The demands and resources arising from shared office spaces

Applied Ergonomics 60 (2017) 103e115

Contents lists available at ScienceDirect

Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo

The demands and resources arising from shared office spaces Rachel L. Morrison a, *, Keith A. Macky b a b

Management Dept., Faculty of Business and Law, Auckland University of Technology, Private Bag 92006, Auckland 1142, New Zealand Independent HR and Research Consultant, New Zealand

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 December 2015 Received in revised form 8 November 2016 Accepted 9 November 2016

The prevalence of flexible and shared office spaces is increasing significantly, yet the socioemotional outcomes associated with these environments are under researched. Utilising the job demandsresources (JD-R) model we investigate both the demands and the resources that can accrue to workers as a result of shared work environments and hot-desking. Data were collected from work experienced respondents (n ¼ 1000) assessing the extent to which they shared their office space with others, along with demands comprising distractions, uncooperative behaviours, distrust, and negative relationships, and resources from co-worker friendships and supervisor support. We found that, as work environments became more shared (with hot-desking being at the extreme end of the continuum), not only were there increases in demands, but co-worker friendships were not improved and perceptions of supervisory support decreased. Findings are discussed in relation to employee well-being and recommendations are made regarding how best to ameliorate negative consequences of shared work environments. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Hot-desking Job demands-resources Co-worker relationships

1. Introduction Research on the impact of the physical working environment on the social systems within organisations has a long pedigree, from the socio-technical approach of the 1950s and 60s and the ‘quality of working life’ movement of the 1970s and 80s (e.g., Davis and Cherns, 1975), through to a more contemporary interest in the impact the built environment has on employee well-being (e.g., Cooper and Boyko, 2009). Much of this historical research has been focused within manufacturing and the factory environment; with a concentration on a limited range of environmental variables such as temperature, air quality, ambient noise levels, lighting, and the design of physical production systems and labour process technologies, including lean manufacturing (Cullinane et al., 2013) and team working (Appelbaum et al., 2000). However, as we move into a new century the dominance of manufacturing has waned in developed economies. Services typically now account for three-quarters of employment in those economies (Boxall and Purcell, 2011), with the majority of these being white-collar workers in office settings. In America for example, more than 70 percent of workers are located in offices

* Corresponding author. E-mail address: [email protected] (R.L. Morrison). http://dx.doi.org/10.1016/j.apergo.2016.11.007 0003-6870/© 2016 Elsevier Ltd. All rights reserved.

(Elsbach and Pratt, 2007). Yet the impact that physical work environments have on office workers is a surprisingly under researched area in mainstream organisational psychology and Human Resource Management (Ashkanasy et al., 2014). And while some research does exist in the ergonomics literature (Kaarlela-Tuomaala et al., 2009) and in publications with a focus on facilities management (Warren, 2003), education (Rytivaara, 2011), and real estate (Voordt, 2004), the topic remains relatively under explored. The last decade has also brought about significant changes in the physical spaces many office workers now find themselves in. Until the turn of the century, most white collar workers remained securely bound to their office and their desk, simply because the tools to do their job were fixed in one place (Felstead et al., 2003). Portable computers, tablets, and smart phones have enabled a significant change in the physical spaces and places that work is now carried out. This, together with the high cost of office space (only human resources are more expensive; McCoy, 2005), has brought about a desire to encourage tele-work (working remotely) (Bentley et al., 2016) and to use physical office space more flexibly. One example of this has been the use of “non-territorial workspaces” or “hot-desks” (also termed “hoteling”) (Elsbach, 2003; p 622), which are found in workplaces where “…staff have no fixed personal workspace and use any available desk as needed” (Felstead et al., 2003, p. 16). These arrangements rely on flexible ICT systems and are characterised by interchangeable workstations or

104

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

the use of portable devices with internet connectivity. The move towards shared work environments and hot-desks is certainly not new, and the limited research to date has been generally positive. For example, flexible and/or shared work environments have been associated with greater employee satisfaction (Cole et al., 2012; Sundstrom et al., 1980), projecting an image of being modern and forward thinking (McElroy and Morrow, 2010), improving flexibility in the use of the physical space (Elsbach, 2003), enabling closer working relationships (Chigot, 2003; McElroy and Morrow, 2010), higher productivity (Cole et al., 2012; Meijer et al., 2009), more easily exchanged knowledge and skills (Ashkanasy et al., 2014; Chigot, 2003; Irving and Ayoko, 2014), increased networking opportunities (Elsbach and Bechky, 2007) and cost-savings (Duffy, 1997; Fawcett and Chadwick, 2007; McElroy and Morrow, 2010; Voordt, 2004; Warren, 2003). These cost savings are generally made by using the available space more intensively and, in addition, the amount of physical space needed does not have to increase in proportion to the number of employees, so organisations can delay acquiring new space as they grow (Elsbach, 2003; Elsbach and Bechky, 2007). On the other hand, there can also be negative consequences of not allowing employees to have their own work space; be it an office, a cubicle, or even a regular spot in an otherwise open-plan room. Sundstrom et al. (1994) for example, report that while employees may tolerate ambient noise from office equipment, overhearing the conversations of others (inevitable in open plan workplaces) is a significant task distraction and source of irritation. It is the potential for negative outcomes that form the focus of this study. Using the job demands-resources (JD-R) model (Bakker and Demerouti, 2007; Demerouti et al., 2001) we investigate both the demands and resources that can accrue to workers as a result of shared work environments and hot-desking. We begin by framing the prior research related to working with others in shared environments (including hot-desking) within JD-R model; following this, the remainder of the paper is conventionally structured with an outline of our methods, participants and findings followed by discussion and conclusions. 2. Theoretical framework: demands and resources within shared workspaces Within the JD-R model, job demands are the aspects of a job that require prolonged cognitive and/or emotional effort, thereby incurring physiological and/or psychological “costs” (Bakker and Demerouti, 2007). Bakker and Demerouti (2007) give examples of demands as including “…high work pressure, an unfavourable physical environment, and emotionally demanding interactions…” (p. 312). For the purposes of the current study, rather than workload per se, we focus specifically on those demands outlined by Bakker and Demerouti (2007) which have been found to be associated with shared work environments. More specifically, we propose that shared office spaces, and particularly hot-desking, place additional demands and increased load on workers by creating an unfavourable physical working environment. Also termed indoor environment quality (IEQ) (Kim and de Dear, 2013), this becomes detrimental to the individual located within it through reduced privacy, increased social distraction, and negative or emotionally demanding interactions with others. Job resources are the aspects of a job that: a) help in achieving work goals, b) reduce the costs associated with job demands, or c) stimulate growth and development (Bakker and Demerouti, 2007). As Bakker and Demerouti (2007) note, research has identified a veritable laundry list of variables under the job resources rubric, including “social support from colleagues, supervisory support, and performance feedback” (p. 311). Our focus here is on the specific

resources which have been posited to come from working in shared spaces; that is: more collegial friendships and increased support from managers and others (Chigot, 2003; Elsbach and Bechky, 2007; Irving and Ayoko, 2014; McElroy and Morrow, 2010). More specifically, we investigate whether an increase in the resources, that flexible and/or shared environments are supposed to provide, are evident. Also relevant here is Bakker and Demerouti's (2007) notion of “dual processes”. On the one hand, demands will incur costs, draining an individual's energy and resulting in strain and health impairment. Resources, on the other hand have “motivational potential” and lead to increased engagement and performance, either through the satisfaction of basic needs (e.g., social support satisfying the “need to belong”) or through the achievement of desired work goals (e.g., supportive colleagues assisting with tasks) (Bakker and Demerouti, 2007). According to the JD-R model, in addition to the direct impact of demands and resources, there will also be an interaction whereby resources will potentially buffer the impact of demands (Bakker et al., 2005; Bakker et al., 2003; Haines et al., 1991). In the current study we examine the social support available in shared workspaces, which is the most well-known variable found to buffer against job strain (see for example; Bakker and Demerouti, 2007; Bakker et al., 2003, 2005; Haines et al., 1991), in relation to increased demands that might also arise from those same workspaces. Through the operation of these dual processes, we see the JD-R model as potentially useful in explaining what might appear to be contradictory findings regarding the impact of shared work environments. Where such environments provide resources to employees, then it would be reasonable to expect positive outcomes from such working arrangements. However, if demands also arise from shared working spaces, and these outweigh or counter those resources, then negative outcomes might also be anticipated. In some circumstances, the research shows positive outcomes (e.g., Chigot, 2003; McElroy and Morrow, 2010), while other researchers describe negative outcomes (e.g., Ashkanasy et al., 2014; Maher and von Hippel, 2005; Sundstrom et al., 1994). This study therefore aims to empirically ascertain if a relationship exists between such work environments and the potential job demands and resources accrued by workers. Below we first discuss the demands associated with open plan workspaces and then outline the possible resources that these spaces might afford; providing hypotheses derived from the literature. 2.1. Demands 2.1.1. Distractions Distractions resulting from a lack of privacy and increased noise are a key source of demand, and have been found to be a significant issue in open plan environments (Ashkanasy et al., 2014; Maher and von Hippel, 2005). Privacy includes both the ability to reduce or control incoming stimuli, and also to limit outgoing information (Altman, 1975; cited in Ashkanasy et al., 2014). In a study of indoor environment quality (IEQ) Kim and de Dear (2013) found that, in open plan offices, noise level and visual privacy were consistently negatively evaluated, but that “sound privacy” was by far the most unsatisfactory IEQ factor. In an open plan office employees have little control over their levels of privacy and this, in itself, becomes a source of job demand. There are consistent findings that distraction caused by overhearing irrelevant conversations is a major issue in open plan office environments and, further, that distraction is negatively linked with employee performance, negative perceptions of the workplace, and/or stress (see for example; Loewen and Suedfeld, 1992; Maher and von Hippel, 2005; Nemecek and Grandjean, 1973; Smith-Jackson and Klein, 2009; Sundstrom

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

et al., 1994). This literature informs our first hypothesis. H1. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report higher levels of distractions from work.

