Introduction

The idea of extending the typical workplace to subterranean environment was initially an option seriously considered (and adopted) by countries with specific problems - such as Japan, a country with cities with very high population density and continuously rising land prices1; or countries with subarctic climates with relatively extreme weather conditions2. However, a combination of recent climatic3, urban4 and macroeconomic5 changes have led many cities to actively examine the possibility of having underground office workspaces (UOW) as a pragmatic solution with many advantages: UOW can free-up more surface land in cities, protect occupants from various harmful environmental conditions, and overall meet the sustainability goal by reducing carbon emission associated with urban sprawl6,7,8 and preserving natural landscapes9. Given the United Nations’ forecast that the global urban population will constitute 68% by 205010, the development and significance of sustainable urban UOWs are gaining momentum as an integral element for accommodating the needs of the expanding urban workforce.

To date, the technological and engineering aspects of UOW have been extensively researched and evolved to meet the physical needs of the space (e.g.,, air ventilation, lighting conditions, environmental safety11,12,13 while prioritizing environmental sustainability (e.g., high energy conservation measures14). Yet sustainable urban development, as the United Nations – Habitat agenda (2016) stated, places a strong emphasis on human well-being and active participation as critical elements in the pursuit of sustainable urban environments15. UOW, in this context, also needs to adhere to the principles of respecting and incorporating the psychological needs of occupants. However, research that looks into the human psychological perspectives of UOW is rather sparse and limited. Hence, the current research aims to address this gap by investigating the psychological mechanisms that support the adoption of UOW within the framework of sustainable urban development.

It is generally assumed, with not much empirical evidence, that the public holds a negative view of underground environments. This view often stems from the feeling of confinement linked to the enclosed nature of underground spaces16, which is inherently associated with diminished sense of control2. Such psychosocial characteristics of underground space are meaningful in the context of understanding people’s reaction to UOW because an environment that limits personal control over the space and prevents effective self-regulation to accomplish needed goals could be perceived negatively and actively avoided17. In line with this, the social ecological perspective in environmental adoption18 states that individual’s capacity to exert control and cope with the environment are crucial factors of environmental evaluation, which subsequently affect people’s willingness to interact with the environment19,20.

Drawing upon the association between the significance of ‘control’ in evaluating spaces and the enclosing features of underground environments that may diminish control16,18,21, it can be inferred that perceived control forms a central aspect of psychological reactions to underground occupancy. This research aims to empirically assess the proposition by investigating the relationship between individuals’ sense of control and the adoption of UOW. A two-fold study design is implemented to examine the phenomenon, and strengthen its ecological validity: in Study 1, we investigate the relationship between a psychological trait that relates to one’s sense of control and the attitude toward UOW and, in Study 2, we examine the contribution of this trait in employees’ actual decisions to work in UOW retrospectively by comparing employees’ sense of control of current underground and well-matched aboveground working populations. Through this, we examine whether the psychological mechanism suggested to be important by a prospective move (Study 1) also persists and plays a central role in the actual adoption of the space (Study 2), enhancing the ecological validity of the identified mechanism.

General perception of underground spaces

Traditionally, an underground space is known as an enclosed space below the surface of the earth. Due to its location, there is no direct path to the external environment unless through designated exits16. Thus, typical but not necessary attributes of built underground environment include lack of windows, loss of contact with nature, and dark space2. These architectural limitations have been suggested to be related to specific profiles of psychological characteristics. For instance, lack of windows prevents the visibility of any landmark around the area, making harder to locate oneself and to way-find within the structure, which could leave occupants feeling lost. Moreover, a number of actions (e.g. opening windows for ventilation, adjustment of curtains to control natural light access) are restricted which could reduce the amount of control occupants have over the environment22. Darker luminosity as well as uniform architecture of the space can further worsen one’s feeling of confidence and certainty within the space2. Additional associations often discussed include lifelessness, isolation, entrapment, and stagnation21,23,24. Although contemporary underground spaces have addressed numerous architectural challenges via engineering and technological advancements, there remains a prevalent public perception rooted in these old-fashioned views of subterranean environments, notably associating them with a loss of control.

