D position, 3: Rear‐Service positions, 4: Marketing position,5:Other | n.a. |
Position levels | Is measured by following question: ‘What’s the level of your position?’ Answer: 1: rank‐and‐file employee, 2: middle manager 3: top manager | n.a. |
Nature of employers | Is measured by following question: ‘What’s the nature of your employer?’ Answer: 1: government 2: public institutions, 3: foreign‐funded enterprise and joint venture, 4: state‐owned enterprise; 5: private enterprise | n.a. |
Age | Answer: 1: under 25, 2: 25–30, 3: 31–35, 4: 36–40, 5: 41–50, 6: over 50 | n.a. |
Gender | Answer: 1: male, 0:female | n.a. |
Education | Answer: 1: no degree to 5: postgraduate degree and above | n.a. |
Marriage & Children | Answer: 1: married, no child, 2: married, have a child or children, 3: single, no child, 4: single, have a child or children | n.a. |
Days of starting work after Chinese New Year | Is measured by following question: ‘How many days since you started to work after Chinese New Year?’ Answer: Numbers | n.a. |
WFH Training | Is measured by following question: ‘Do you ever have training experience working from home? (e.g., remote work apps, training on communications via online tools),’ Answer: Dummy, 1: yes; 0: no | n.a. |
Our analysis consists of three steps. In Step 1, to test our hypothesis 1, we applied entropy balance and weighted mean difference Welch's t ‐test (mean after entropy balance matching) methods to compare the self‐evaluated job performance between WFH and WFO employees. Following the approach of recent papers on labor economics and health (Hetschko, Schöb and Wolf 2016 ; Kunze and Suppa 2017 ; Nikolova, 2019 ), our strategy includes 1) data preprocessing to form comparable groups of individuals as treatment and control group (treatment group: WFH employees; control group: RTW employees) by applying entropy balance, and 2) estimating the treatment effect after matching by Welch's t ‐test. We also reconfirmed the regression result (Hainmueller 2012 ).
In Step 2, we investigated the direct and mediating effects of job control and job demand on job performance (hypotheses 2 and 3). We applied the quasi‐Bayesian Monte Carlo method to test the mediating effect of job demand and job control, which is a technique to increase the robustness of the mediating test by employing a strategy of numerous repeated re‐sampling to build an empirical approximation of the sampling distribution and examine the indirect effects by constructing the confidence intervals (CIs; Imai, Keele and Tingley 2010 ). We used the package ‘Mediation’ for causal mediation analysis. In addition, to confirm the validity and reliability of mediating hypotheses results, we used structural equation modeling (SEM) as robustness check, with package ‘lavaan’ to assess the mediating effect of job control and job demand on the relationship between WFH and job performance.
In Step 3, to test the moderating effect of social support, we applied hierarchical regressions at the final step by following the classical approaches to seek the significance of interactions in a set of model tests.
All the analysis is conducted with software R.
Before testing the hypotheses, a benchmark test of a binary correlation matrix is presented in Table 2 . The overall coefficient is not high, and a variance inflation factor was performed at below 10, demonstrating low multicollinearity.
Variables correlation matrix
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1. | Job performance – quality | |||||||||||||||||||||||||||||||||
2. | Job performance – productivity | .41 | ||||||||||||||||||||||||||||||||
3. | WFH | .29 | −.12 | |||||||||||||||||||||||||||||||
4. | Job control | .18 | .24 | .06 | ||||||||||||||||||||||||||||||
5. | Job demand | .13 | .25 | −.12 | .18 | |||||||||||||||||||||||||||||
6. | Social support | .16 | .27 | .00 | .23 | .16 | ||||||||||||||||||||||||||||
7. | Effective communication | −.11 | −.17 | .06 | −.11 | −.05 | −.08 | |||||||||||||||||||||||||||
8. | Daily working hours | .03 | .04 | −.04 | −.15 | −.19 | −.14 | −.10 | ||||||||||||||||||||||||||
9. | Difference of working hours | −.03 | −.06 | −.04 | −.04 | .04 | −.05 | −.02 | .00 | |||||||||||||||||||||||||
10. | Working experiences | .00 | .13 | −.14 | .12 | .03 | .11 | −.15 | −.02 | .01 | ||||||||||||||||||||||||
11. | Daily number of colleagues work with | .01 | .13 | −.13 | .02 | .07 | .03 | −.11 | .06 | −.02 | .17 | |||||||||||||||||||||||
12. | Daily number of leaders work with | .01 | .11 | .00 | .04 | .11 | −.02 | −.06 | .12 | −.08 | .10 | .51 | ||||||||||||||||||||||
13. | Daily number of departments work with | −.06 | .06 | −.06 | .06 | .04 | .01 | −.02 | .11 | −.07 | .07 | .36 | .44 | |||||||||||||||||||||
14. | Daily commuting time | .03 | .04 | .06 | −.03 | −.06 | −.01 | −.02 | .08 | −.04 | .05 | .10 | .14 | .11 | ||||||||||||||||||||
15. | Management | .03 | .06 | .03 | .10 | .04 | .00 | .04 | .04 | .01 | .04 | .12 | .13 | .25 | .03 | |||||||||||||||||||
16. | Research | .04 | .05 | −.03 | .05 | .08 | .02 | −.07 | −.02 | −.05 | −.01 | .02 | .08 | −.03 | .02 | −.25 | ||||||||||||||||||
17. | Service | −.08 | −.01 | −.12 | −.01 | .05 | −.04 | .07 | −.01 | .05 | .01 | −.04 | −.06 | .01 | .03 | −.17 | −.09 | |||||||||||||||||
18. | Marketing | −.01 | −.05 | .01 | −.08 | −.02 | −.01 | −.01 | −.02 | .02 | .02 | .01 | −.01 | −.06 | −.03 | −.32 | −.25 | −.13 | ||||||||||||||||
19. | Other | −.03 | −.05 | .07 | −.04 | −.07 | .01 | .02 | −.01 | .01 | −.05 | −.08 | −.09 | −.05 | −.02 | −.26 | −.15 | −.07 | −.17 | |||||||||||||||
20. | Position levels | .02 | .08 | −.04 | .12 | .10 | .06 | .07 | −.02 | −.04 | .27 | .21 | .22 | .25 | .02 | .31 | .08 | −.13 | −.06 | −.15 | ||||||||||||||
21. | Government | .02 | .02 | −.04 | .01 | .01 | .01 | .05 | −.01 | −.01 | −.12 | −.09 | −.02 | −.04 | −.04 | .03 | .01 | .05 | −.09 | .06 | .01 | |||||||||||||
22. | Public institutions | .04 | .02 | .08 | .03 | .06 | .05 | .00 | −.01 | .01 | −.03 | −.02 | .05 | .07 | −.02 | .01 | .03 | .05 | −.04 | .08 | .01 | −.06 | ||||||||||||
23. | Foreign‐funded enterprise and joint venture | .05 | .02 | −.01 | .04 | .00 | .03 | −.01 | −.01 | −.06 | .03 | .11 | .09 | .09 | .06 | .10 | .08 | −.03 | −.06 | −.06 | .12 | −.07 | −.15 | |||||||||||
24. | State‐owned enterprise | −.06 | .01 | −.03 | .00 | .04 | .06 | −.03 | −.02 | .01 | .09 | .02 | .04 | .05 | .05 | .01 | .01 | .03 | .00 | −.04 | −.09 | −.07 | −.16 | −.19 | ||||||||||
25. | Private enterprise | −.03 | −.04 | .00 | −.07 | −.08 | −.10 | .01 | .03 | .04 | −.04 | −.06 | −.11 | −.13 | −.07 | −.09 | −.07 | −.04 | .08 | −.02 | −.03 | −.16 | −.37 | −.43 | −.45 | |||||||||
26. | Age | .01 | .11 | −.06 | .09 | .06 | .09 | −.14 | −.02 | −.02 | .72 | .10 | .12 | .06 | .04 | .05 | −.03 | −.01 | .02 | −.01 | .29 | −.05 | .10 | −.02 | .10 | −.11 | ||||||||
27. | Gender | .02 | .03 | −.04 | .01 | .10 | .09 | −.03 | −.03 | .03 | .08 | .01 | .02 | −.03 | .00 | −.04 | .18 | −.12 | .04 | .01 | .14 | .03 | .01 | −.04 | .03 | .01 | .10 | |||||||
28. | Education | .06 | .02 | −.02 | .09 | −.03 | −.04 | −.04 | −.05 | −.06 | −.05 | .09 | .13 | .10 | .01 | .12 | .24 | −.17 | −.14 | −.10 | .17 | .03 | .09 | .09 | −.02 | −.13 | −.09 | −.02 | ||||||