2.1.2. Uncooperative behaviours Several studies have focused on co-worker interaction in workplaces featuring “activity based working” (Koetsveld and Kamperman, 2011, p. 304) “flexible and non-dedicated workspaces” (Elsbach, 2003, p.622) and “open-plan offices” (Kim and de Dear, 2013, p. 18). Part of the rationale for adopting flexible and/or open plan work environments is to enable cooperation, networking and group work (Irving and Ayoko, 2014). The assumption is that a shared work environment will facilitate communication and interaction which will, in turn, improve performance and productivity (see for example: Brand and Smith, 2005; Chigot, 2003; Kupritz, 2003; cited in Kim and de Dear, 2013; Sundstrom et al., 1980). However, there is now evidence that the opposite is true; that is, that shared work environments have a negative impact on interpersonal interactions, including a decline in cooperative behaviours (Brennan et al., 2002). Open plan offices have been found to not only increase distraction and reduce privacy, but also to increase employees' use of coping strategies such as withdrawal (KaarlelaTuomaala et al., 2009), to negatively impact team member relations (Brennan et al., 2002), to make cooperation less pleasant (Kaarlela-Tuomaala et al., 2009), and to decrease communication when compared to private offices (Hatch, 1990; cited in McElroy and Morrow, 2010). Further, in terms of IEQ, private offices outperform open-plan workspaces in almost all aspects of IEQ (privacy, noise, visual distraction) (Kim and de Dear, 2013). While these authors found minimal benefits in terms of the ease with which colleagues could communicate with one another, these were far outweighed by the penalties of decreased IEQ. These findings, suggesting an increase in uncooperative behaviours (and therefore increased demands), in open plan office environments informs our second hypothesis. H2. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report higher levels of uncooperative behaviour from colleagues.

2.1.3. Negative relationships and distrust Research investigating negative workplace relationships generally supports the notion that workplace factors including workload, negative affect, and increased uncertainty and distrust are all associated with co-worker incivility and a deterioration of interpersonal relationships (see for example; Andersson and Pearson, 1999; Hutton, 2006; Labianca and Brass, 2006; Moerbeek and Need, 2003; Morrison and Nolan, 2007; Pearson et al., 2001). Research into shared work environments has been linked to these factors. This research is discussed below and informs the next two hypotheses. 2.1.3.1. Personalisation and territoriality: extra effort and uncertainty. While “…office work is increasingly becoming detached from individual and personalised cubes of space, marked by a walled cell or by an allocated desk…” (Felstead et al., 2003, p.22), there are those who value the certainty of knowing what and where their workspace will be for the coming day. Workplaces such as call centres (where employees typically work in their own small cubicle) are frequently decorated with family photos, plants, and children's drawings in an effort by employees to mark a territory as their own,

105

to make work seem more like “home”, or to stamp their own identity on the space. Research from the United States indicates that 70%e90% of employees who have their own workspace, personalise it (Wells and Thelen, 2002; Wells et al., 2007). Personalising a workspace may been thought of as “territorial behaviour” (Brown, 2009) and the ubiquity of this behaviour in society ) is (e.g., gang graffiti, locks on doors, or a jacket over a chair in a cafe evidence of the fundamental importance territory has for many people. Brown (2009) states that organisations are “fertile grounds for territoriality” (p. 44) and, because an employee's workspace is central to their experience of work, many will be motivated to both mark and protect even temporary spaces in a workplace. Physical working arrangements, such as hot-desking, that prevent territorial behaviours seem therefore to work against something fundamentally human. Coping with such an “unnatural” situation might reasonably be expected to increase demands on employees. There is also evidence that hot-desking creates additional demands for those employees forced to seek a free workspace each day, or even sometimes several times a day (Hirst, 2011). In addition to simply finding a workspace, another demand arises from the effort employees go to in order to satisfy their “need for differentiation” (Elsbach, 2003, p. 625). Elsbach (2003) found that workers, when stripped of the opportunity to personalise their workspace, would frequently engage in a variety of often prohibited behaviours, and expend a great deal of energy, over and above their assigned work, to assert their individuality. These activities were “effortful” and included squatting in offices that were meant to be vacant, reconstructing a specific area or territory for their group, displaying portable artefacts (e.g., pictures of their kids) or engaging in salient behaviours (e.g., talking about weekend activities). Often these displays were prohibited; breaking rules and norms of their organisation (such as saving spaces for others, not letting people use particular workstations, or “hard gluing” pictures onto their equipment). Brown (2009) describes an organisation which tried to create a non-territorial office; in response, employees “…quickly started claiming favorite spaces and reacted when others tried to use the supposedly “communal” resources.” (p. 44). 2.1.3.2. Psychosocial and psychological outcomes: negative affect. People usually decorate or modify their work environment to reflect their own identity; and personalisation has been found to have numerous positive psychological and psychosocial outcomes. Wells et al. (2007) describes the functions of workplace personalisation, claiming it will express an employee's personality, individuality, and uniqueness (Heidmets, 1994; Wells, 2000). In addition it has been found to buffer stress, evoke positive emotion (Scheiberg, 1990; Wells, 2000), and give workers a sense of control and agency over their work environment (Heidmets, 1994). Elsbach (2003) examined the loss of office personalisation in terms of the threat it poses to an employee's identity, finding that being prevented from personalising may reduce positive emotion, increase stress and lower the sense of control workers experience at work. Even where a shared environment may not, in itself, force workers to move every day (for example where there are enough desks for all employees at a given time) hot-desking may be rigorously promoted and reinforced by rules and “office etiquettes” (Allen et al., 2005; cited in Hirst, 2011) which include clear desk policies aiming to keep offices free of ‘stuff’ (Bradley and Hood, 2003; Meijer et al., 2009). With the removal of both status and group identity indicators, people have no way signal their individuality or identity at work; individuals are not able to satisfy their need for differentiation (Elsbach, 2003) which has been shown to be important for the positive emotion, commitment, and motivation of group members (Sheldon and Bettencourt, 2002).

106

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

In sum, the research described above supports our notion that shared office environments, and in particular hot desking, will be associated with a) increased demands from both distraction and workload b) negative emotion, and c) increased uncertainty and distrust. As stated previously, these factors are reliably associated a deterioration of interpersonal relationships at work. In addition, there have been findings suggesting that open plan arrangements can lead directly to a worsening in co-worker relations (Brennan et al., 2002). The literature outlined in the previous two sections informs our next two hypotheses. H3. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased distrust between employees. H4. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased negative interpersonal relationships and interactions between employees.

supervisor support, PSS) (Eisenberger et al., 2002; Kottke and Sharafinski, 1988). PSS has been found to be antecedent to POS (Eisenberger et al., 2002; Rhoades and Eisenberger, 2002) and outcomes of both PSS and POS are generally positive, including organisational commitment, job related affect, job involvement and the desire to remain (for a review of this literature see; Rhoades and Eisenberger, 2002). The majority of studies on PSS focus on the outcomes rather than the antecedents of this variable. While there is some research that suggests that open plan working arrangements result in less supervisor feedback and support (Oldham and Brass, 1979), most research in this area generally supports the notion that increased accessibility, contact, and opportunity to communicate with supervisors facilitates perceptions of support (Beehr et al., 1990; Bippus et al., 2001). This literature informs our sixth hypothesis. H6. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased supervisor support.

2.2. Resources

3. Method

2.2.1. Co-worker friendships As noted earlier, part of the rationale for the more open forms of physical office arrangement has been to increase levels of sociability and communication between co-workers. Friendly relationships with co-workers can provide both functional support and positive psychosocial and psychological outcomes. Research in this area has highlighted the positive effects of co-worker friendships (e.g., Morrison, 2004, 2009; Richer et al., 2002; Riordan and Griffeth, 1995). Positive outcomes of friendship have been found to include improved worker wellbeing (Nielsen, Jex and Adams, 2000), increased communication (Kramer, 1996), social support (Buunk et al., 1993), trust and co-operation, which has in turn been found to influence work related attitudes and behaviours (Foote, 1985; Krackhardt and Stern, 1988; Riordan and Griffeth, 1995). Proximity and propinquity have long been accepted precursors to friendship (Nahemow and Lawton, 1975) and, within organisations, these factors play a key role in both the opportunities for friendship and the prevalence of friends. Chigot (2003) argues that high-density and open plan offices facilitate more productive interactions among organisational members; for example, by improving communication flow and fostering closer working relationships. Similarly, McElroy and Morrow (2010) found that employees experienced more positive perceptions of culture and improved work-related attitudes when in an open plan environment. Nielsen et al. (2000) also found that both friendship prevalence and friendship opportunities were significantly correlated to job satisfaction, affective commitment, and turnover intention. However findings since then have highlighted the relatively greater importance of friendship opportunities (Morrison, 2006), which includes the chance to communicate and interact freely with colleagues e something generally assumed to be afforded by shared workspaces. We therefore hypothesize the following:

3.1. Sample and procedure

H5. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report more collegial friendships.