More recent research has attempted to analyze psychological reactions to underground work environments systematically by identifying the foundational attitudinal factors that underpin these reactions. According to Lee et al. (2019)25, these psychological reactions are primarily rooted in three key attitudinal factors: calmness/safety, negative affect, and confinement. The calmness/safety factor generally reflect positive attitudes towards underground workspaces, such as feelings of comfort, protection, and stability, which presumably stem from the protection underground structures can offer from various uncertainties aboveground. Negative affect captures negative emotional responses to underground workspaces, including fear, nervousness, and loneliness, which stem from the perceived isolation, restricted spaces, and potential risks associated with such settings. Lastly, confinement represents thoughts and understanding associated with the confining attributes of underground spaces, such as believing that the space lacks air or the space is associated with negative cultural connotations that are related to being confined below the ground level (e.g., ‘staying inside a tomb’). This foundational analysis suggests that both positive and negative attitudes toward underground built environments are tied to the concept of encapsulation. Here, encapsulation can represent either a protective separation from the broader community, providing safety and comfort by shielding occupants from external stimuli, or a sense of confinement rooted in traditional perceptions of underground spaces—characterized by darkness, dampness, isolation, and restricted movement—that expose occupants to unfavorable environmental conditions21,25. The extent to which these aspects are perceived to a greater or less degree may vary depending on individuals’ characteristics, experiences, and overall tendencies in perceiving their environment. Building on the earlier discussion highlighting the importance of perceived control in evaluating UOW, we propose that Locus of Control (LoC)—a personality trait reflecting an individual’s sense of control—will influence their attitudes toward UOW, affecting the level of Safety/Comfort (positive), Negative Affect (negative), and Confinement (negative) felt toward the environment.

Locus of control: perception of control

LoC is a type of personality trait that reflects the degree to which people believe that they can determine their own environment and/or bring desired outcomes as opposed to external forces26,27. When LoC is located internally, the person believes the source of control is within themselves and that events in their life are determined by their own actions, reflecting a sense of mastery and self-efficacy. On the other hand, when LoC is located externally, individuals believe that control lies outside of themselves, attributing outcomes in their life to external factors such as luck, fate, or the influence of others. This external orientation often leads to a perception of limited personal agency and a reliance on external circumstances to shape their experiences. In the literature on LoC, scholars often divide this construct into two independent components: internal LoC (reflecting the extent to which individuals believe control resides within themselves) and external LoC, (representing the extent to which individuals believe control lies outside of themselves). This distinction is used to capture these two aspects in detail, as they may not always correlate, thereby providing a clearer and more accurate understanding of a person’s overall sense of control28.

LoC plays a crucial role in shaping various life outcomes, with internal and external orientations influencing different aspects of an individual’s experiences. An internal LoC, where individuals believe they control their life outcomes, is linked to positive outcomes such as higher well-being, academic success, and resilience to adversity29. It fosters motivation, achievement, and life satisfaction, acting as a buffer against negative events like illness or crime victimization, although it may not mitigate long-term economic impacts30. Internal LoC also supports pro-social behaviors and better coping, with factors like parental supervision aiding its development in children31. Conversely, an external LoC, where individuals perceive their lives as controlled by external forces, is often linked to negative outcomes such as poor academic performance, stress, and mental health challenges32,33. Together, these findings reveal the complex role of LoC, with internal LoC generally promoting adaptive behaviors and external LoC highlighting vulnerabilities in certain contexts.

While there has been limited research on the connection between LoC and responses to the physical environment, one study found that individuals with a more external LoC orientation often arrange their personal office spaces to minimize their visibility to visitors. This behavior is likely a defensive measure aimed at reducing the potential stress from unpredictable interactions with visitors34. In other words, individuals with a greater degree of external LoC, who feel less capable of managing situations, actively re-arranged their workspace to create an environment that offered greater control over interactions with visitors, thereby improving their ability to manage stress during their occupancy. Retrospectively, giving more control over the environment to individuals was suggested to increase satisfaction with the space35, which contributes to the future engagement with an environment positively36. Thus, individuals’ beliefs about their capacity to control situations and surroundings, as reflected by their LoC, may significantly influence how they form attitudes toward their immediate physical environment and how they interact with it. In particular, in environments like UOW, which inherently feature architectural elements that can diminish one’s sense of control, LoC may serve as a central psychological determinant of whether a person is able to adapt to and accept such spaces.