29. | Married, no child | .05 | −.03 | .05 | −.06 | −.13 | −.03 | −.04 | .10 | .00 | −.06 | −.05 | −.04 | −.09 | .07 | −.04 | .03 | .00 | −.03 | .09 | −.06 | .12 | −.03 | .05 | −.09 | .02 | −.08 | .03 | .08 | |||||
30. | Married, have a child or children | −.05 | .06 | −.13 | .12 | .13 | .12 | −.07 | −.04 | −.03 | .57 | .13 | .09 | .13 | −.01 | .11 | .01 | .02 | −.02 | −.14 | .31 | −.10 | .07 | −.01 | .12 | −.09 | .52 | .03 | −.03 | −.40 | ||||
31. | Single, no child | .02 | −.04 | .11 | −.08 | −.05 | −.11 | .10 | −.02 | .03 | −.57 | −.11 | −.07 | −.08 | −.04 | −.09 | −.02 | −.03 | .03 | .09 | −.30 | .02 | −.05 | −.02 | −.07 | .09 | −.51 | −.04 | −.02 | −.22 | −.08 | |||
32. | Single, have a child or children | −.03 | −.04 | .00 | −.05 | −.01 | .01 | .01 | .00 | .00 | .05 | .00 | .03 | −.01 | .01 | −.02 | −.03 | .02 | .07 | −.03 | .07 | −.01 | −.02 | .02 | .02 | .00 | .06 | −.03 | −.02 | −.02 | −.08 | −.05 | ||
33. | Days of starting work after Chinese New Year | −.01 | .04 | −.11 | .08 | .06 | −.01 | −.01 | −.02 | .03 | .09 | .03 | .05 | .01 | .02 | .00 | .07 | .01 | −.03 | −.03 | .04 | .06 | −.05 | .07 | .00 | −.04 | .03 | .05 | .09 | .03 | .01 | −.03 | −.01 | |
34. | WFH training | .08 | .07 | .08 | .06 | .14 | .13 | −.02 | .02 | .03 | −.02 | .05 | .11 | .12 | .02 | .10 | .08 | −.02 | .00 | −.09 | .16 | .03 | .03 | .07 | −.02 | −.10 | −.01 | .09 | .01 | −.04 | .09 | −.08 | .05 | −.11 |
Before proceeding to test hypothesis 1 in Step 1, we first applied the entropy balance and weighted mean difference (mean after entropy balance matching) methods. The quality of entropy balance matching combined with a data description is summarized in Table 4 . Before matching, WFH employees worked for <2.7 h daily on average compared with their pre‐daily working hours. Employees who had returned to work worked <0.53 h on average than their current daily work. After matching, this difference was reduced. WFH employees are used to having less colleagues to work with (mean: WFH = 2.91, RTW = 3.18), are less likely to work at back office (mean: WFH = 0.10, RTW = 0.18), are younger (mean: WFH = 2.62, RTW = 2.77), are less likely to be married and have a child or children (mean: WFH = 0.52, RTW = 0.56), and are more likely to be single and without a child or children (mean: WFH = 0.36, RTW = 0.26). In addition, WFH employees indicated that they started working after Chinese New Year a day later than WFO employees (mean: WFH = 9.36, RTW = 11.09). In particular, WFH employees experienced better interpersonal communication than RTW employees (mean: WFH = 2.74, RTW = 2.67). In entropy balance matching, we matched all conditioning variables, and the bias of each matched variables was reduced to nearly 0, supporting good quality of entropy balance matching. Moreover, differences in mean and variance between the treatment and control groups were largely reduced after weighting (see in Table A2 ).
Causal mediation analysis of job control and job demand
via Job Control | via Con2 | via Con5 | via Job Demand | via Dem3 | via Dem4 | |
---|---|---|---|---|---|---|
Regression on job performance – quality | ||||||
Mediating effect | .14*** | .02^ | .12*** | −.02** | −.03** | −.03* |
Direct effect | .45*** | .46*** | .45*** | .50*** | .50*** | .49*** |
Total effect | .59*** | .48*** | .57*** | .48*** | .48*** | .48*** |
Prop. mediated | 23.72%** | 4.16%^ | 21.05%** | 4.33%* | 5.3%* | 6%^ |
Regression on job performance – productivity | ||||||
Mediating effect | .03* | . 01* | .05*** | −.03*** | −.03** | −.001 |
Direct effect | −.19*** | −.17*** | −.21*** | −.12*** | −.13** | −.16*** |
Total effect | −.17*** | −.17** | −.17*** | −.15*** | −.16*** | −.16*** |
Prop. mediated | 16.4%* | 5.88%^ | 29.41%*** | 21.25% | 20.11%** | 4.9% |
^ p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001.
Descriptive statistics before treatment, selected covariate variables, before and after matching
Treated | Controls unmatched | Controls matched | Standardized bias % | |||||
---|---|---|---|---|---|---|---|---|
= 442 | = 419 | = 419 | ||||||
Mean | Variance | Mean | Variance | Mean | Variance | Unmatched | Matched | |
Effective communication | 2.74 | .37 | 2.67 | .30 | 2.74 | .33 | .12 | .00 |
Difference of working hours | −2.73 | 117.00 | −.53 | 72.15 | −2.73 | 97.22 | .29 | .00 |
Daily working hours | 3.43 | 1.40 | 3.77 | 1.30 | 3.43 | 1.40 | .09 | .00 |
Working experiences | 3.56 | 1.31 | 3.55 | 1.43 | 3.56 | 1.55 | .01 | .00 |
Daily number of colleagues work with | 2.91 | 1.08 | 3.18 | 1.04 | 2.91 | .85 | .27 | .00 |
Daily number of leaders work with | 2.14 | .64 | 2.15 | .59 | 2.14 | .56 | .01 | .00 |
Daily number of departments work with | 2.34 | .61 | 2.44 | .63 | 2.34 | .55 | .13 | .00 |
Daily commuting time | 2.18 | .67 | 2.09 | .58 | 2.18 | .64 | .12 | .00 |
Management | .41 | .24 | .38 | .24 | .40 | .24 | .12 | .00 |
Research | .20 | .16 | .22 | .17 | .20 | .16 | .06 | .00 |
Service | .10 | .09 | .18 | .15 | .10 | .09 | .05 | .00 |
Marketing | .31 | .22 | .31 | .21 | .31 | .22 | .23 | .00 |
Other | 1.40 | .35 | 1.44 | .32 | 1.40 | .34 | .01 | .00 |
Position levels | .15 | .13 | .10 | .09 | .15 | .13 | .07 | .00 |
Government | .02 | .02 | .03 | .03 | .02 | .02 | .08 | .00 |
Public institutions | .16 | .15 | .10 | .12 | .16 | .15 | .17 | .00 |
Foreign‐funded enterprise and joint venture | .15 | .13 | .16 | .13 | .15 | .13 | .03 | .00 |
State‐owned enterprise | .16 | .13 | .18 | .15 | .16 | .13 | .06 | .00 |
Private enterprise | .50 | .25 | .51 | .25 | .50 | .25 | .01 | .00 |
Age (under 25) | .44 | .25 | .39 | .24 | .43 | .25 | .15 | .00 |
Age (25–30) | .25 | .19 | .33 | .22 | .27 | .20 | .10 | .00 |
Age (31–35) | .10 | .09 | .11 | .10 | .10 | .10 | .18 | .00 |
Age (36–40) | .07 | .07 | .07 | .06 | .07 | .07 | .06 | .00 |
Age (41–45) | .02 | .02 | .02 | .02 | .02 | .02 | .01 | .00 |
Age (over 45) | .02 | .14 | .02 | .14 | .02 | .14 | .01 | .00 |
Gender (male) | .41 | .24 | .46 | .25 | .41 | .24 | .09 | .00 |
Education (no degree) | .05 | .05 | .02 | .02 | .05 | .04 | .14 | .00 |
Education (primary school) | .15 | .13 | .15 | .13 | .15 | .13 | .02 | .00 |
Education (high school) | .69 | .21 | .73 | .20 | .71 | .21 | .08 | .00 |
Education (undergraduate) | .11 | .32 | .10 | .30 | .11 | .32 | .04 | .00 |
Education (postgraduate degree and above) | .001 | .05 | 0 | 0 | – | – | – | – |
Married, no child | .11 | .10 | .09 | .08 | .11 | .10 | .10 | .00 |
Married, have a child or children | .52 | .25 | .65 | .23 | .53 | .25 | .26 | .00 |
Single, no child | .36 | .23 | .26 | .19 | .36 | .23 | .22 | .00 |
Single, have a child or children | .00 | .00 | .00 | .00 | .00 | .00 | .00 | .00 |
Days of starting work after Chinese New Year | 9.36 | 69.62 | 11.09 | 59.76 | 9.36 | 51.00 | .22 | .00 |
WFH training | .71 | .21 | .64 | .23 | .71 | .21 | .15 | .00 |
Then, we verified hypothesis 1 by measuring the ATT under the balanced matching conditions in Table 3 . After matching, the results for hypothesis 1 are presented in Tables 5 and and6. 6 . The results show that WFH employees are more satisfied with quality (mean: WFH = 4.56, RTW = 4.11, p < 0.01). In addition, WFH employees feel less satisfied with productivity (mean: WFH = 3.86, RTW = 4.05, p < 0.01). Hypotheses 1a and 1b were supported.