2.2.2. Supervisor support Workers will have perceptions of the extent to which they are supported and valued by their organisation (perceived organisational support, POS) (Eisenberger et al., 1986) and will similarly develop general views concerning the degree to which their supervisors value them and care about their well-being (perceived

A total of 1000 Australian participants in permanent employment, aged 18 and over, were recruited in 2014 using a Qualtrics survey panel of workers who had voluntarily agreed to participate in research studies. Participants were told their responses would be kept confidential and were compensated for their time, either by cash incentives or redeemable points which can be used towards flights, discounted hotel stays, or products. Qualtrics keeps its pricing strategy proprietary (Brandon et al., 2013) so precise compensation estimates are not available. While responses were anonymous to the researchers, the means by which panel respondents are tracked by the Qualtrix software means that multiple submissions by a single person are not possible and each response is unique. The respondent characteristics are summarised in Table 1. Gender was fairly evenly represented among the respondents and most were employed full time. The average age of the respondents was just under 47 and, while the standard deviation appears high relative to the mean, an examination of the age distribution shows

Table 1 Respondent characteristics. Sample Characteristic

N ¼ 1000

Male Female Full-time employed Part-time employed

45.0% 55.0% 64.7% 35.3%

Usually works with others in a team Usually works independent of others Works sometimes in a team or on own

40.4% 24.4% 35.2%

Has own office Shares office with one or two others Works in open-plan office with own workstation Works in open-plan office and hot-desks Works mainly at clients in whatever space given Works mainly at home or on-the-road Other arrangement

16.0% 11.2% 40.2% 9.2% 9.9% 9.3% 4.2%

Age in Years Years tenure with current employer Years in current role Years in current career

Mean ¼ 46.84 SD ¼ 12.67 Mdn ¼ 6.29 range 0.8e50 Mdn ¼ 5.17 range 0.8e50 Mdn ¼ 12.50 range 0.8e56

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

it approximates a normal distribution bounded with a range from 18 to 77 years. However, as is typical for such variables, years with current employer, in current role and current career are positively skewed. A diverse range of 695 distinct occupations were stated by respondents, with no single occupation dominating. The largest single occupational cluster was administration roles (n ¼ 50), followed by sales (n ¼ 38), customer service (n ¼ 30), and manager (unspecified) (n ¼ 22). In total, 76.6% of respondents indicated they worked in an office environment (see Table 1), whether that be a personal office, a shared office, open plan with workstation, or hot desking. In the analyses that follow, those reporting “other” (4.2%) were excluded. Respondents reporting that they worked mainly at a clients' in whatever space available to them, or who worked at home or on the road (19.2% in total) were retained; being included in the analysis of the various work environments (see Fig. 1). 3.2. Measures In addition to demographics, the online questionnaire assessed the extent to which participants shared office workspaces (they responded to items 1. I mainly work at home or on-the-road, 2. I have my own office, 3. I share an office with one or two other people, 4. I work mostly at clients' workplaces in whatever space they give me, 5. I work in an open plan office space but have my own desk/work space/ work station, 6. I work in an open plan office and hot-desk; first in first served). They were also asked the extent to which they worked independently (participants responded to items 1. I usually work with others in a team environment, 2. I sometimes work in a team and sometimes on my own, 3. I usually work independently of others on my own tasks). To create the measures of the variables related to our four demand hypotheses (distrust, uncooperative behaviour, distractions, and negative relationships) and two resource hypotheses (co-worker friendship and supervisor support) we followed practices outlined by prior researchers and scholars who have discussed the process of developing and validating measures of constructs (e.g., Rich et al., 2010; Schwab, 1980) and used a sequential, three phased, mixedmethod process. In brief, the first phase generated an initial pool of 54 potential survey questionnaire items reflective of an employee's social demands and resources in their workplace. This included an analysis of the prior literature on social capital at the individual level, including social support (Cohen and Wills, 1985; Hill et al., 1989; Sarason et al., 1987), interpersonal trust and reciprocity (Cummings and Bromiley, 1996; Kouvonen et al., 2008; McAllister, 1995; Sapp et al., 2010; Suzuki et al., 2010), relationship quality and satisfaction (Bridge and Baxter, 1992; Crosby et al., 1990; White et al., 2012), and workplace friendliness (Morrison, 2004; Nielsen, Jex and Adams, 2000). In addition, aspects of negative interpersonal behaviour at work were examined, including undermining (Duffy et al., 2002; Duffy et al., 2006; Duffy et al., 2012), bullying (Einarsen, 1999; Einarsen and Skogstad, 1996), jealousy and envy (Vecchio, 2000), and workplace deviance (Bennett and Robinson, 2000). In addition, two focus group sessions, one of MBA students and the other comprising 28 professionals in five discussion groups, were used to generate additional items. The second phase, conducted within a month, involved 32 participants and included workplace researchers and work experienced employees recruited using opportunity sampling. The main selection requirement was that participants had a workplace social network from which they might acquire social capital/benefit and or liabilities/demands. Using an online Q-sort activity (where participants are asked to combine statements into meaningful clusters), they sorted the 54 potential survey questionnaire items into pre-labelled categories: two relating to the resources variables

107

and the remaining four relating to the demand variables. Another category was left blank for those items participants were unable to categorize. Q-sorts, are a useful qualitative method to test if statements previously generated represent a construct (Grey, 2014). In addition, to identify items unlikely to have sufficient variance to differentiate between people, respondents were asked to indicate the extent to which they agreed or disagreed that each item described their own workplace social network using a 7-point Likert-scale, scored 1 (Strongly Disagree) to 7 (Strongly Agree). For an item to be retained, more than 50% of the respondents needed to agree an item belonged to a particular Q-sort category. In addition, items were dropped for having insufficient variance if more than 80% disagreed or strongly disagreed that the item described their workplace social network. From these analyses, 42 items were retained for use in the online Qualtrics survey questionnaire (measuring constructs relating to our six hypotheses; co-worker friendship, supervisor support and the four social demand components). In the third phase of developing the measures for this study, responses from the 1000 respondents to the online survey were analysed using the principal axis factor analysis feature available in IBM SPSS Statistics (V22). Because extant literature suggests that the four aspects of social demand are positively related to each other, the direct oblimin method of oblique rotation was used, allowing the factors to correlate. As Table 2 shows, the rotated pattern matrix has six clear factors, corresponding to the two resource and four demand variables, and explaining a total of 61.94% of the variance (a higher loading in Table 2 reflects an item which has more importance within a given factor, thereby allowing us to cluster items according to construct similarity). As is normative for factor analysis, a minimum factor loading of 0.4 was used to determine if an item was retained on a given factor. Table 2 also shows the wording of all demand and resource items used, with those in italics being the ones dropped from subsequent analyses due to poor factor loadings. Responses to all items were obtained on a 5-point Likert-type scale coded from 1 (Strongly Disagree) to 5 (Strongly Agree). As Table 2 shows Factor 1 contained 10 items with a loading above 0.4 that form the demand variable negative relationships. All corrected item-total correlations (which indicate how well an item measures a given construct) were above 0.6 (range 0.68e0.87) and the coefficient alpha of the scale was 0.95 (a high alpha provides further evidence that all items measure the same construct). Factor 4 has three items with factor loadings above 0.4 which relate to the demand variable of distrust. Coefficient alpha was 0.82 and itemtotal correlations ranged from 0.58 to 0.74. Factor 5 contains six items with factor loadings above 0.4 that all relate to the notion that others in a workplace can serve as distractions from work and put demands on us. These six items measure the demand variable, distractions (coefficient alpha ¼ 0.84; item-total correlations ranging from 0.53 to 0.72). The final employee demand variable, uncooperative behaviour, is measured by the six items loading above 0.4 on Factor 6 (see Table 2). Coefficient alpha for the six item measure was 0.90 and the inter-item correlation ranged from 0.68 to 0.80, indicating strong internal item consistency. The remaining two factors in Table 2 deal with the resource variables. Factor 2 contains 7 items loading at 0.4 or above that relate to having co-worker friendships in the workplace (coefficient alpha ¼ 0.89; inter-item correlation range ¼ 0.61 to 0.78). And finally, Factor 3 contains five items measuring the degree to which an employee perceives that they have supervisory support at work (coefficient alpha ¼ 0.95; inter-item correlation range ¼ 0.80 to 0.90). Because the data for all variables were collected at the same

108

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

Fig. 1. Mean score as a function of shared work environment.