The present study and hypotheses

This research investigates the link between individuals’ LoC and the adoption of UOW through a two-part study: Study 1 investigates whether LoC influences attitudes toward UOW, while Study 2 examines its role in actual decisions to work in UOW by comparing current underground and aboveground employees’ degree of LoC (refer to the flowchart in Fig. 1 for an overview of the general progression of the entire research). Drawing upon the discussion, individuals with an externally oriented LoC might view UOW less favorably, perceiving its inherent features that are unfamiliar and confining (which restricts control) as more challenging and harder to manage compared to those with an internally oriented LoC, who typically show greater resilience to external environments. Accordingly, we posit the following hypotheses on the relationship between internal and external LoC42, and the three factors of the attitude toward UOW (calmness/safety, negative affect, and confinement) proposed by Lee et al.25. The hypothesis models are presented in Fig. 2:

Fig. 1
figure 1

Flow chart of an overview of the general progression of the entire research.

Full size image
Fig. 2
figure 2

Hypothesized models for study 1.

Full size image

Hypothesis 1a

Internal LoC is negatively related to negative affect toward underground workspaces.

Hypothesis 1b

Internal LoC is negatively related to confinement related thoughts about underground workspaces.

Hypothesis 1c

Internal LoC is positively related to calmness / safety perception toward underground workspaces.

On the other hand, we predict the opposite pattern of relationships for external LoC as below:

Hypothesis 2a

External LoC is positively related to negative affect toward underground workspaces.

Hypothesis 2b

External LoC is positively related to confinement related thoughts about underground workspaces.

Hypothesis 2c

External LoC is negatively related to calmness / safety perception toward underground workspaces.

Moreover, it is speculated that LoC not only shapes how individuals perceive and form attitudes toward UOW but also influences their choice of work location based on their perceived capacity to manage and cope with UOW according to their level of LoC. Viewed retrospectively, this suggests that the degree of LoC among today’s workforce varies based on their office location—underground or above ground—reflecting their perceived ability to manipulate their environments to achieve desired outcomes. Following this, we present the hypotheses on the levels of internal and external LoC of the two working populations based on the location of office (see Fig. 3):

Fig. 3
figure 3

Hypothesized models for study 2.

Full size image

Hypothesis 3a

Individuals working in underground offices have a greater degree of internal LoC compared to those working in comparable aboveground offices.

Hypothesis 3b

Individuals working in underground offices have a lesser degree of external LoC compared to those working in comparable aboveground offices.

In addition, we suggest that the duration of employment serves as a critical moderator influencing the LoC variation of employees based on office location. As employment duration increases, LoC variation is likely to widen as prolonged interaction with the office environment enables employees to fully evaluate their adaptability, ultimately influencing their continued occupancy at the workplace. Employees who cannot adapt to and cope with UOW due to a relatively higher external LoC and lower internal LoC are expected to be selected out further (e.g. leave the current employer or re-locate). In other words, over time, employees in UOW with a more internally oriented LoC, who have a stronger belief in their ability to control their situations and environment, are more likely to remain in the workplace. Thus, we present the next hypothesis on the moderating effect of employment duration on the predictability of work location in LoC levels.

Hypothesis 4

As employment duration increases, the differences in internal and external LoC between employees working in underground and aboveground offices become more pronounced, aligning with the directions described in Hypotheses 3a and 3b.

Study 1 aimed to investigate and establish the link between individuals’ LoC orientation and the attitudes toward UOW. A large-scale survey was conducted with the general public of Singapore to examine the relationship between internal and external LoC with each attitude factor, calmness/safety, negative affect, and confinement25, toward a hypothetical deep UOW to gain an insight into the psychological reaction of prospective employees of UOW.

Study1

Method

Sample

The sample, comprised of 1093 Singaporean citizens and permanent residence holders (women: 548 (50.1%)), was recruited through a specialized market research company, Qualtrics, and was administered an online survey using the Qualtrics platform. In terms of ethnicity, 71.6% were Chinese, 14.9% were Malay, 10.2% were Indian, and 3.2% were others, which was matched to the ethnic diversity of Singapore. Mean age was 36.07 years (SD = 11.21).

Ethics statement

Informed consent was obtained from all the survey participants, the anonymity of the participants was ensured and collected data was kept confidential. Before starting the study, the Institutional Review Board of Nanyang Technological University (the central ethics committee) reviewed and approved our study procedures. The study followed the Declaration of Helsinki guidelines for all procedures involving human subjects.