Treatment effect of WFH before and after entropy balance matching
Treated group | Controls unmatched | Treatment effect (unmatched) | Controls matched | Treatment effect (matched) | |||
---|---|---|---|---|---|---|---|
Mean | Mean | Mean difference | ‐Test | Mean | Mean difference | ‐Test | |
Job performance – quality | 4.56 | 4.11 | .45 | 8.92*** | 4.11 | .45 | 8.83*** |
Job performance – productivity | 3.86 | 4.05 | −.19 | −3.41*** | 4.03 | −.17 | −3.1** |
Job control | 3.67 | 3.59 | .08 | 1.81* | 3.58 | .09 | 2.10* |
Con1 | 3.62 | 3.567 | .05 | .68 | 3.59 | .03 | .38 |
Con2 | 3.24 | 3.01 | .03 | 2.04* | 3.06 | .18 | 2.36* |
Con3 | 3.76 | 3.84 | −.08 | −1.12 | 3.83 | −.07 | −1.09 |
Con4 | 3.61 | 3.67 | −.06 | −.78 | 3.71 | −.01 | .21 |
Con5 | 3.69 | 3.17 | .52 | 6.85*** | 3.16 | .53 | 7.14*** |
Con6 | 3.60 | 3.60 | 0 | −.02 | 3.53 | .07 | .028 |
Job demand | 3.37 | 3.48 | −.11 | −3.50*** | 3.46 | −.09 | −3.07*** |
Dem1 | 2.64 | 2.63 | .01 | .16 | 2.64 | 0 | −.12 |
Dem2 | 3.14 | 3.29 | −.15 | −2.27* | 3.25 | −.11 | −1.67 |
Dem3 | 3.57 | 3.75 | −.18 | −2.77*** | 3.73 | −.16 | −2.47* |
Dem4 | 3.25 | 3.50 | −.25 | −3.53*** | 3.48 | −.23 | −3.26** |
Dem5 | 3.10 | 3.18 | −.08 | −1.21 | 3.15 | −.05 | −.65 |
Dem6 | 3.72 | 3.87 | −.15 | −1.84 | 3.87 | −.15 | −1.84 |
Dem7 | 3.69 | 3.72 | −.03 | −.25 | 3.77 | −.08 | −.96 |
Dem8 | 3.85 | 3.87 | −.02 | −.25 | 3.81 | .04 | .59 |
Dem9 | 4.12 | 4.19 | −.07 | .24 | 4.17 | −.05 | −.88 |
Social support | 4.17 | 4.17 | .00 | −.08 | 4.17 | .00 | −.14 |
* p < 0.05; ** p < 0.01; *** p < 0.001.
Regressions on satisfaction with job performance (quality)
M1 | M2 | M3 | M4 | M5 | M6 | M7 | |
---|---|---|---|---|---|---|---|
WFH | .49 (.05)*** | .48 (.05)*** | .53 (.05)*** | .50 (.05)*** | .48 (.05)*** | .53 (.05)*** | .52 (.05)*** |
Mediators | |||||||
Job control | .20 (.04)*** | .23 (.21) | .27 (.22) | ||||
Job demand | .33 (.06)*** | .32 (.31) | .25 (.32) | ||||
Social support | .16 (.03)*** | .19 (.18) | .15 (.25) | .21 (.27) | |||
Interactions | |||||||
Job control * Social support | −.02 (.05) | .01 (.07) | |||||
Job demand * Social support | .00 (.07) | −.03 (.05) | |||||
Conditioning variables | |||||||
Effective communication | −.16 (.05)*** | −.14 (.05)** | −.15 (.04)*** | −.14 (.04)** | −.13 (.04)** | −.14 (.04)** | −.13 (.04)** |
Daily working hours | .00 (.00) | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)** |
Difference of working hours | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) |
Working experiences | .04 (.04) | .03 (.03) | .05 (.03) | .03 (.03) | .03 (.03) | .04 (.03) | .04 (.03) |
Daily number of colleagues work with | .03 (.03) | .04 (.03) | .04 (.03) | .03 (.03) | .04 (.03) | .04 (.03) | .04 (.03) |
Daily number of leaders work with | −.02 (.04) | −.02 (.04) | −.04 (.04) | −.01 (.04) | −.02 (.04) | −.03 (.04) | −.03 (.04) |
Daily number of departments work with | −.07 (.04). | −.08 (.04)* | −.07 (.04) | −.08 (.04)* | −.08 (.04)* | −.07 (.04) | −.08 (.04)* |
Daily commuting time | .00 (.03) | .01 (.03) | .01 (.03) | .00 (.03) | .00 (.03) | .01 (.03) | .01 (.03) |
Management | .09 (.08) | .08 (.08) | .07 (.08) | .10 (.08) | .09 (.08) | .09 (.08) | .08 (.08) |
Research | .07 (.09) | .07 (.09) | .05 (.09) | .08 (.09) | .08 (.09) | .06 (.09) | .06 (.09) |
Service | −.02 (.10) | −.02 (.10) | −.04 (.10) | .01 (.10) | .00 (.10) | −.02 (.10) | −.02 (.10) |
Marketing | .02 (.08) | .03 (.08) | .01 (.08) | .03 (.08) | .04 (.08) | .03 (.08) | .04 (.08) |
Position levels | −.14 (.10) | −.13 (.10) | −.13 (.10) | −.14 (.10) | −.13 (.10) | −.14 (.10) | −.13 (.10) |
Government | .00 (.05) | −.02 (.05) | −.02 (.05) | −.01 (.05) | −.02 (.05) | −.02 (.05) | −.03 (.05) |
Public institutions | .15 (.25) | .15 (.25) | .13 (.25) | .12 (.25) | .13 (.25) | .11 (.25) | .12 (.25) |
Foreign‐funded enterprise and joint venture | −.06 (.16) | −.04 (.16) | −.08 (.16) | −.09 (.16) | −.07 (.16) | −.10 (.15) | −.08 (.15) |
State‐owned enterprise | .01 (.17) | .04 (.17) | .02 (.17) | .00 (.17) | .02 (.17) | .01 (.17) | .03 (.17) |
Private enterprise | −.15 (.17) | −.11 (.17) | −.14 (.17) | −.17 (.17) | −.14 (.17) | −.17 (.17) | −.14 (.17) |
Age | −.07 (.16) | −.03 (.16) | −.06 (.16) | −.07 (.16) | −.04 (.16) | −.06 (.16) | −.03 (.16) |
Gender | .03 (.03) | .03 (.03) | .03 (.03) | .03 (.03) | .03 (.03) | .03 (.03) | .03 (.03) |
Education | .01 (.05) | .01 (.05) | .00 (.05) | .00 (.05) | .01 (.05) | −.01 (.05) | .00 (.05) |
Married, no child | .04 (.05) | .03 (.05) | .06 (.05) | .05 (.05) | .04 (.05) | .07 (.04) | .06 (.04) |
Married, have a child or children | −.11 (.09) | −.12 (.09) | −.17 (.09) | −.11 (.09) | −.12 (.09) | −.17 (.09) | −.17 (.09) |
Single, no child | −.07 (.09) | −.09 (.09) | −.11 (.09) | −.05 (.09) | −.07 (.09) | −.09 (.09) | −.10 (.09) |
Single, have a child or children | −.50 (.39) | −.43 (.39) | −.54 (.38) | −.48 (.39) | −.43 (.38) | −.52 (.38) | −.48 (.38) |
Days of starting work after Chinese New Year | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) |
WFH training | .07 (.06) | .06 (.06) | .03 (.06) | .04 (.06) | .04 (.06) | .01 (.06) | .00 (.06) |
square | .15 | .17 | .17 | .17 | .18 | .19 | .20 |
Adjust square | .12 | .14 | .14 | .14 | .15 | .16 | .17 |
‐value | 5.00 | 5.67 | 6.01 | 5.74 | 5.84 | 6.28 | 6.21 |
‐Value | .00 | .00 | .00 | .00 | .00 | .00 | .00 |
^ p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001; Standard errors in parentheses.