time using the same method, there is potential for the methodological artefact of common method variance (CMV) to exist. To

reduce the possibility of CMV we followed the advice of Conway and Lance (2010) by including multiple reverse-phrased items in

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

109

Table 2 Principal axis factor analysis with oblique rotation pattern matrix. With reference to your place of work and the relationships you have with others who work there, please indicate the extent to Factor which you agree or disagree with the statements below. 1 2 Some people I work with have insulted me Some people I work with have talked down to me Some people I work with made me feel incompetent Some people I work with have hurt my feelings Some people I work with have undermined my efforts to be successful on the job Some people I work with criticize the way I handle things on the job in a way that is not helpful People at work have spread of gossip and rumours about me Some people I work with try to make me look bad in front of others Some people I work with have not given as much help as they promised People I work with succeed by stepping on other people Some co-workers are not above “bending the facts” to create the impression they want

3

0.918 0.918 0.902 0.879 0.804 0.658 0.636 0.506 0.435 0.422 0.388

I have formed strong friendships at work I socialise with co-workers outside the workplace I do not feel that anyone I work with is a true friend (R) The people I work with are friends as well as co-workers I can count on my workmates to really care about me, regardless of what is happening I can talk freely to others at work about difficulties I am having and know that they will want to listen I can count on my workmates to help me feel more relaxed when I am under pressure or tense Things would be a lot easier if people were only friends or only work associates instead of trying to be both (R)

4

5

6

0.338 0.838 0.753 0.734 0.710 0.450 0.401 0.400 0.372

My manager treats me with kindness and consideration My managers shows concern for my rights as an employee I can trust my supervisor In our interactions, my supervisor considers my feelings We have a 'we are together' attitude

-0.331 -0.351 0.983 0.961 0.856 0.827 0.619

People I work with cannot be trusted to do as they say People I work with are mostly just looking out for themselves I can't be too careful in dealing with the people I work with I suspect that people I work with are actively working against me (for example sabotage, withholding information)

-0.304 0.564 0.550 0.414 0.374

My workmates sometimes take up time I would rather spend on my job My friends at work often keep me from my job requirements I often wish I could get on with my work without having to interact so much with my peers I am often distracted by others at work Some of the people I work with expect me to prioritize our friendship relationship over my work commitments Some of the people I work with are very needy, demanding a great deal of my attention It requires extra effort to maintain both the friendship side and the work side of my relationships in this organisation

-0.858 -0.774 -0.644 -0.534 -0.517 -0.481 -0.365

Most people I work with cooperate with each other (R) Most of the time, people I work with try to be helpful (R) Most people I work with can be trusted (R) My co-workers and I assist each other in accomplishing assigned tasks (R) I am able to work with my co-workers to collectively solve problems (R) People keep each other informed about work-related issues in the work unit (R) R My co-workers can be relied upon to keep their promises (R)

0.320

-0.364

0.757 0.706 0.585 0.544 0.533 0.432 0.363

Notes: For clarity, factor loadings below 0.3 are not shown. (R) ¼ reverse scored items. Italicised items were dropped from further analysis.

order to the reduce response bias due to inattention and acquiescence (Podsakoff et al., 2012). We also made use of a Qualtrix feature for online questionnaires that includes random questions requiring participants to make a specified response. This reduces the potential for participants to endorse all items using the same

response option. Such procedural steps go some way to obviating common-method concerns. In addition, a Harmon's single factor test for common method variance was performed, finding a forced factor containing all the scale items used in the current study accounted for 25.1% of the variance; well short of indicating the

Table 3 Variable correlations and descriptive statistics. Variables

M

SD

1

2

3

4

5

6

7

8

1 2 3 4 5 6 7 8 9

3.10 3.32 3.43 3.18 2.52 4.60 4.96 e 46.84

1.12 1.53 1.44 1.19 1.01 1.25 1.48 e 12.58

0.94 0.94 0.81 0.76 0.77 -0.54 -0.67 0.00 -0.16

0.95 0.68 0.61 0.63 -0.48 -0.61 0.02 -0.14

0.82 0.55 0.68 -0.50 -0.57 0.00 -0.13

0.84 0.43 -0.25 -0.38 -0.03 -0.18

0.90 -0.65 -0.71 -0.00 -0.08

0.89 0.57 0.10 -0.03

0.95 0.07 0.01

-0.15

Social liabilities (aggregate) Negative relationships Distrust Distractions Uncooperative behaviour Co-worker friendships Supervisor support Gender (1 F; 0 M) Age in Years

Note: All correlations > 0.07 significant at p < 0.01 (1-tailed); correlations ¼ 0.07 significant at p < 0.05 (1-tailed). High scores indicate increased social demands and/or increased support and friendship. N ¼ 958. Coefficient alphas are shown in bold on the diagonal.

0.116 2.86 3.09 0.009 0.105 2.50 3.68 0.003 0.122 2.69 3.73 0.002 0.150 3.31 3.04 0.010 0.159 4.73 2.90 0.013 93

Notes: N ¼ 958. High scores indicate increased demands and/or increased support and friendship.

3.04 2.62 0.023 0.154 5.05 4.42 0.001

4.29 4.53 92 99

0.131

0.115 0.111 3.38 3.15 0.105 0.101 2.77 2.70 0.121 0.117 3.25 3.15 0.149 0.143 3.79 3.61 0.158 0.152 3.70 3.28 0.153 0.147 4.52 4.64

4.73 4.86 4.53 160 112 402

I have my own office I share an office with one or two others I work in an open plan office but have my own workspace I work in an open plan office and hotdesk I work mostly at clients workplaces in whatever space they give me I mainly work at home or on the road F (5,950) p

0.130 0.125

2.99 2.95 3.18 0.079 0.094 0.050 2.33 2.33 2.55 0.092 0.109 0.058 3.21 3.15 3.27 0.113 0.134 0.071 3.26 3.14 3.48 0.120 0.143 0.076 3.18 3.13 3.41 0.116 0.138 0.073 5.16 5.30 4.96

SE M M SE M SE M M M

SE

Supervisor support Co-worker friendships N Working arrangement

Given the relationships found between the demand and resource variables, and for age and gender (see Table 3), the multivariate analysis of covariance (MANCOVA) function available in IBM SPSS (V22) was used to test the hypotheses concerning differences between office environments in the dependent variables, while controlling for gender and age as covariates. Initial multivariate tests were statistically significant, justifying the inclusion of the shared office space independent variable (trace ¼ 0.069, p ¼ 0.000) with gender (trace ¼ 0.028, p ¼ 0.000) and age (trace ¼ 0.043, p ¼ 0.000) control variables in the full model. Table 4 reports the estimated marginal means (controlling for age and gender) for each of the demand and resource variables as a function of the different office working environments, together with the univariate between-subjects F-tests for these means. Bonferroni corrections were applied to all significance levels to reduce the potential for Type I errors arising from multiple statistical tests. In addition, Levene's tests of the equality of the error variances were not significant indicating that this MANCOVA assumption has been met. Significant multivariate effects were found for all four demand variables (see Table 4). To test the hypotheses more directly, multivariate least significant difference (LSD) contrast tests on the marginal means were performed. Fig. 1 shows the relationship between shared work environments and social liabilities (see Table 4 for means and standard deviations). The figure demonstrates that these demands significantly increase as work environments are more “shared” (the lower five lines, including the full set of social liability items, which combines the four demand sub-constructs). Hypotheses 1 to 4 are supported. Other than working at home or on the road, the work environment with the lowest level of demands in the form of distrust, distractions, uncooperative behaviours and negative interactions was found to be having one's

Table 4 Demand and resource estimated marginal means and standard errors by office working arrangements.

4.1. Office environments and social demands

SE

Negative interactions and behaviours

Distrust of others

Distractions

SE

Uncooperative behaviour

Table 3 shows means, standard deviations and Cronbach's coefficient alpha for each of the variables used to test H1eH6, together with their inter correlations. In each case scale scores were the averages of the items, with higher scores indicating higher levels of demands and resources. In addition a combined demands variable, that we have termed “social liabilities” and including distrust, uncooperative behaviour, distraction and negative relationships (coefficient alpha ¼ 0.94) was analysed in relation to shared work environments. As Table 3 shows, all four demand variables are significantly correlated with each other, although not to a level to suggest multicollinearity. They are also negatively associated with the resource variables, such that those with higher levels of supervisor support and co-worker friendships are also more likely to report lower levels of distraction, negative relationships, uncooperative behaviour and distrust in others. Gender is unrelated to any of the demand variables, while weak correlations are found for the resource variables such that female respondents are slightly more likely to report higher supervisor support and more co-worker friendships. Conversely, weak correlations are found for age with the four demand variables but not the two resource ones, with older respondents slightly more likely to report lower levels of distrust, negative relationships, uncooperative behaviour and social distractions at work.

SE

4. Results and discussion

M

Combined social liability items

presence of common method variance bias in these data.