Measure

Locus of control

LoC was measured using Levenson’s locus of control scales37. The internal LoC scale (α = 0.76) measures the extent to which individuals believe reinforcements as contingent on their own behavior, hence, the degree to which control is internal. The external (chance) LoC scale (α = 0.78) measures the belief that experiences and outcomes are controlled by chance or luck, reflecting the extent to which control is determined by external factors. One of the items in each scale was written in the context of car accident for the general population. However, given the low private car ownership rate in Singapore38, these context-based items were omitted. Participants rated on a 5-point Likert scale (1 = Strongly Disagree, 5 = Strongly Agree).

Attitudes toward working in underground workspace

Perception toward working underground was measured by the underground workspaces questionnaire (UWSQ; Lee et al., 2019). The UWSQ assesses pre-occupant attitudes toward underground workspaces, and is comprised of three factors, namely, calmness/ safety, negative affect, and confinement. The calmness/ safety factor measures the view on positive aspects of underground spaces, reflecting both affective and practical considerations (α = 0.88). The negative affect factor consists of items that depict negative affective states (α = 0.87). In other words, the question asks how the respondent would feel if they were working in an underground office (e.g. nervous, lonely, and distracted). The confinement factor reflects negative beliefs in relation to the elements of confinement of the underground workspace (α = 0.86), such as concern about air quality and associated negative cultural beliefs (e.g. burial and evil). Participants rated 16 items in total on 7-point Likert scales (1 = Strongly Disagree to 7 = Strongly Agree).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results and short discussion for study 1

Table 1 presents the means, standard deviations, and simple correlations between each attitude factor of UWSQ (Calmness/ Safety, Negative Affect, Confinement), and internal and external LoC. For the main analyses, linear regressions were conducted to determine whether internal and external LoC (predictor variables) predict the attitudes toward working underground (dependent variables): Calmness / Safety, Negative Affect, and Confinement. Age and gender were included as control variables because they are widely recognized in social, behavioral, and psychological research as key demographic factors influencing outcomes. Controlling for these variables ensures that observed effects are not confounded by demographic differences. Moreover, in the context of architectural studies, research indicates that women often perceive environments in greater detail, prioritising aspects such as cleanliness, aesthetics, and safety, while men may focus more on connectivity39. Gender differences also extend to sensory and neurological responses to architectural settings40. Likewise, age influences environmental perceptions, with older adults placing greater importance on personal control, comfort, and space usage, whereas younger individuals value environments that support teamwork41. Thus, including age and gender accounts for potential variability in responses and enhances the robustness of the analysis.

First, negative affect toward the underground office was tested as the outcome variable. The results showed that external LoC (β = 0.222) was associated with negative affect, supporting Hypothesis 1b, but internal LoC was not associated with negative affect, hence Hypothesis 1a was not supported. This implies that people who exhibit greater external LoC in particular had more negative affect toward the underground office (see Fig. 4; Table 2).

Next, perceived confinement toward the underground office was tested as the outcome variable. External LoC was a significant predictor (β = 0.161) of perceived confinement, supporting Hypothesis 2b. As in the previous analysis, internal LoC was not associated with perceived confinement toward the underground office so our Hypothesis 2a was not supported (see Fig. 4; Table 3).

Lastly, calmness / safety perception toward the underground office was tested as the outcome variable. The results revealed that there was no main effect of any of the LoC variables, which suggests that LoC variables were not associated with the perception of calmness and safety toward the underground office. Therefore. Hypothesis 3a and 3b were not supported (see Fig. 4; Table 4).

The findings generally support our hypotheses, showing that the degree of LoC influences attitudes toward UOW. Specifically, individuals with a greater degree of external LoC, who feel overpowered by external forces, exhibit more negative feelings toward working in UOW (negative affect) and perceive the environment as more confining (confinement). On the other hand, internal LoC had no predictive values in attitudes toward UOW, suggesting that an individuals’ belief in their ability to manage situations plays a less critical role in shaping attitudes compared to the perception of being influenced by external forces (external LoC). Lastly, perceived comfort and safety regarding UOW was not influenced by LoC.