Regressions on satisfaction with job performance (productivity)
M1 | M2 | M3 | M4 | M5 | M6 | M7 | |
---|---|---|---|---|---|---|---|
WFH | −.16 (.05)** | −.18 (.05)*** | −.11 (.05)* | −.15 (.05)** | −.17 (.05)*** | −.11 (.05)* | −.13 (.05)** |
Mediators | |||||||
Job control | .30 (.05)*** | .22 (.21) | .32 (.22) | ||||
Job demand | .46 (.06)*** | .11 (.30) | .05 (.31) | ||||
Social support | .27 (.03)*** | .22 (.18) | .00 (.24) | .05 (.26) | |||
Interactions | |||||||
Job control * Social support | .00 (.05) | .03 (.05) | |||||
Job demand * Social support | −.07 (.07)* | −.08 (.07)* | |||||
Conditioning variables | |||||||
Effective communication | −.21 (.05)*** | −.18 (.05)*** | −.21 (.05)*** | −.19 (.05)*** | −.17 (.04)*** | −.19 (.04)*** | −.17 (.04)*** |
Daily working hours | .00 (.00) | .00 (.00) | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)** |
Difference of working hours | .00 (.00)* | .00 (.00)* | .00 (.00)** | .00 (.00) | .00 (.00) | .00 (.00)* | .00 (.00)* |
Working experiences | .06 (.04) | .05 (.04) | .07 (.04)* | .05 (.03) | .04 (.03) | .06 (.03) | .05 (.03) |
Daily number of colleagues work with | .05 (.03) | .05 (.03) | .06 (.03) | .04 (.03) | .05 (.03) | .05 (.03) | .06 (.03) |
Daily number of leaders work with | .02 (.04) | .02 (.04) | .00 (.04) | .04 (.04) | .03 (.04) | .01 (.04) | .01 (.04) |
Daily number of departments work with | −.04 (.04) | −.05 (.04) | −.03 (.04) | −.05 (.04) | −.06 (.04) | −.05 (.04) | −.05 (.04) |
Daily commuting time | .05 (.03) | .06 (.03) | .06 (.03) | .04 (.03) | .05 (.03) | .06 (.03) | .06 (.03)* |
Management | .13 (.08) | .11 (.08) | .10 (.08) | .15 (.08) | .13 (.08) | .12 (.08) | .12 (.08) |
Research | .07 (.09) | .07 (.09) | .03 (.09) | .09 (.09) | .09 (.09) | .05 (.09) | .05 (.09) |
Service | −.04 (.10) | −.04 (.10) | −.07 (.10) | .01 (.10) | .00 (.10) | −.03 (.10) | −.04 (.09) |
Marketing | −.09 (.08) | −.08 (.08) | −.10 (.08) | −.06 (.08) | −.06 (.08) | −.07 (.08) | −.06 (.08) |
Position levels | −.12 (.10) | −.10 (.10) | −.12 (.10) | −.13 (.10) | −.11 (.10) | −.12 (.10) | −.11 (.10) |
Government | .08 (.06) | .05 (.05) | .05 (.05) | .06 (.05) | .04 (.05) | .04 (.05) | .02 (.05) |
Public institutions | .10 (.26) | .09 (.25) | .08 (.25) | .05 (.25) | .05 (.25) | .03 (.25) | .04 (.24) |
Foreign‐funded enterprise and joint venture | .06 (.16) | .09 (.16) | .03 (.16) | .01 (.16) | .04 (.16) | .00 (.15) | .03 (.15) |
State‐owned enterprise | −.08 (.18) | −.04 (.17) | −.06 (.17) | −.11 (.17) | −.07 (.17) | −.08 (.17) | −.05 (.17) |
Private enterprise | −.02 (.18) | .03 (.17) | −.02 (.17) | −.07 (.17) | −.03 (.17) | −.06 (.17) | −.02 (.17) |
Age | −.03 (.17) | .02 (.16) | −.01 (.16) | −.03 (.16) | .01 (.16) | −.01 (.16) | .02 (.16) |
Gender | .06 (.04) | .06 (.04) | .06 (.03) | .07 (.03) | .07 (.03) | .07 (.03)* | .07 (.03)* |
Education | −.07 (.06) | −.05 (.05) | −.08 (.05) | −.08 (.05) | −.07 (.05) | −.08 (.05) | −.07 (.05) |
Married, no child | −.03 (.05) | −.04 (.05) | .00 (.05) | .00 (.05) | −.01 (.04) | .02 (.04) | .01 (.04) |
Married, have a child or children | −.05 (.09) | −.08 (.09) | −.15 (.09) | −.05 (.09) | −.07 (.09) | −.14 (.09) | −.14 (.09) |
Single, no child | .16 (.09) | .13 (.09) | .10 (.09) | .18 (.09) | .15 (.09) | .13 (.09) | .11 (.09) |
Single, have a child or children | −.48 (.40) | −.37 (.39) | −.54 (.39) | −.44 (.39) | −.36 (.38) | −.49 (.38) | −.43 (.38) |
Days of starting work after Chinese New Year | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) |
WFH training | .13 (.06)* | .11 (.06) | .07 (.06) | .07 (.06) | .07 (.06) | .02 (.06) | .02 (.06) |
square | .11 | .15 | .16 | .17 | .19 | .22 | .23 |
Adjust square | .08 | .12 | .13 | .14 | .16 | .18 | .2 |
‐value | 3.56 | 5.1 | 5.5 | 5.78 | 6.41 | 7.22 | 7.48 |
‐Value | .00 | .00 | .00 | .00 | .00 | .00 | .00 |
* p < 0.05; ** p < 0.01; *** p < 0.001; Standard errors in parentheses.
Changes in job demand and job control can be observed from Step 1 in Tables 5 and and6. 6 . Under balanced matching conditions, WFH employees experience a significantly higher level of job control (ATT: WFH = 3.67, RTW = 3.58, p < 0.05). More specifically, such change is noteworthy in the job control of ‘talking right’ (con2; ATT: WFH = 3.24, RTW = 3.06, p < 0.05) and job control of ‘working rate’ (con5; ATT: WFH = 3.69, RTW = 3.16, p < 0.001). In terms of job demand, WFH employees experience a significantly lower level than RTW employees (mean: WFH = 3.37, RTW = 3.46, p < 0.001). The difference is obviously observed in terms of ‘long periods of intense concentration’ (dem3; ATT: WFH = 3.57, RTW = 3.73, p < 0.05) and ‘hecticness of the job’ (dem4; ATT: WFH = 3.25, RTW = 3.48, p < 0.01). These results imply that WFH may lead to changes in job control and job demand, which may intermediately affect job performance.
Therefore, in the second step, we tested the mediating effect by applying the quasi‐Bayesian Monte Carlo method in Table 4 . The results show that in terms of quality, the mediating effect of job control and job demand is confirmed as statistically significant (job control = 0.14, p < 0.001; job demand = −0.02, p < 0.01). The proportion of mediating effect on total effect is around 23.72% and 4.33%. We also tested the mediating effect of the important items of job control and job demand. We find that the job control on ‘working rate’ (con5; 0.12, p < 0.10, prop. mediated = 21.05%), job demand on ‘long periods of intense concentration’ (dem3; −0.03, p < 0.01, prop. mediated = 5.3%), and ‘hecticness of the job’ (dem4; −0.03, p < 0.05, prop. mediated = 6%) positively mediate the relationship between WFH and satisfaction with quality.
In terms of productivity performance, the mediating effect of job control and job demand is supported (job control = 0.03, p < 0.05, prop. mediated = 16.4.5%; job demand = −0.03, p < 0.01, prop. mediated = 21.25%). However, it is noticeable, unlike in the domain of quality, that the mediating effect of job control and job demand contributes to the direct impact of WFH. Such mediating effect trades off the direct influence of WFH on satisfaction with productivity. Items such as job control on ‘working rate’ (con5; 0.01, p < 0.05, prop. mediated = 5.88%) and job demand on ‘long periods of intense concentration’ (dem3; −0.03, p < 0.01, prop. mediated = 20.11%) mediate the relationship between WFH and satisfaction with productivity. In this case, hypotheses 3 and 4 are fully supported.
In addition, the robustness check results via SEM analysis (both classical and bootstrap approach is used) is consistent with the quasi‐Bayesian Monte Carlo analysis. Accordingly, hypotheses 3 and 4 are supported as well (see details in Tables A3 and andA4 A4 ).