0.087 0.104 0.055

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

0.098 0.117 0.062

110

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

own office, followed by sharing an office with one or two others. The working environments with the highest levels of demands were open plan ones, with hot-desking having the most demands (see Fig. 1 and Table 4). Each hypothesis related to these social liabilities is addressed briefly below: H1. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report higher levels of distractions from work. H1 was supported with distraction being significantly worse in all the shared office arrangements when compared to those working at home or on the road. Logically these remote workers have fewer opportunities to interact with colleagues in general, and so were removed from the potential distraction from others in their organisation. H2. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report higher levels of uncooperative behaviour from colleagues. H2 was supported. For uncooperative behaviour, those in open plan offices, either hot-desking or with their own workstations, report higher levels of uncooperative behaviours than those with their own offices or who share an office with one or two others (hot-desking and working in an open plan office with an assigned workstation were not significantly different from each other for this variable). Those working on client premises report similar levels of uncooperative behaviour to open plan office environments. H3. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased distrust between employees. H3 was supported. For distrust, the hot-desking condition was found to be significantly worse than either having one's own office or sharing with one or two others. As with uncooperative behaviour, hot-desking and working in an open plan office with an assigned workstation were not found to differ significantly in terms of distrust levels. Furthermore, those working on client premises report similar levels of distrust to those who work in open plan offices. H4. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased negative interpersonal relationships between employees. H4 was supported. Those in a hot-desking arrangement reported significantly more negative relationships than those in their own office, those sharing an office with one or two others, those working on client's premises or at home. Again there was no significant difference between hot-desking and working in an open plan office with one's own workstation on this variable. For the combined social liability measure, SPSS post hoc comparisons, using Bonferroni corrections to reduce the potential for Type I errors, found those who hot-desk and those in open plan offices with their own workstations to have significantly higher social liabilities when compared to those with their own offices or who share their offices with one or two others. Furthermore, those working at home or on the road report significantly fewer social liabilities. Thus, consistent with the JD-R model, when in shared workspaces and/or hot-desking environments, the four liability variables, both individually and as a combined measure, are likely to place additional demands and increased load on workers by creating an unfavourable working environment thereby incurring physiological and/or psychological “costs” (Bakker and Demerouti, 2007).

111

4.2. Shared office environments and social resources Table 4 shows an overall significant difference in both coworkers friendships and supervisor support as a function of office working arrangements. Fig. 1 demonstrates that supervisor support and co-worker friendships generally decrease as work environments are more shared (the top two lines in Fig. 1). Each is discussed below. H5. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report more co-worker friendships. Pairwise LSD post hoc comparisons show that co-worker friendships are actually lower in hot desking and open plan office arrangements, when compared to those with their own offices or who share offices. They are even significantly lower than those who mainly work at home or on the road. Hypotheses 5 is therefore rejected. It would seem that being in a shared environment does not facilitate friendships at work. H6. Those in shared office environments (with hot-desking being at the extreme end of the continuum) will report increased supervisor support. There was no support for the sixth hypothesis. Fig. 1 demonstrates that supervisor support decreases as work environments become more shared and the pairwise LSD comparisons between the marginal means supports this. Those in hot-desking office environments reported statistically lower support compared to those who share offices and who work on client premises. This is in the direction opposite to what was hypothesised. It seems that shared work environments are associated with lower quality supervision/ manager relationships. The finding that perceptions of supervisor support worsen in shared environments may be related to the notion that, when workers do not see their supervisors every day, the time they do have with them is perceived to be of higher quality. In this context, they may have a supervision meeting with focused attention rather than simply working nearby their manager on a daily basis. It is also possible that many of the factors related to social demands at work also negatively impact on supervision relationships in shared work environments, making them less satisfactory. The employee, the supervisor, or both, may be irritated, distracted and attempting to combat this by withdrawing (Kaarlela-Tuomaala et al., 2009), thereby worsening the supervisory relationship. In relation to the JD-R model it seems that shared work environments, and in particular hot-desking increase demands without buffering by variables widely accepted as resources. Shared environments did not improve friendship opportunities and, in addition, were associated with perceptions of less supportive supervision.

4.3. Collaborative working In terms of working collaboratively with others, no significant difference was found for the combined social liability score between those who worked independently, in a team or both (F (2,997) ¼ 0.06, p ¼ 0.95). However, as discussed above, significant differences were found as a function of the physical working arrangements participants were in. This implies that it is not working closely with others per se. that has a negative impact. When colleagues work closely with others in a team, on projects, or towards common goals (for example) we found no evidence of increased negative social impacts; it seems only when workers are hot-

112

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

desking or simply working nearby (as opposed to “with”) others that these demands are increased. Perhaps, as Kaarlela-Tuomaala et al. (2009) suggest, it is possible that we would have different results if the respondents in the current study needed to continuously interact to do their jobs. In terms of the resources we measured, there was no significant difference in perceptions of supervisor support between those who worked independently, in a team or both (F (2, 997) ¼ 2.77, p ¼ 0.63) but there were significant findings in terms of co-worker friendships (F (2, 997) ¼ 9.47, p ¼ 0.00) indicating that those who worked closely with others (either usually or sometimes) reported better co-worker friendship than those who usually worked independently. In contrast, as Table 4 and Fig. 1 shows, co-worker friendship did not vary significantly across physical working arrangement. This suggests that working collaboratively in team environment facilitates co-worker friendship whereas simply doing one's own work nearby others in a shared environment does not.

4.4. Limitations and directions for future research One potential limitation of the current study is the use of data from a survey panel (in this case Qualtrics), however the increasing use and external validity of data obtained from survey panels such as Qualtrics and SurveyMonkey has received some research attention; a recent review (Brandon et al., 2013) finding “…that participants who are recruited using non-traditional methods generate data that are as externally valid as those provided by more traditional participant recruitment methods.” (p.21). Another limitation of the current study is that we did not gather data specifically about the proportion of their work day respondents spent in an office environment. Future work in this area might explore this in order to evaluate the effects of relative exposure to these potentially adverse conditions. In addition, although we measured both demands and resources as they relate to shared work environments, we did not measure employee outcomes of these, such as stress, directly. A direction for future research would be to look at whether workplace social liability variables mediate or moderate associations between shared environments and stress. The finding that having one's own office, or sharing with just one or two others, is associated with significantly improved perceptions of supervisor support adds to the relatively sparse literature on antecedents of perceived supervisor support. Future research could examine whether having the physical distance/autonomy that offices provide improves perceived supervisor support for all workers and perhaps explore further why this is so. Another direction for future research could be to explore ways to further transform office design; to retain the flexibility and cost saving of flexible work arrangements while avoiding the negative social, psychological, and physical outcomes. Some organisations are already exploring ways to make this work. For example the Catholic Commission for Employment Relations has adopted a “post hot-desking” office design; moving on from their hot desk arrangements to embrace “zones” including a collaboration zone  style seating, where people can sit informally, and with cafe communicate freely. These exist alongside task zones with ergonomically designed chairs and desks. In addition there is a “cockpit” (a small separate office) available for workers who need to be isolated and uninterrupted. The creation of these zones around the office, designed specifically for particular work gives employees the autonomy to move between them, depending on the specific task they are performing (Patty, 2015).

5. Conclusion and recommendations As a result of the changing face of education in modern schools worldwide (such as innovative student groupings, blended learning environments and flexible work spaces in class rooms) (see for example; Giles, 2006; Punie, 2007; Reh et al., 2011; Rytivaara, 2011; Strayer, 2012; Thompson, 2015), there may well be a generation of “new” employees (those yet to enter the workforce) who will be used to working with many different colleagues and moving from space to space as their task, or as space constraints, dictate. These future workers may have little or no expectation that their employer will provide either privacy, or their own designated space. Whether or not this is true in the future however, it seems that for now, shared work environments, and in particular hot-desking, are associated with increases in distraction, negative relationships, uncooperative behaviours and distrust. This may be due to amplified demands in terms of workload in hot-desk environments (i.e., finding and personalising a space), in addition to the uncertainly associated with not having an assigned workstation, or the increased load from distraction in a shared space. Withdrawal as a response to this load may account for the uncooperative behaviours and lack of reciprocity reported. Though we did not measure stress directly in the current study, we found that there were increases in variables widely accepted as “demands” without any offset or buffering by variables widely accepted as “resources”. Shared environments did not improve friendship opportunities/supervisor support (and in fact were associated with perceptions of less supportive supervision). In almost all our measured variables the “best” case was having one's own office or sharing with one or two others. In addition to the negative social impacts of shared work spaces, our results align with those of Kaarlela-Tuomaala et al. (2009) who found that the benefits which are often claimed to be associated with shared spaces were not apparent, finding instead that cooperation became less pleasant and information flow did not change. In sum, we agree with Kaarlela-Tuomaala et al.’s conclusion, that “…the open-plan office is not recommended for professional workers.” (p. 1423). However, we acknowledge that the inexorable move towards shared office environments and hot-desking, along with the perceived flexibility and genuine cost saving they represent, is unlikely to be reversed. Though perhaps, in time, and with more research in this area, there will increased consideration of these policies in the context of worker well-being. 5.1. Ergonomic principles Prior to a discussion of what might be done to ameliorate any unintended negative consequences of these work environments, it is necessary to also consider the work done by those experiencing adverse effects; functional recommendations cannot be made without a proper assessment of the roles and demands of their jobs. A systematic approach within the ergonomic principle of assessment is needed. As it is outside the scope of the current study, we have yet to ascertain if there are some roles, jobs, professions, or industries in which the adverse effects of shared space are greater; for example, are there particular aspects of work that make some employees more vulnerable? Similarly, the inclusion of individual differences should to be a focus of future work (Caple, 2008); for example, does personality, gender or age mean some workers are comparatively more vulnerable? Instead, the focus of the current study has been on an aspect of the physical environment within which work is performed e the office e rather than on the nature of that work, or the people doing the work. Thus, future research could deepen our understanding of