Table 1 Means, Standard Deviations, and correlations.
Full size table
Table 2 Summary of regression analysis for the effect of locus of control on calmness/safety perception.
Full size table
Table 3 Summary of regression analysis for the effect of locus of control predicting negative affect.
Full size table
Table 4 Summary of regression analysis for the effect of locus of control on perceived confinement.
Full size table
Fig. 4
figure 4

The Relationship between (a) Negative Affect, (b) Confinement, and (c) Perceived Calmness and Safety toward Underground Workspaces and Locus of Control (Internal and External). External locus of control positively predicted two negative attitude factors (negative affect and confinement) associated with working in UOW. However, it did not predict the positive attitude factor (calmness and safety). Internal locus of control did not predict any of the attitude factors associated with working in UOW.

Full size image

Study 2

Study 1 revealed that the general public with an externally oriented LoC generally holds negative attitudes toward the idea of working in UOW. However, a key question remains: does LoC influence the actual decision to work in UOW? Study 2 aims to fill in this gap by examining the LoC of the current working population of underground and comparable aboveground offices to retrospectively determine whether differences in LoC are associated with their actual work environment choices. Thus, the internal and external LoC of the current underground and aboveground employees who work in the same industry were compared, based on the assumption that their choice of the current work location reflects the employees’ psychological profiles that enabled the adoption of UOW. If people with predominantly external LoC (feeling controlled by outside forces) are reluctant to work underground, it is expected that the current underground working population will possess relatively internal LoC (feeling in control of their own lives) overall.

Method

Sample and procedure

Data were collected from a cohort of 217 participants between November 2016 and April 2017 at two time points. Initially, 75 workers from underground offices and 142 workers from aboveground offices participated in the study, but 7 participants from underground and 13 participants from aboveground offices dropped out of the study by the second round due to personal issues. In terms of ethnicity, 74.7% were Chinese, 9.2% were Malay, 12.0% were Indian, and 4.1% were others, which was representative of the ethnic diversity of Singapore. The mean age of participants was 38.40 (SD = 11.67). Our sample was predominantly male (69.1%, N = 150) due to the nature of the industry chosen (transport industry). Participants were invited to participate in the study via meetings, workplace posters, and emails. Those who expressed interest in participation were screened for eligibility based on the following criteria: aged 21 years and above, should be working for at least four hours per day at their stated workplace, and not pregnant. Participants were deemed ineligible if they made at least one trip per month to countries in a different time zone from Singapore in the past 6 months. Importantly, all our participants worked in the same sector of semi-governmental companies, with very similar job functions (office desk-bound job) with almost identical organizational characteristics (e.g. work hours and hiring practices), which greatly accounts for uncontrolled organizational differences.

Ethics statement

Informed consent was obtained from all the survey participants, the anonymity of the participants was ensured and collected data was kept confidential. Before starting the study, the Institutional Review Board of Nanyang Technological University (the central ethics committee) reviewed and approved our study procedures. The study followed the Declaration of Helsinki guidelines for all procedures involving human subjects.

Materials

Work site

Four office buildings supporting two companies within the same industry in Singapore were selected as the study sites. These organizations had offices in either under- or aboveground environments or both. Indoor environmental quality parameters were all within the recommended levels (see Table 5 for general characteristics of different office layouts).

Table 5 Characteristics of office layouts for underground vs. aboveground.
Full size table
Locus of control

As in Study 1, LoC was measured using Levenson’s locus of control scale42. Again, the internal LoC scale (α = 0.79) that measures the degree to which control is internal, and the external LoC scale (α = 0.78) which measures the belief that experiences, and outcomes are controlled by chance or luck (externally oriented), were administered.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results and short discussion for study 2

Univariate general linear models (GLM) were used to examine the differences in LoC of those who work in underground and aboveground offices: the location of the office (2: aboveground vs. underground) was the independent variable and the two types of LoC (Internal and external) were dependent variables. Specifically, for each of LoC (internal and external), we conducted an analysis of covariance (ANCOVA) that compared the effects of underground versus aboveground location on LoC after controlling for age and gender as in Study 1. In addition, we included the number of months the employees worked (duration of work) at the current location of company as a covariate to control for the effect of duration of work.