Robustness check of mediation effect by structure equation modelling
Description | χ | GFI | NNFI | CFI | RMSEA | SRMR | ||
---|---|---|---|---|---|---|---|---|
Accept values | >.90 | >.90 | >.95 | <.05 | <.08 | |||
M1 | Full items model | 1592.36 | 467 | .795 | .992 | .994 | .053 | .049 |
M2 | Dropped items model | 394.16 | 194 | .915 | .998 | .999 | .035 | .024 |
M3 | Dropped items model (bootstrap) | 394.16 | 194 | .915 | .998 | .999 | .035 | .024 |
M4 | Mean | 15.456 | 1 | .991 | .991 | 1 | .13 | .006 |
M5 | Mean (bootstrap) | 15.456 | 1 | .991 | .991 | 1 | .13 | .006 |
M1 | M2 | M3 | M4 | M5 | ||||||
---|---|---|---|---|---|---|---|---|---|---|
Quality | Productivity | Quality | Productivity | Quality | Productivity | Quality | Productivity | Quality | Productivity | |
Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | Path coefficient | |
WEF | .44 (.05)*** | −.19 (.05)*** | .45 (.05)*** | −.18 (.06)** | .45 (.05)*** | −.18 (.06)** | .47 (.05)*** | −.14 (.05)** | .47 (.05)*** | −.14 (.06)** |
Mediator | ||||||||||
Job control | .24 (.05)*** | .38 (.06)*** | .20 (.05)*** | .34 (.06)*** | .20 (.06)*** | .34 (.07)*** | .17 (.04)*** | .26 (.04)*** | .17 (.04)*** | .26 (.05)*** |
Job demand | .07 (.07) | .11 (.08) | .21 (.07)*** | .27 (.07)*** | .21 (.08)** | .27 (.08)*** | .29 (.06)*** | .37 (.06)*** | .29 (.06)*** | .37 (.07)*** |
Mediation effect | ||||||||||
Via job control | .03 (.01)* | .04 (.02) | .04 (.01)* | .06 (.02)** | .04 (.02)* | .06 (.03)* | .02 (.01)* | .30 (.01)* | .02 (.01)* | .02 (.01)* |
Via job demand | −.01 (.01) | −.01 (.01) | −.03 (.01)* | −.03 (.01)** | −.03 (.01)* | −.03 (.02)* | −.03 (.01)** | −.04 (.01)** | −.03 (.01)** | −.03 (.011)** |
Control variables | ||||||||||
Effective communication | −.12 (.04)** | −.15 (.05)*** | −.12 (.04)*** | −.16 (.05)** | −.12 (.05)** | −.16 (.05)*** | −.11 (.04)** | −.15 (.05)*** | −.11 (.04)* | −.15 (.04)** |
Daily working hours | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00)* | .00 (.00) | .00 (.00) | .00 (.00)*** | .00 (.00)** | .00 (.00) | .00 (.00) |
Working experiences | .03 (.03) | .06 (.04) | .05 (.03) | .08 (.04)* | .05 (.04) | .08 (.04)* | .04 (.03) | .07 (.03) | .04 (.04) | .07 (.04) |
Daily number of colleagues work with | .04 (.03) | .06 (.03) | .04 (.03) | .06 (.03) | .04 (.03) | .06 (.03) | .04 (.03) | .06 (.03) | .04 (.03) | .06 (.03) |
Daily number of leaders work with | −.02 (.04) | .02 (.04) | −.04 (.04) | .01 (.04) | −.04 (.04) | .01 (.05) | −.04 (.04) | .01 (.04) | −.04 (.04) | .01 (.05) |
Daily number of departments work with | −.08 (.04) | −.03 (.04) | −.08 (.04)* | −.03 (.04) | −.08 (.04) | −.03 (.04) | −.07 (.04) | −.01 (.04) | −.07 (.04) | −.01 (.04) |
Daily commuting time | .01 (.03) | .03 (.03) | .01 (.03) | .03 (.03) | .01 (.03) | .03 (.04) | .02 (.03) | .04 (.03) | .02 (.03) | .04 (.04) |
Management | −.02 (.07) | .01 (.08) | −.02 (.07) | .01 (.08) | −.02 (.07) | .01 (.07) | −.02 (.07) | .01 (.08) | −.02 (.07) | .01 (.07) |
Research | −.01 (.08) | .02 (.08) | −.02 (.08) | .00 (.08) | −.02 (.08) | .00 (.09) | −.02 (.08) | .01 (.08) | −.02 (.08) | .01 (.08) |
Service | −.09 (.08) | −.04 (.09) | −.10 (.08) | −.06 (.09) | −.10 (.08) | −.06 (.09) | −.10 (.08) | −.05 (.08) | −.10 (.08) | −.05 (.09) |
Marketing | −.01 (.07) | −.07 (.07) | −.01 (.07) | −.06 (.07) | −.01 (.07) | −.06 (.07) | −.02 (.07) | −.07 (.07) | −.02 (.07) | −.07 (.07) |
Other | −.12 (.09) | −.09 (.10) | −.13 (.09) | −.10 (.10) | −.13 (.10) | −.10 (.10) | −.11 (.09) | −.08 (.09) | −.11 (.10) | −.08 (.10) |
Position levels | .00 (.05) | .01 (.05) | −.01 (.05) | −.01 (.05) | −.01 (.05) | −.01 (.06) | .00 (.05) | .01 (.05) | .00 (.05) | .01 (.05) |
Government | .15 (.21) | .18 (.22) | .13 (.21) | .15 (.22) | .13 (.21) | .15 (.20) | .14 (.20) | .16 (.21) | .14 (.19) | .16 (.20) |
Public institutions | .02 (.13) | .08 (.14) | .00 (.13) | .06 (.14) | .00 (.09) | .06 (.09) | .01 (.13) | .06 (.14) | .01 (.09) | .06 (.09) |
Foreign‐funded enterprise and joint venture | .07 (.15) | .01 (.15) | .05 (.15) | −.01 (.15) | .05 (.11) | −.01 (.11) | .06 (.14) | .00 (.15) | .06 (.11) | .00 (.11) |
State‐owned enterprise | −.08 (.15) | .03 (.15) | −.10 (.15) | .01 (.15) | −.10 (.11) | .01 (.11) | −.10 (.14) | −.01 (.15) | −.10 (.11) | −.01 (.11) |
Private enterprise | −.02 (.14) | .03 (.15) | −.03 (.14) | .02 (.14) | −.03 (.10) | .02 (.10) | −.03 (.13) | .01 (.14) | −.03 (.10) | .01 (.10) |
Age | .01 (.03) | .02 (.04) | −.01 (.03) | .01 (.04) | −.01 (.03) | .01 (.04) | .01 (.03) | .02 (.03) | .01 (.03) | .02 (.04) |
Gender | .02 (.05) | .00 (.05) | .01 (.05) | −.01 (.05) | .01 (.05) | −.01 (.06) | .01 (.05) | −.02 (.05) | .01 (.05) | −.02 (.06) |
Education | .04 (.04) | −.04 (.05) | .05 (.04) | −.04 (.05) | .05 (.05) | −.04 (.05) | .06 (.04) | −.02 (.04) | .06 (.05) | −.02 (.05) |
Married, have a child or children | −.12 (.09) | −.06 (.10) | −.14 (.09) | −.08 (.10) | −.14 (.09) | −.08 (.10) | −.15 (.09) | −.09 (.09) | −.15 (.08) | −.09 (.09) |
Single, no child | −.02 (.09) | .12 (.10) | −.04 (.09) | .10 (.10) | −.04 (.09) | .10 (.10) | −.04 (.09) | .10 (.10) | −.04 (.09) | .10 (.09) |
Single, have a child or children | −.36 (.37) | −.39 (.39) | −.49 (.37) | −.58 (.39) | −.49 (.23)* | −.58 (.19)** | −.38 (.36) | −.41 (.38) | −.38 (.20) | −.41 (.21) |
Days of starting work after Chinese New Year | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) | .00 (.00) |
WFH training | .07 (.06) | .09 (.06) | .05 (.06) | .06 (.06) | .05 (.06) | .06 (.07) | .05 (.05) | .06 (.06) | .05 (.06) | .06 (.06) |
Tables 5 and and6 6 present the results of the moderating analysis of social support by applying hierarchical regressions. The results from the first four regression models consider the direct impact of WFH, job control, job demand and social support on self‐reported job performance as benchmark (Models 1–4 in Tables 5 and and6). 6 ). Models 5–7 test the moderating effect of employers' social support on the relationships between job control, job demand and social support with job performance. We initially find that the social support is significantly positively related to satisfaction with quality (0.16, p < 0.001) and productivity (0.27, p < 0.001). Toward the moderating effect of employers' anti‐epidemic policy, we find the interaction terms of job demand*social support to be only significant on the regressions on satisfaction of productivity (−0.07, p < 0.05). That is, hypothesis 4a is supported.
Overall, the results of testing the hypotheses are shown in Table 7 and Figure 2 .