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

these issues by employing more focused surveys, engaging in ergonomic assessments of specific organisations, analysing the tasks and work processes of affected individuals, and capturing rich, qualitative data using focus groups and interviews. Only then might we have confidence to select from the various types of interventions and apply them to a given situation. A “design centred approach” (Caple, 2008) where user needs are central, might suggest that office space be designed, not for cost and space saving efficiencies, but to reflect the well-being of the user (in this case the office worker). An ideal situation would be to evaluate the worker, analyse the work, and create an “ideal” workspace thereby obviating the need to minimize demands later on (Hale et al., 2007). The “hierarchy of controls” should influence decisions regarding which interventions or changes are most appropriate. The most desirable levels, “eliminating” or “isolating”, would imply a situation where distraction or interaction from nearby others is removed altogether, or was never there in the first place. As we have shown in the current study, having one's own office or working from home is indeed the “best” situation in terms of the demands we measured. However, given that this is not realistic for many organisations, the recommendations below are aimed at “minimizing” the demands. 5.2. If you're going to do it, do it right A range of interventions and measures to ameliorate possible negative consequences of shared office space have been proposed in recent years. These could be considered once an assessment of specific work environments, individuals, and work processes has taken place to provide a detailed understanding of the work context. First, clear desk requirements and banning the personalisation of work spaces are two policies that have little obvious benefit, but potentially quite negative consequences for workers (Brown, 2009; Elsbach, 2003). These policies could be re-thought wherever possible, as it seems that the personalisation of workspaces fulfils some quite fundamental human needs. Second, visual distraction from nearby co-workers can be dramatically reduced with the use of panels or book shelves (Smith-Jackson and Klein, 2009) or “green” walls of plants, termed “living wall systems” (LWS) (Perini and Rosasco, 2013). Researchers investigating the benefits of LWS struggle to directly measure the psycho-social benefit of LWS due to the “…impossibility to assign a value to the positive impact of vegetation.” (Perini and Rosasco, 2013, p. 118). That said, both aesthetic improvement and sound reduction are commonly cited benefits of LWS (Rakhshandehroo et al., 2015). Third, auditory distraction in shared environments may be reduced with noise reduction equipment such as noise cancelling headphones; shown to reduce auditory interference (Smith-Jackson and Klein, 2009). Both visual screens and noise reducing equipment are relatively low-cost, flexible ways to improve the IEQ for office workers. It is worth noting, however, that interventions such as headphones that put the onus onto individuals to manage their own IEQ, and other associated demands, are not as appropriate as wider policy, practice, and office designs aimed at improving the environment in general. Organisations should evaluate technologies and equipment within the context of the whole workstation, the type of work being done, and individual preferences. Thus, a contingency approach is needed with the expectation that provision of headphones (for example) will suit some organisational members but may be unacceptable, or even more demanding, to others. Case studies of hot-desk and/or flexible workspaces designed with worker well-being in mind can also provide options regarding both physical workspace options and ways of working. For example Pitt and Bennett (2008) describe a large office redesigned to include

113

not only hot-desking, but also “touchdown areas” (unassigned desks which allow for quick access to information), “bookable offices” (spaces that provide privacy that can be booked in advance), “collaborative workspaces” (formal rooms to facilitate group work which include teleconferencing capabilities) and finally “break-out workspaces” (soft-seating and coffee tables arranged to allow for spontaneous, informal collaborative work). This last initiative relates to the notion that there needs to be a “critical density of spontaneous interaction” (Haynes and Price, 2004, p.11) for this type of activity-based workplace initiative to succeed. Too much and the distractions will outweigh any potential collaborative benefits. Too little and the benefits are not evident. Industry, social/societal context, worker individual differences, sector, and type of work will all no doubt play a role in what that critical density will be for a given workplace or organisation. Fulton (2005) outline other best practices in terms of hotdesking arrangements. Relevant to the current study, they suggest providing some sort of “home base” for hot-deskers as a way to ensure they feel included, and to ameliorate feelings of being displaced, or stress from seeking a work station. This could take the form of an inviting lounge or common room with lockers to store equipment or hard copies of documents. In addition there should be information technology initiatives to ensure ease of transitioning from one workstation to another; this could include seamless Wi-Fi when computers are undocked, computers and software with ‘plug and play’ capability, access to electronic networks and intranets, and timely access to paper-based information (Fulton, 2005). The reduction in perceived supervisor support in shared environments perhaps indicates that the assumption that simply being nearby one's supervisor obviates the need for formal feedback or supervision meetings is misguided. Alternatively it might suggest that seeing one's manager too frequently puts extra strain on the supervisory relationship in some way. A recommendation may be to have time for dedicated supervision or team meetings set aside, possibly along with screens and headphones for privacy within the supervision relationship, as above. Ultimately, providing options for workers to suit both individual preferences and the type of work being carried out, is likely to be the key to successful implementation of shared office environments and hot-desking in situations where the provision of private offices is deemed inappropriate or too expensive. Thus, ergonomists and others tasked with the design and implementation of such systems, should consider user-testing of the work environment and employee participation in workspace design as part of a user-centred approach to the introduction of shared workspaces (Cullen, 2007). In addition, should changes be made, or recommendations applied to improve workers' experiences, these must to be trialled, monitored and evaluated. It is not enough to simply make changes aimed at improving the IEQ in a shared workspace. Without ongoing assessment, it is impossible to know whether these changes are the right ones, done in the correct way, or to the right extent. Further, as the workforce evolves and as work processes change, ongoing, evidence based, assessment of the efficacy of interventions should continue, encouraging employee-participation and feedback.

Acknowledgements This research was made possible with a grant from the Work Research Institute (WRI) within the Faculty of Business and Law, AUT University, NZ.