First, we examined the differences of internal LoC of underground vs. aboveground office workers. The results ANCOVA are presented below (Table 6; Fig. 5), suggesting that underground office workers had higher internal LoC score compared to aboveground workers, F(1, 212) = 4.803, p = .029, after controlling for age, gender, and employment duration. Thus, H3a was supported.

We then examined whether employment duration moderated the effect of office location on internal LoC, to determine if internal LoC levels became more extreme in the direction predicted by Hypothesis 3a as employment duration increased. A simple moderation analysis was performed using PROCESS macro in SPSS (a statistical tool designed to test complex models, such as moderation and mediation43). The dependent variable for analysis was internal LoC, and the independent variable was the location of office (underground vs. aboveground). Age and gender were entered as covariates to control for the effects. The results showed that the interaction between the location of office and work duration was non-significant [B = − 0.0009, 95% C.I. (-0.0024, 0.0007), p = .2592], showing that the length of employment at the current location did not moderate the relationship between office location and internal LoC. Thus, H4 was not supported.

Table 6 ANCOVA results for Internal Locus of Control.
Full size table

Moving on to the effect of office location on external LoC, we examined the differences in external LoC of underground office workers versus. aboveground office workers. The results of ANCOVA are shown below (Table 7; Fig. 5), indicating that the workers in underground offices scored lower in external LoC compared to the workers in aboveground offices F(1,192) = 4.548, p = .034, after controlling for age, gender, and work duration, supporting H3b.

Again, we checked whether employment duration moderated the predictability of office location on internal LoC using PROCESS macro. The dependent variable for the analysis was external LoC, and the independent variable was the location of office, with age and gender entered as covariates to control for the effects. The results showed that the interaction between the location of office and employment duration was non-significant [B = − 0.0001, 95% C.I. (-0.0017, 0.0019), p = .9028], confirming that there was no moderating effect driven by the length of employment at the current location. Thus, H4 was not supported.

In short, the location of office (underground vs. aboveground) predicted both internal and external LoC of the current employees, such that those who work in underground offices exhibited more internally oriented LoC (i.e., high internal LoC and low external LoC) compared to those who work in aboveground offices. There was no moderation effect by employment duration for either LoC, suggesting that the actual experience of working in an underground office did not amplify the differences in LoC in the predicted directions.

Table 7 ANCOVA results for external locus of control.
Full size table
Fig. 5
figure 5

Bar charts showing the differences in internal (a) and external (b) locus of control across under- vs. aboveground office workers. Underground office employees have higher internal locus of control and lower external locus of control.

Full size image

Discussion

This research examined whether individuals’ general sense of control, as reflected by the degree of internal and external LoC levels, is a fundamental psychological trait that influences the attitudes and acceptance of working in an UOW. Through an examination of the general public’s attitude towards a hypothetical move to UOW (Study 1), as well as an observational study involving current aboveground and underground employees (Study 2), our findings suggest that individuals’ sense of personal control serves as a significant predictor of the adoption of UOW. Specifically, Study 1 demonstrated that people who have externally oriented LoC had negative attitude toward UOW, while Study 2 demonstrated that those who currently work in UOW exhibited relatively internal LoC. These results collectively suggest that individuals who feel less capable of exerting control over their lives may find occupying UOW challenging and actively avoid such spaces.

Implications

The current research provides theoretical implications for understanding how psychological profiles affect interactions with the physical environment. Although both studies generally supported the main hypothesis that people with externally oriented LoC, believing that their lives are largely controlled by external forces or circumstances, are reluctant to the idea of working underground, the findings of Study 1 showed that only external LoC was a significant predictor of the attitudes toward UOW, when in Study 2, both internal and external LoC were shown to be significant predictors of UOW occupancy. These results shed light on an important aspect that distinguishes psychological processes underlying mere attitudes versus active interaction with a physical environment.