Results of hypotheses
Hypotheses | Findings | Accept/Reject |
---|---|---|
H1a: Employees who work from home are more satisfied with their job performance | Significance only can be seen in terms of Quality (8.83***) (Evidence from Table ) | Partly accept |
H1b: Employees who work from home are less satisfied with their job performance | Significance only can be seen in terms of Productivity (−3.1**) (Evidence from Table ) | Partly accept |
H2: Job demand, at least in part, negatively mediates the relationship between WFH and job performance | Job demand negatively mediates, in part, between the WFH and the job performance (Productivity: .02*, 12.5%; Quality: .14***, 23.72%) (Evidence from Table ) | Accept |
H3: The relationship between WFH and job performance is mediated, in part, by job control | Job control negatively mediates, in part, between the WFH and the job performance (Productivity: .03**, 15.78%; Quality: .08***, 14.28%) (Evidence from Table ) | Accept |
H4a: Social support negatively moderates the relationship between job demand and job performance | Interaction term job demand*social support is significant on the regressions on satisfaction of productivity (−.10*). (Evidence from Tables and 6) | Partly accept |
H4b: Social support positively moderates the relationship between job control and job performance | Non‐significance (Evidence from Tables and 6) | Reject |
* p < 0.05; ** p < 0.01; *** p < 0.001.
The hypotheses results presented in the conceptual framework
In responding to the inconsistent findings on the impact of WFH on job performance, the present paper found that WFH helps promote job performance in terms of quality but leads to poor job performance in terms of productivity, which indicates that WFH may not always play an ‘either‐or’ (positive or negative) role, as previous theories suggest. To explore the causal mechanism underpinning the findings, based on the JDCS model, we found that WFH affects job performance via job demand and control path, moderated by social support, which indicate that WFH leads to flexibility, and employees have more autonomy to work at any timepoint per day to finalize their job. They usually choose the timepoint to conduct work when they have a desirable working condition, consequently cultivate focus, concentration and creativity (Hunter 2019 ). Accordingly, job quality can be enhanced. Despite a good job quality, WFH employees devote higher job demand. Thus, it is not conducive to job productivity than WFO employees. In addition, we found the positive moderating role of social support from organizations to enhance job performance during epidemic crisis.
The present paper aims to contribute in several ways. Our study extends the JDCS model under the context of COVID‐19 by investigating whether WFH can render the change in job control and job demand and exert influence on employees' job performance with the moderating effect of employers' support. The JDCS model can also help explain why WFH plays a mixed role to affect job performance. Prior studies have mainly qualitatively discussed changes to the way that individuals work during the COVID‐19 pandemic (Wang et al. 2021 ), the advantages and disadvantages of enforced WFH (Hallman et al. 2021 ; Purwanto 2020 ), ICT functions that enable to offer affordance to satisfy WFH targets (Waizenegger et al. 2020 ), and the way to provide a resource for WFH (Hafermalz and Riemer 2021 ). Research that indicates why WFH can affect employees' work‐related outcomes, particularly with empirical evidence, is limited. By applying a sample collected in China, we investigated two paths (i.e. job demand and job control) and a boundary condition (support) of the relationship between WFH and job performance.
Our results show that job control and job demand positively mediate the relationship between WFH and job performance. The increased job control and decreased job demand by applying WFH can be considered one of the main reasons WFH helps enhance job quality. This finding is notable because this study tends to clarify the mixed mechanism that WFH affects work‐related outcomes from the perspective of job characteristics and provides a theoretical framework. In terms of job productivity, we find that the increased job control and decreased job demand trade off the negative effect of WFH on productivity. Therefore, when explaining why WFH compared with WFO varies in job performance, the verified mediating effect of job control and job demand underpinned by the JDCS model can only account for job quality enhancement, rather than sufficiently support why WFH lowers job productivity.
The present paper also articulates the specific job control (‘talking right’ and ‘work rate’) and job demand (‘a long time of intense concentration’ and ‘hecticness of the job’) items are vital factors in performance enhancements. On the basis of such findings, we can presume that the ‘talking right’ enhanced by WFH implies that the enforced ‘physical distance’ may shorten the ‘power distance’ inscribed in hierarchical structure, because ICT enables communication flattening information transmitting in traditional stratified management. Reciprocally, such physical distance reduces redundant commands from managers, and workplace distractions trigger WFH employees to have more autonomy on ‘working rate’. Thereafter, in the wake of alleviations on ‘a long time of intense concentration’ and ‘hecticness of the job’, performance is enhanced.
Furthermore, we applied entropy balance matching, a method that has been regarded with more advantages for controlling self‐selection bias in quasi‐experiment research. Future studies could also adopt entropy balance matching to control self‐selection from process control, especially in the crisis context.
Empirically, post COVID‐19, WFH may become a vital HRM strategy. According to the Gartner CFO Survey (2020), 74% of companies plan to shift some of their employees to remote working temporarily. Our findings may imply several valuable tips for organizational employers and employees if one wants to accommodate employees to WFH for the long term. We suggest that sustained and pragmatic WFH policy in terms of ‘set working hours’ and ‘taking regular breaks’ should be designed to reduce job demands, such as ‘a long time of intense concentration’ and ‘hecticness of the job’. Furthermore, employers may leave employees more empowerment on scheduling, enhance the equality among different hierarchy people, and avoid lengthy and discursive commands while working to improve the ‘talking right’ and ‘work rate’ autonomy for employees. In addition, social support is found to be a critical boundary condition between WFH and job characteristics. Thus, it is vital that sound and feasible epidemic policies, such as providing personal protective equipment, a financial sponsored program, psychological counselling and support, are put in place and executed as crucial responsibilities (Shani and Pizam 2009 ). And finally, employers need to be aware that more resources should be available for increased virtual collaboration needs as WFH has now taken hold and will be around for a long time in the future.
First, even though in the present study we have controlled for a wide range of variables that may potentially relate to job performance, inevitably, it still misses some relevant variables. For example, even though we have involved communication factors under control, technology fatigue may still contribute significantly on change of job demands and subsequently affect job performance (Yang et al. 2021 ). Second, our dataset is a cross‐sectional one and we asked employees to rate job performance rather than multilevel respondents. The absence of lagged performance data restricts the possibility of examining the long‐term effect of WFH on job performance and relationships between the variables of interest. As already noted, the current sample was collected at the early period of ending epidemic lockdown. By applying the cross‐sectional model, identifying the potential time variance (e.g. honeymoon effect) from the targeted relationship is difficult. Thus, future studies should adopt panel data and compare the present study to test for robustness.
This work was supported by the National Natural Science Foundation of China (grant number 72102033); Shanghai 2020 Science and Technology Innovation Action Plan (grant number 21692102600); the Fundamental Research Funds for the Central Universities of China (grant number N2206012); the Humanities and Social Science Foundation of the Ministry of Education of China (grant number 21YJC630153); the Social Science Foundation of Liaoning in China (grant number L21CGL013).
We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled, ‘Working from home vs. working from office in terms of job performance during COVID‐19 pandemic crisis: evidence from China’.
Jingjing Qu is an associate professor at Shanghai AI Lab, China. Her research interests include artificial intelligence governance, artificial intelligence technology innovation and well‐being of entrepreneurs.
Jiaqi Yan is a lecturer at School of Business and Administration of Northeastern University. He received his PhD degree from Tongji University and studied as a joint PhD student at the University of Sydney. His research interests include human resource management, hospitality management and entrepreneurship.
The White House 1600 Pennsylvania Ave NW Washington, DC 20500
Today, the Council of Economic Advisers under the leadership of Chair Jared Bernstein released the 2024 Economic Report of the President , the 78 th report since the establishment of CEA in 1946. The 2024 Report brings economic evidence and data to bear on many of today’s most significant issues and questions in domestic and international economic policy:
Chapter 1, The Benefits of Full Employment , which is dedicated to the late Dr. William Spriggs, examines the labor market, distributional, and macroeconomic impacts of full employment, with a particular focus on the benefits for economically vulnerable groups of workers who are much more likely to be left behind in periods of weak labor markets.
Chapter 2, The Year in Review and the Years Ahead , describes macroeconomic and financial market trends in 2023 and presents the Federal government’s FY 2024 macroeconomic forecast.
Chapter 3, Population, Aging, and the Economy , explains how long-run trends in fertility and mortality are shaping the U.S. population and labor force.
Chapter 4, Increasing the Supply of Affordable Housing, explores the causes and consequences of the nation’s longstanding housing shortage and how the Biden-Harris administration’s policy agenda can significantly increase the production of more affordable housing.
Chapter 5, International Trade and Investment Flows, presents key facts about long-term trends in U.S. international trade and investment flows, including the role of global supply chains, and highlights the benefits and costs of global integration for American workers.
Chapter 6, Accelerating the Clean Energy Transition , applies a structural change framework to explain the factors that can accelerate the transition towards a clean energy economy.