114

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115

References Andersson, L.M., Pearson, C.M., 1999. Tit for tat? The spiraling effect of incivility in the workplace. Acad. Manag. Rev. 24 (3), 452e471. http://dx.doi.org/10.5465/ AMR.1999.2202131. Appelbaum, E., Bailey, T., Berg, P., Kalleberg, A., 2000. Manufacturing Advantage: Why High-performance Systems Pay off. ILR Press, Ithaca: NY. Ashkanasy, N.M., Ayoko, O.B., Jehn, K.A., 2014. Understanding the physical environment of work and employee behavior: an affective events perspective. J. Organ. Behav. 35, 1169e1184. http://dx.doi.org/10.1002/job.1973. Bakker, A.B., Demerouti, E., 2007. The job demands-resources model: state of the art. J. Manag. Psychol. (22), 309e328. http://dx.doi.org/10.1108/ 02683940710733115. Bakker, A.B., Demerouti, E., Euwema, M.C., 2005. Job resources buffer the impact of job demands on burnout. J. Occup. Health Psychol. 10, 170e180. Bakker, A.B., Demerouti, E., Taris, T., Schaufeli, W.B., Schreurs, P., 2003. A multigroup analysis of the Job Demands-Resources model in four home care organizations. Int. J. Stress Manag. 10, 16e38. Beehr, T.A., King, L.A., King, D.W., 1990. Social support and occupational stress: talking to supervisors. J. Vocat. Behav. 36 (1), 61e81. Bennett, R.J., Robinson, S.L., 2000. Development of a measure of workplace deviance. J. Appl. Psychol. 85, 349e360. http://dx.doi.org/10.1037/00219010.85.3.349. Bentley, T.A., Teo, S.T., McLeod, L., Bosua, R., Gloet, M., Tan, F., 2016. The role of organisational support in teleworker wellbeing: a socio-technical systems approach. Appl. Ergon. 52, 207e215. Bippus, A.M., Brooks, C.F., Plax, T.G., Kearney, P., 2001. Students' perceptions of parttime and tenured/tenure-track faculty: accessibility, mentoring, and extra-class communication. J. Assoc. Commun. Adm. 30, 13e23. Boxall, P., Purcell, J., 2011. Strategy and Human Resource Management, third ed. Palgrave Macmillan, New York. Bradley, S., Hood, C., 2003. Delivering minimalist workplaces that improve corporate agility. J. Facil. Manag. 2 (1), 68e84. Brand, J.L., & Smith, T.J. (2005). Effects of reducing enclosure on perceptions of occupancy quality, job satisfaction, and job performance in open-plan offices. presented at the meeting of the Human Factors and Ergonomics Society 49th Annual Meeting. Brandon, D.M., Long, J.H., Loraas, T.M., Mueller-Phillips, J., Vansant, B., 2013. Online instrument delivery and participant recruitment services: emerging opportunities for behavioral accounting research. Behav. Res. Account. 26 (1), 1e23. Brennan, A., Chugh, J.S., Kline, T.J.B., 2002. Traditional versus open office design: a longitudinal field study. Environ. Behav. 34 (3), 279e299. Bridge, K., Baxter, L.A., 1992. Blended relationships: friends as work associates. West. J. Commun. 56 (3), 200e225. http://dx.doi.org/10.1080/ 10570319209374414. Brown, G., 2009. Claiming a corner at work: measuring employee territoriality in their workspaces. J. Environ. Psychol. 29 (1), 44e52. Buunk, B.P., Doosje, B.J., Liesbeth, G., Jans, J.M., Hopstaken, L.E.M., 1993. Perceived reciprocity, social support, and stress at work: the role of exchange and communal orientation. J. Pers. Soc. Psychol. (65), 801e811. Caple, D., 2008. Emerging challenges to the ergonomics domain. Ergonomics 51 (1), 49e54. http://dx.doi.org/10.1080/00140130701800985. Chigot, P., 2003. Controlled transparency in workplace design: balancing visual and acoustic interaction in office environments. J. Facil. Manag. 2, 121e130. Cohen, S., Wills, T.A., 1985. Stress, social support, and the buffering hypothesis. Psychol. Bull. 98 (2), 310e357. http://dx.doi.org/10.1037/0033-2909.98.2.310. Cole, R.J., Bild, A., Oliver, A., 2012. The changing context of knowledge-based work: consequences for comfort, satisfaction and productivity. Intell. Build. Int. 4 (3), 182e196. http://dx.doi.org/10.1080/17508975.2012.695950. Conway, J.M., Lance, C.E., 2010. What reviewers should expect from authors regarding common method bias in organizational research. J. Bus. Psychol. 25 (3), 325e334. http://dx.doi.org/10.1007/s10869-010-9181-6. Cooper, R., Boyko, C.R., 2009. The effect of the physical environment on mental wellbeing. In: Cooper, G.L., Field, J., Goswami, U., Jenkins, R., Sahakian, B.J. (Eds.), Mental Capital and Wellbeing, pp. 967e1006. Crosby, L.A., Evans, K.R., Cowles, D., 1990. Relationship quality in services selling: an interpersonal influence perspective. J. Mark. 54 (3), 68e81. http://dx.doi.org/ 10.2307/1251817. Cullen, L., 2007. Human factors integration e bridging the gap between system designers and end-users: a case study. Saf. Sci. 45, 621e629. http://dx.doi.org/ 10.1016/j.ssci.2007.01.002. Cullinane, S.-J., Bosak, J., Flood, Demerouti, 2013. Job design under lean manufacturing and its impact on employee outcomes. Organ. Psychol. Rev. 3, 41e61. Cummings, L.L., Bromiley, P., 1996. The organizational trust inventory (OTI): development and validation. In: Kramer, R.M., Tyler, T.R. (Eds.), Trust in Organizations. Sage, Thousand Oaks, CA, pp. 302e330. Davis, L.E., Cherns, A.B., 1975. The Quality of Working Life: Problems, Prospects, and the State of the Art. Free Press, New York. Demerouti, E., Bakker, A.B., Nachreiner, F., Schaufeli, W.B., 2001. The job demands e resources model of burnout. J. Appl. Psychol. 86, 499e512. Duffy, F., 1997. The New Office (London). Duffy, M.K., Ganster, D.C., Pagon, M., 2002. Social undermining in the workplace. Acad. Manag. J. 45 (2), 331e351. http://dx.doi.org/10.2307/3069350.

Duffy, M.K., Ganster, D.C., Shaw, J.D., Johnson, J.L., Pagon, M., 2006. The social context of undermining behavior at work. Organ. Behav. Hum. Decis. Process. 101, 105e121. Duffy, M.K., Scott, K.L., Shaw, J.D., Tepper, B.J., Aquino, K., 2012. A social context model of envy and social undermining. Acad. Manag. J. 55 (3), 643e666. http:// dx.doi.org/10.1016/j.obhdp.2006.04.005. Einarsen, S., 1999. The nature and causes of bullying at work. Int. J. Manpow. 20 (1/ 2), 16e27. http://dx.doi.org/10.1108/01437729910268588. Einarsen, S., Skogstad, A., 1996. Bullying at work: epidemiological findings in public and private organizations. Eur. J. Work Organ. Psychol. 5 (2), 185e201. http:// dx.doi.org/10.1080/13594329608414854. Eisenberger, R., Huntington, R., Hutchison, S., Sowa, D., 1986. Perceived organizational support. J. Appl. Psychol. 71, 500e507. Eisenberger, R., Stinglhamber, F., Vandenberghe, C., Sucharski, I., Rhoades, L., 2002. Perceived supervisor support: contributions to perceived organizational support and employee retention. J. Appl. Psychol. 87 (3), 565e573. Elsbach, K.D., 2003. Relating physical environment to self-categorizations: identity threat and affirmation in a non-territorial office space. Adm. Sci. Q. 48 (4), 622e654. Elsbach, K.D., Bechky, B.A., 2007. It's more than a desk: working smarter through leveraged office design. Calif. Manag. Rev. 49 (2). Elsbach, K.D., Pratt, M.G., 2007. Physical environment in organizations. Acad. Manag. Ann. 181e223. Fawcett, W., Chadwick, A., 2007. Space-time management and office floorspace demand. J. Corp. Real Estate 9 (1), 5e24. Felstead, A., Jewson, N., Walters, S., 2003. The Changing Place of Work. ESRC Future of Work Programme. Working Paper 28. Foote, N.N., 1985. The Friendship Game. Image Books, New York. Fulton, C., 2005. Hotdesking. In: Mobile Information Systems II. Springer, pp. 163e170. Giles, C., 2006. The sustainability of innovative schools as learning organizations and professional learning communities during standardized reform. Educ. Adm. Q. 41 (1), 124e156. http://dx.doi.org/10.1177/0013161X05278189. Grey, D.E., 2014. Doing Research in the Real World, third ed. Sage, Los Angeles. Haines, V.A., Hurlbert, J.S., Zimmer, C., 1991. Occupational stress, social support, and the buffer hypothesis. Work Occup. 18, 212e235. n, U., 2007. Safe by design: where are we now? Saf. Sci. 45, Hale, A., Kirwan, B., Kjelle 305e327. http://dx.doi.org/10.1016/j.ssci.2006.08.007. Haynes, B., Price, I., 2004. Quantifying the complex adaptive workplace. Facilities 22 (1/2), 8e18. Heidmets, M., 1994. The phenomenon of personalization of the environment: a theoretical analysis. J. Russ. East Eur. Psychol. 32 (3), 41e85. Hill, S.E.K., Bahniuk, M.H., Dobos, J., Rouner, D., 1989. Mentoring and other communication support in the academic setting. Group Organ. Stud. 14 http:// dx.doi.org/10.1177/105960118901400308, 355e336. Hirst, A., 2011. Settlers, vagrants and mutual indifference: unintended consequences of hot-desking. J. Organ. Change Manag. 24 (6), 767e788. http:// dx.doi.org/10.1108/09534811111175742. Hutton, S.A., 2006. Workplace incivility: state of the science. J. Nurs. Adm. 36 (1), 22e27. Irving, G., & Ayoko, O. (2014). An exploratory study of the connection between office environments and group cognition. presented at the meeting of the 28th Conference of the Australian and New Zealand Academy of Managemnt (ANZAM) 2014, Sydney, Australia. Kaarlela-Tuomaala, A., Helenius, R., Keskinen, E., Hongisto, V., 2009. Effects of acoustic environment on work in private office rooms and open-plan offices e longitudinal study during relocation. Ergonomics 52 (11), 1423e1444. Kim, J., de Dear, R., 2013. Workspace satisfaction: the privacy-communication tradeoff in openplan offices. J. Environ. Psychol. 36, 18e26. Koetsveld, R.v., Kamperman, L., 2011. How flexible workplace strategies can be made successful at the operational level. Corp. Real Estate J. 303e319. October. Kottke, J.L., Sharafinski, C.E., 1988. Measuring perceived supervisory and organizational support. Educ. Psychol. Meas. 48, 1075e1079. Kouvonen, A., Oksanen, T., Vahtera, J., Stafford, M., Wilkinson, R., Schneider, J., et al., 2008. Low workplace social capital as a predictor of depression - the Finnish public sector study. Am. J. Epidemiol. 167 (10), 1143e1151. http://dx.doi.org/ 10.1093/aje/kwn067. Krackhardt, D., Stern, R.N., 1988. Informal networks and organizational crises: an experimental simulation. Soc. Psychol. Q. 51 (2), 123e140. Kramer, M.W., 1996. A longitudinal study of peer communication during job transfers. The impact of frequency, quality, and network multiplexity on adjustment. Hum. Commun. Res. 23 (1), 59e86. Kupritz, V., 2003. Accommodating privacy to facilitate new ways of working. J. Archit. Plan. Res. 20 (2). Labianca, G., Brass, D.J., 2006. Exploring the social ledger: negative relationships and negative asymmetry in social networks in organisations. Acad. Manag. Rev. 31 (3), 596e614. http://dx.doi.org/10.5465/AMR.2006.21318920. Loewen, L., Suedfeld, P., 1992. Cognitive and arousal effects of masking office noise. Environ. Behav. 24 (3), 381e395. Maher, A., von Hippel, C., 2005. Individual differences in employee reactions to open-plan offices. J. Environ. Psychol. 25 (2), 219e229. McAllister, D.J., 1995. Affect- and cognition-based trust as foundations for interpersonal cooperation in organizations. Acad. Manag. J. 38 (1), 24e59. http:// dx.doi.org/10.2307/256727. McCoy, J.M., 2005. Linking the physical work environment to creative context.

R.L. Morrison, K.A. Macky / Applied Ergonomics 60 (2017) 103e115 J. Creat. Behav. 39, 169e191. McElroy, J.C., Morrow, P.C., 2010. Employee reactions to office redesign: a naturally occurring quasi-field experiment in a multi-generational setting. Hum. Relat. 63, 609e636. Meijer, E.M., Frings-Dresen, M.H.W., Sluiter, J.K., 2009. Effects of office innovation on office workers' health and performance. Ergonomics 52 (9), 1027e1038. Moerbeek, H.H.S., Need, A., 2003. Enemies at work: can they hinder your career? Soc. Netw. 25 (1), 67e82. http://dx.doi.org/10.1016/S0378-8733(02)00037-0. Morrison, R., 2004. Informal relationships in the workplace: associations with job satisfaction, organisational commitment and turnover intentions. N. Z. J. Psychol. 33 (3), 114e128. Morrison, R.L. (2006, 1e4 July). All friends are not created equal. An evaluation of the Workplace Friendship Scale: Is there ecological validity in measuring the “Prevalence of Friends” at work? Symposium conducted at the meeting of the Australian Centre for Research in Employment and Work (ACREW) Conference. “Socially responsive, socially responsible approaches to employment and work”, Monash University Prato Centre, Italy. Morrison, R.L., 2009. Are women tending and befriending in the workplace? Gender differences in the relationship between workplace friendships and organisational outcomes. Sex. Roles 60 (1), 1e13. Morrison, R.L., Nolan, T., 2007. Negative relationships in the workplace e a qualitative study. Qual. Res. Account. Manag. 4 (3) http://dx.doi.org/10.1108/ 11766090710826646. Nahemow, L., Lawton, M.P., 1975. Similarity and propinquity in friendship formation. J. ol Pers. Soc. Psychol. 32 (2). Nemecek, J., Grandjean, E., 1973. Results of an ergonomic investigation of largespace offices. Hum. Factors 15 (2), 111e124. Nielsen, I.K., Jex, S.M., Adams, G.A., 2000. Development and validation of scores on a two dimensional Workplace Friendship Scale. Educ. Psychol. Meas. 60 (4), 628e643. http://dx.doi.org/10.1177/00131640021970655. Oldham, G.R., Brass, D., 1979. Employee reactions to an open-plan office: a naturally occurring quasi-experiment. Adm. Sci. Q. 24 (2), 267e284. Patty, A., 2015, December 1. Hot-desking May Lead to More Strain Injuries. The Sydney Morning Herald. Retrieved from: http://www.smh.com.au/business/ workplace-relations/hotdesking-may-lead-to-more-strain-injuries-20151127gla06j.html. Pearson, C.M., Anderson, L.M., Wegner, J.W., 2001. When workers flout convention: a study of workplace incivility. Hum. Relat. 54, 1387e1419. http://dx.doi.org/ 10.1177/00187267015411001. Perini, K., Rosasco, P., 2013. Costebenefit analysis for green façades and living wall systems. Build. Environ. 70, 110e121. Pitt, M., Bennett, J., 2008. Workforce ownership of space in a space sharing environment. J. Facil. Manag. 6 (4), 290e302. http://dx.doi.org/10.1108/ 14725960810908154. Podsakoff, P.M., MacKenzie, S.B., Podsakoff, N.P., 2012. Sources of method bias in social science research and recommendations on hot to control it. Annu. Rev. Psychol. 63, 539e569. http://dx.doi.org/10.1146/annurev-psych-120710100452. Punie, Y., 2007. Learning spaces: an ict-enabled model of future learning in the knowledge-based society. Eur. J. Educ. 42 (2), 185e199. http://dx.doi.org/ 10.1111/j.1465-3435.2007.00302.x. Rakhshandehroo, M., Yusof, M.J.M., Najd, M.D., 2015. Green façade (vertical greening): benefits and threats. Appl. Mech. Mater. 747. Reh, S., Rabenstein, K., Fritzsche, B., 2011. Learning spaces without boundaries? Territories, power and how schools regulate learning. Soc. Cult. Geogr. 12 (1), 83e98. http://dx.doi.org/10.1080/14649365.2011.542482. Rhoades, L., Eisenberger, R., 2002. Perceived organizational support: a review of the literature. J. Appl. Psychol. 87 (4), 698e714.

115

Rich, B.L., Lepine, J.A., Crawford, E.R., 2010. Job engagement: antecedents and effects on job performance. Acad. Manag. J. 53 (3), 617e635. Richer, S.F., Blanchard, U., Vallerand, R.J., 2002. A motivational model of work turnover. J. Appl. Soc. Psychol. 32 (10), 2089e2113. Riordan, C.M., Griffeth, R.W., 1995. The opportunity for friendship in the workplace: an underexplored construct. J. Bus. Psychol. 10 (2), 141e154. Rytivaara, A., 2011. Flexible grouping as a means for classroom management in a heterogeneous classroom. Eur. Educ. Res. J. 10 (1), 118e128. http://dx.doi.org/ 10.2304/eerj.2011.10.1.118. Sapp, A.L., Kawachi, I., Sorensen, G., LaMontagne, A.D., Subramanian, S.V., 2010. Does workplace social capital buffer the effects of job stress? A cross-sectional, multilevel analysis of cigarette smoking among US manufacturing workers. J. Occup. Environ. Med. 52 (7), 740e750. http://dx.doi.org/10.1097/ JOM.0b013e3181e80842. Sarason, I.G., Sarason, B.R., Shearin, E.N., Pierce, G.R., 1987. A brief measure of social support: practical and theoretical implications. J. Soc. Pers. Relat. 4, 497e510. http://dx.doi.org/10.1177/0265407587044007. Scheiberg, S.L., 1990. Emotions on display. Am. Behav. Sci. 33, 330e338. Schwab, D.P., 1980. Construct validity in organizational behavior. In: Staw, B.M., Cummings, L.L. (Eds.), Research in Organizational Behavior. JAI Press, Greenwich, CT, pp. 3e43. Sheldon, K.M., Bettencourt, B.A., 2002. Psychological need-satisfaction and subjective well-being within social groups. Br. J. Soc. Psychol. 41, 25e38. Smith-Jackson, T.L., Klein, K.W., 2009. Open-plan offices: task performance and mental workload. J. Environ. Psychol. 29, 279e289. http://dx.doi.org/10.1016/ j.jenvp.2008.09.002. Strayer, J.F., 2012. How learning in an inverted classroom influences cooperation, innovation and task orientation. Learn. Environ. Res. 15 (2), 171e193. http:// dx.doi.org/10.1007/s10984-012-9108-4. Sundstrom, E., Burt, R.E., Kamp, D., 1980. Privacy at work: architectural correlates of job satisfaction and job performance. Acad. Manag. J. 23 (1), 101e117. Sundstrom, E., Town, J., Rice, R., Osborn, D., Brill, M., 1994. Office noise, satisfaction, and performance. Environ. Behav. 26 (2), 195e222. Suzuki, E., Takao, S., Subramanian, S.V., Komatsu, H., Doi, H., Kawachi, I., 2010. Does low workplace social capital have detrimental effect on workers' health? Soc. Sci. Med. 70 (9), 1367e1372. http://dx.doi.org/10.1016/j.socscimed.2010.01.014. Thompson, A.D., 2015. Preparing our teacher education students for new schools. J. Digit. Learn. Teach. Educ. 31 (3), 83e84. http://dx.doi.org/10.1080/ 21532974.2015.1059710. Vecchio, R.P., 2000. Negative emotion in the workplace: employee jealousy and envy. Int. J. Stress Manag. 7 (3), 161e179. http://dx.doi.org/10.1023/A: 1009592430712. Voordt, T.J.M.v.d, 2004. Productivity and employee satisfaction in flexible workplaces. J. Corp. Real Estate 6 (2), 133e148. Warren, C.M.J., 2003. New working practice and office space density: a comparison of Australia and the UK. Facilities 21 (13/14), 306e314. Wells, M., 2000. Office clutter or meaningful personal displays: the role of office personalization in employee and organizational well-being. J. Environ. Psychol. 20, 239e255. Wells, M., Thelen, L., 2002. What does your workspace say about you? The influence of personality, status, and workspace on personalization. Environ. Behav. 34, 300e321. Wells, M.M., Thelen, L., Ruark, J., 2007. Workspace personalization and organizational culture does your workspace reflect you or your company? Environ. Behav. 39 (5), 616e634. http://dx.doi.org/10.1177/0013916506295602. White, C.D., Campbell, K.S., Kacmar, K.M., 2012. Development and validation of a measure of leader rapport management: the lrm scale. Inst. Behav. Appl. Manag. 121e149.