Chapter 7, An Economic Framework for Understanding Artificial Intelligence , uses an economic framework to explore when, how, and why AI may be adopted, adapting standard economic models to explore AI’s potential effects on labor markets, while examining policy decisions that will affect social and macroeconomic outcomes.
We'll be in touch with the latest information on how President Biden and his administration are working for the American people, as well as ways you can get involved and help our country build back better.
Opt in to send and receive text messages from President Biden.
If 2023 was the year the world discovered generative AI (gen AI) , 2024 is the year organizations truly began using—and deriving business value from—this new technology. In the latest McKinsey Global Survey on AI, 65 percent of respondents report that their organizations are regularly using gen AI, nearly double the percentage from our previous survey just ten months ago. Respondents’ expectations for gen AI’s impact remain as high as they were last year , with three-quarters predicting that gen AI will lead to significant or disruptive change in their industries in the years ahead.
This article is a collaborative effort by Alex Singla , Alexander Sukharevsky , Lareina Yee , and Michael Chui , with Bryce Hall , representing views from QuantumBlack, AI by McKinsey, and McKinsey Digital.
Organizations are already seeing material benefits from gen AI use, reporting both cost decreases and revenue jumps in the business units deploying the technology. The survey also provides insights into the kinds of risks presented by gen AI—most notably, inaccuracy—as well as the emerging practices of top performers to mitigate those challenges and capture value.
Interest in generative AI has also brightened the spotlight on a broader set of AI capabilities. For the past six years, AI adoption by respondents’ organizations has hovered at about 50 percent. This year, the survey finds that adoption has jumped to 72 percent (Exhibit 1). And the interest is truly global in scope. Our 2023 survey found that AI adoption did not reach 66 percent in any region; however, this year more than two-thirds of respondents in nearly every region say their organizations are using AI. 1 Organizations based in Central and South America are the exception, with 58 percent of respondents working for organizations based in Central and South America reporting AI adoption. Looking by industry, the biggest increase in adoption can be found in professional services. 2 Includes respondents working for organizations focused on human resources, legal services, management consulting, market research, R&D, tax preparation, and training.
Also, responses suggest that companies are now using AI in more parts of the business. Half of respondents say their organizations have adopted AI in two or more business functions, up from less than a third of respondents in 2023 (Exhibit 2).
Most respondents now report that their organizations—and they as individuals—are using gen AI. Sixty-five percent of respondents say their organizations are regularly using gen AI in at least one business function, up from one-third last year. The average organization using gen AI is doing so in two functions, most often in marketing and sales and in product and service development—two functions in which previous research determined that gen AI adoption could generate the most value 3 “ The economic potential of generative AI: The next productivity frontier ,” McKinsey, June 14, 2023. —as well as in IT (Exhibit 3). The biggest increase from 2023 is found in marketing and sales, where reported adoption has more than doubled. Yet across functions, only two use cases, both within marketing and sales, are reported by 15 percent or more of respondents.
Gen AI also is weaving its way into respondents’ personal lives. Compared with 2023, respondents are much more likely to be using gen AI at work and even more likely to be using gen AI both at work and in their personal lives (Exhibit 4). The survey finds upticks in gen AI use across all regions, with the largest increases in Asia–Pacific and Greater China. Respondents at the highest seniority levels, meanwhile, show larger jumps in the use of gen Al tools for work and outside of work compared with their midlevel-management peers. Looking at specific industries, respondents working in energy and materials and in professional services report the largest increase in gen AI use.
The latest survey also shows how different industries are budgeting for gen AI. Responses suggest that, in many industries, organizations are about equally as likely to be investing more than 5 percent of their digital budgets in gen AI as they are in nongenerative, analytical-AI solutions (Exhibit 5). Yet in most industries, larger shares of respondents report that their organizations spend more than 20 percent on analytical AI than on gen AI. Looking ahead, most respondents—67 percent—expect their organizations to invest more in AI over the next three years.
Where are those investments paying off? For the first time, our latest survey explored the value created by gen AI use by business function. The function in which the largest share of respondents report seeing cost decreases is human resources. Respondents most commonly report meaningful revenue increases (of more than 5 percent) in supply chain and inventory management (Exhibit 6). For analytical AI, respondents most often report seeing cost benefits in service operations—in line with what we found last year —as well as meaningful revenue increases from AI use in marketing and sales.
As businesses begin to see the benefits of gen AI, they’re also recognizing the diverse risks associated with the technology. These can range from data management risks such as data privacy, bias, or intellectual property (IP) infringement to model management risks, which tend to focus on inaccurate output or lack of explainability. A third big risk category is security and incorrect use.
Respondents to the latest survey are more likely than they were last year to say their organizations consider inaccuracy and IP infringement to be relevant to their use of gen AI, and about half continue to view cybersecurity as a risk (Exhibit 7).
Conversely, respondents are less likely than they were last year to say their organizations consider workforce and labor displacement to be relevant risks and are not increasing efforts to mitigate them.
In fact, inaccuracy— which can affect use cases across the gen AI value chain , ranging from customer journeys and summarization to coding and creative content—is the only risk that respondents are significantly more likely than last year to say their organizations are actively working to mitigate.
Some organizations have already experienced negative consequences from the use of gen AI, with 44 percent of respondents saying their organizations have experienced at least one consequence (Exhibit 8). Respondents most often report inaccuracy as a risk that has affected their organizations, followed by cybersecurity and explainability.
Our previous research has found that there are several elements of governance that can help in scaling gen AI use responsibly, yet few respondents report having these risk-related practices in place. 4 “ Implementing generative AI with speed and safety ,” McKinsey Quarterly , March 13, 2024. For example, just 18 percent say their organizations have an enterprise-wide council or board with the authority to make decisions involving responsible AI governance, and only one-third say gen AI risk awareness and risk mitigation controls are required skill sets for technical talent.
The latest survey also sought to understand how, and how quickly, organizations are deploying these new gen AI tools. We have found three archetypes for implementing gen AI solutions : takers use off-the-shelf, publicly available solutions; shapers customize those tools with proprietary data and systems; and makers develop their own foundation models from scratch. 5 “ Technology’s generational moment with generative AI: A CIO and CTO guide ,” McKinsey, July 11, 2023. Across most industries, the survey results suggest that organizations are finding off-the-shelf offerings applicable to their business needs—though many are pursuing opportunities to customize models or even develop their own (Exhibit 9). About half of reported gen AI uses within respondents’ business functions are utilizing off-the-shelf, publicly available models or tools, with little or no customization. Respondents in energy and materials, technology, and media and telecommunications are more likely to report significant customization or tuning of publicly available models or developing their own proprietary models to address specific business needs.
Respondents most often report that their organizations required one to four months from the start of a project to put gen AI into production, though the time it takes varies by business function (Exhibit 10). It also depends upon the approach for acquiring those capabilities. Not surprisingly, reported uses of highly customized or proprietary models are 1.5 times more likely than off-the-shelf, publicly available models to take five months or more to implement.
Gen AI is a new technology, and organizations are still early in the journey of pursuing its opportunities and scaling it across functions. So it’s little surprise that only a small subset of respondents (46 out of 876) report that a meaningful share of their organizations’ EBIT can be attributed to their deployment of gen AI. Still, these gen AI leaders are worth examining closely. These, after all, are the early movers, who already attribute more than 10 percent of their organizations’ EBIT to their use of gen AI. Forty-two percent of these high performers say more than 20 percent of their EBIT is attributable to their use of nongenerative, analytical AI, and they span industries and regions—though most are at organizations with less than $1 billion in annual revenue. The AI-related practices at these organizations can offer guidance to those looking to create value from gen AI adoption at their own organizations.
To start, gen AI high performers are using gen AI in more business functions—an average of three functions, while others average two. They, like other organizations, are most likely to use gen AI in marketing and sales and product or service development, but they’re much more likely than others to use gen AI solutions in risk, legal, and compliance; in strategy and corporate finance; and in supply chain and inventory management. They’re more than three times as likely as others to be using gen AI in activities ranging from processing of accounting documents and risk assessment to R&D testing and pricing and promotions. While, overall, about half of reported gen AI applications within business functions are utilizing publicly available models or tools, gen AI high performers are less likely to use those off-the-shelf options than to either implement significantly customized versions of those tools or to develop their own proprietary foundation models.
What else are these high performers doing differently? For one thing, they are paying more attention to gen-AI-related risks. Perhaps because they are further along on their journeys, they are more likely than others to say their organizations have experienced every negative consequence from gen AI we asked about, from cybersecurity and personal privacy to explainability and IP infringement. Given that, they are more likely than others to report that their organizations consider those risks, as well as regulatory compliance, environmental impacts, and political stability, to be relevant to their gen AI use, and they say they take steps to mitigate more risks than others do.
Gen AI high performers are also much more likely to say their organizations follow a set of risk-related best practices (Exhibit 11). For example, they are nearly twice as likely as others to involve the legal function and embed risk reviews early on in the development of gen AI solutions—that is, to “ shift left .” They’re also much more likely than others to employ a wide range of other best practices, from strategy-related practices to those related to scaling.
In addition to experiencing the risks of gen AI adoption, high performers have encountered other challenges that can serve as warnings to others (Exhibit 12). Seventy percent say they have experienced difficulties with data, including defining processes for data governance, developing the ability to quickly integrate data into AI models, and an insufficient amount of training data, highlighting the essential role that data play in capturing value. High performers are also more likely than others to report experiencing challenges with their operating models, such as implementing agile ways of working and effective sprint performance management.
The online survey was in the field from February 22 to March 5, 2024, and garnered responses from 1,363 participants representing the full range of regions, industries, company sizes, functional specialties, and tenures. Of those respondents, 981 said their organizations had adopted AI in at least one business function, and 878 said their organizations were regularly using gen AI in at least one function. To adjust for differences in response rates, the data are weighted by the contribution of each respondent’s nation to global GDP.
Alex Singla and Alexander Sukharevsky are global coleaders of QuantumBlack, AI by McKinsey, and senior partners in McKinsey’s Chicago and London offices, respectively; Lareina Yee is a senior partner in the Bay Area office, where Michael Chui , a McKinsey Global Institute partner, is a partner; and Bryce Hall is an associate partner in the Washington, DC, office.
They wish to thank Kaitlin Noe, Larry Kanter, Mallika Jhamb, and Shinjini Srivastava for their contributions to this work.
This article was edited by Heather Hanselman, a senior editor in McKinsey’s Atlanta office.
Related articles.
We’re a group of 3,000 researchers inventing what’s next in computing at labs across the world. Learn more about us and our work below.
IMAGES
VIDEO
COMMENTS
The phenomenon of working from home (WFH) is linked to a variety of megatrends that companies have confronted over many years. These include demographic change, leading to shortages of skilled workers in many regions and professions; the individualisation of needs and lifestyles as a result of changing values; and most particularly, the digitalisation of the world of work (Schmoll and Süß ...
This thesis is a research on effects of remote work on workers and the workplace and how workers react to a switch to remote work. The research was carried out in re- ... employees to working from home away from the office to avoid unnecessary contact, risk-ing a spread of the virus among the workforce. After the start of the pandemic, all busi-
Working from home also seems to increase productivity and work hours, which may be due to less time commuting, less time on social activities, or other benefits of controlling one's work environment (Choudhury, Foroughi, and Larson 2021). There are some downsides to working from home, including reduced connectedness to coworkers.
The flexibility allowed by the mobilization of technology disintegrated the traditional work-life boundary for most professionals. Whether working from home is the key or impediment to academics' efficiency and work-life balance became a daunting question for both scientists and their employers. The recent pandemic brought into focus the merits and challenges of working from home on a level ...
The coronavirus disease 2019 (COVID-19) pandemic has impacted the world economy in various ways. In particular, the drastic shift to telework has dramatically changed how people work. Whether the new style of working from home (WFH) will remain in our society highly depends on its effects on workers' productivity. However, to the best of our knowledge, the effects of WFH on productivity are ...
Central issues around the organisational and societal phenomenon of working from home are linked to the resources and strains of employees. This has direct consequences for the leadership and management of human resources. In this article, we review the results of research contributions available in this issue and at the same time show that ...
working environments where they have a personal connection to the organization's mission and vision and where they feel the work culture is familial. The taxonomy derived from this research could provide organizational leaders with techniques to engage and inspire the talent of remote workers to create positive and sustainable social change.
During the COVID-19 pandemic, working from home has unquestionably become one of the most extensively employed techniques to minimize unemployment, keep society operating, and shield the public from the virus. However, the impacts of work-from-home (WFH) on employee productivity and performance is not fully known; studies on the subject are fragmented and in different contexts. The purpose of ...
April 1, 2022. Abstract. W e study employee productivity (output per hour work ed) before and during the working from. home period of the Covid-19 pandemic, using personnel and analytics data from ...
This study synthesizes the empirical research on Work From Home (WFH) during COVID-19 from 2020 to 2022. The study addressed two objectives; (1) to find the common WFH practices during the COVID-19 2020-2022 period and (2) to find the areas that are not common in WFH practices in the research landscape during COVID-19 in 2020-2022 and (3) the areas that need attention in future research on ...
While work from home during 2020 and 2021 was not typical of working from home prior to the pandemic, it was better than expected for many, and attitudes and norms have changed. Workers and firms are now embarking on a second wave of experimentation — negotiating, trialling and adjusting — to see what best works for them.
The Covid-19 pandemic sparked what economist Nicholas Bloom calls the " working-from-home economy .". While some workers may have had flexibility to work remotely before the pandemic, this ...
As a first step, the present paper explored how. working from home affects researchers' efficiency and well-being. Our results showed that while the pandemic-related lockdown decreased the work ...
1. Introduction. As a consequence of the COVID-19 pandemic, about 72% of employees worldwide were required to switch overnight to working from home (WFH) [].According to a Survey Monkey report, more than 89% of employees surveyed (n = 9059) were satisfied with their WFH arrangements [].However, a Martec Group 2020 study reported that only 32% of respondents (n = 1214) were satisfied with their ...
Productivity increased 4.4 percent when employees moved from working at home on a limited basis to the location of their choice. Based on a patent's average value, this productivity gain could add $1.3 billion of value to the US economy each year, the researchers estimate. Sources: US Patent and Trademark Office; " (Live) and Work from ...
Research Proposal PDF Available. A Study on Work from Home, Motivation & Productivity of Employees in Indian Population during COVID-19 Pandemic. October 2020;
The first phase of this research identified eight key themes from the review of existing research on homeworking during the COVID-19 lockdown: Increased productivity among homeworkers is often achieved through work intensification. For some workers, homeworking can provide a more productive environment because there are fewer distractions.
The study results indicate that working from home has an influence on productivity of the employees. While influences of some of the factors are either positive or negative, the influence of some of the factors depend on the characteristics and attitude of the employees and the circumstances. 1. Introduction.
Research proposal examples. Writing a research proposal can be quite challenging, but a good starting point could be to look at some examples. We've included a few for you below. Example research proposal #1: "A Conceptual Framework for Scheduling Constraint Management".
The proposal, rejected by U.S. military research agency DARPA, describes the insertion of human-specific cleavage sites into SARS-related bat coronaviruses. Sharon Lerner, Maia Hibbett. September ...
The extensive research on work-life conflict, should help us examine the issue and to develop coping strategies applicable for academics' life. The Boundary Theory [26, 51, 52] proved to be a useful framework to understand the work-home interface. According to this theory, individuals utilize different tactics to create and maintain an ideal ...
May 27, 2024 9:00 pm ET. Listen to article. (2 minutes) More Americans are profoundly lonely, and the way they work—more digitally linked but less personally connected—is deepening that sense ...
Nail the Tone. The tone of your proposal should reflect both your personal brand and the client's culture. But to be on the safe side, err more towards mirroring the client. As my mother always ...
Mission. The Purdue On-Campus Writing Lab and Purdue Online Writing Lab assist clients in their development as writers—no matter what their skill level—with on-campus consultations, online participation, and community engagement. The Purdue Writing Lab serves the Purdue, West Lafayette, campus and coordinates with local literacy initiatives.
Professional development for large groups. Give your teams access to the training they need to craft grant winning proposals and fundraise effectively. Our experts can put together a training package to suit your organization's goals. Candid Learning offers information and resources that are specifically designed to meet the needs of ...
Abstract. This research aims to study the challenges and remedies of Work From Home (WFH) during the COVID-19 pandemic. Many organizations around the globe are forced to let employees work from ...
Literature review. WFH is a working arrangement in which employees fulfill the essential responsibilities that their job entails while remaining at home using ICT (International Labor Organization 2020, 5).Although a slight difference exists among terms such as WFH, teleworking, telecommuting and remote working, these concepts are largely interchangeable.
Today, the Council of Economic Advisers under the leadership of Chair Jared Bernstein released the 2024 Economic Report of the President, the 78th report since the establishment of CEA in 1946 ...
If 2023 was the year the world discovered generative AI (gen AI), 2024 is the year organizations truly began using—and deriving business value from—this new technology. In the latest McKinsey Global Survey on AI, 65 percent of respondents report that their organizations are regularly using gen AI, nearly double the percentage from our ...
A global community. A global. community. We're a group of 3,000 researchers inventing what's next in computing at labs across the world. Learn more about us and our work below. Filter by.