When a person evaluates an environment without commitment of interaction, as in Study1 (where participants only had to report their attitudes toward a hypothetical UOW), one’s concern for action capacity within the environment may become less important because there is no need for one to realistically project their capacity to manage and survive the environment17,18. In this context, internal LoC, representing an individual’s belief in their ability to influence outcomes, diminishes in significance in the evaluation process because the person does not expect to exert control over the environment (i.e., UOW). On the other hand, external LoC, which reflects an individual’s perception of external constraints, is still likely to remain significant regardless of the level of interaction commitment, as this psychological construct originates from one’s evaluation of external factors26. Thus, external LoC is expected to consistently predict evaluations of a physical environment, whether these evaluations involve interaction commitment or not, whereas internal LoC is significant only in cases involving interaction commitment. Accordingly, the findings of Study 2, where participants had already evaluated and committed to interacting with their physical environment, logically show that both internal and external LoC hold predictive value in their choice of office environment. To some extent, these results align with prior studies that suggested external LoC tends to be a more reliable predictor of subjective wellbeing28,44. Thus, it is recommended that scholars and practitioners in built environment pay more attention to external LoC as a key psychological factor in understanding prospective occupants’ reaction to different environments.

The readers of this article should be careful in interpreting the current results on the association between LoC and UOW. While we argue that those with more externally oriented LoC are reluctant to accept UOW, presumably due to their perceived inability to control the environment, this proposition does not necessarily mean that working at UOW actually depletes one’s control. Study 2 found that employment duration did not amplify the differences in LoC between the two office locations, suggesting that the UOW environment did not cause employees to leave their jobs or struggle to adapt due to their LoC. This suggests that UOW may not have as much of an impact on their ability to function or adjust to the environment as expected based on their pre-conceived beliefs or perceptions. Moreover, research examining various types of workers’ satisfaction in both underground and aboveground settings in the past found no difference in satisfaction with the physical work environment (e.g., lighting, air quality, humidity) between underground workers and their aboveground counterparts, regardless of job type45. Although it is too early to conclude, pre-conceived perception and beliefs about UOW might be the critical factor that underlies various negative psychological reactions and behavioral intentions toward UOW rather than the actual experience of being UOW.

Our studies contribute significantly to theoretical advancements in the psychology of underground spaces. Past research has highlighted an association of UOW with a generic, non-specific negative psychological sentiment. We here move forward by identifying a more precise, theoretically established and measurable mechanism to effectively answer the longstanding question of the exact reasons underlying the negative psychological associations – perception of control. The current research highlights that a very specific aspect of personality that relates to perceived control predicts how one views and adopts UOW: the consistency of key results across different populations in the two studies enhances external validity and generalizability, indicating that the mechanism applies in both prospective and actual occupants. While past literature on underground space has constantly brought the concept of perceived control as an essential psychological element that underlies various reactions toward underground spaces2,16, it has not been directly tested until today. Thus, by outlining LoC as an essential trait that influences people’s attitudes and adoption of UOW, our research extends somewhat limited literature of the psychology of underground spaces by empirically testing and confirming the crucial role of individuals’ perceived control in understanding people’s reaction toward underground environment.

In addition, from the perspective of LoC research, this research is one of the very few available studies that demonstrates the independent and direct effect of LoC on people’s preference for physical environment. Earlier literature on LoC mostly discussed LoC as a key predictor for various well-being outcomes (e.g., job attitudes, affect, motivation and behavioral performance46). However, our findings demonstrate that LoC does not only affect one’s reactions to situations but physical environments that they might need to interact. In this regard, LoC should be regarded as an important psychological variable in understanding how people navigate and utilize an environment, especially an environment that people might be unfamiliar with like UOW.

Interventions and practical applications

Even though LoC is a psychological trait, it can certainly be enhanced and “primed”, especially by architectural or other interventions. Currently, the public seems to associate UOW with the traditional idea of underground environment that limits one’s control, which has a pronounced effect on people with predominantly external LoC when deciding whether to adopt UOW. In essence, this implies that only a restricted number of people may ultimately embrace UOW as their workplace (if they had the freedom to choose), which could be unfavorable for enterprises located in subterranean settings. To reduce such bias, it is recommended to rebrand publicly accessible underground spaces to challenge and change outdated perceptions and beliefs about underground environments. Existing underground structures could be repurposed for a wider range of uses beyond their traditional roles as car parks, storage areas, or simple transitional spaces, enabling them to serve more diverse and innovative functions. For instance, urban farming, particularly hydroponics, is a potential use for multi-storey car-parking structures, offering social, environmental, and economic benefits47. The Lowline project in New York City, is also a good example that demonstrates the innovative use of abandoned underground spaces for public gardens, addressing urban challenges and enhancing city life by transforming idle spaces into green infrastructure with ecological, economic, and social benefits48. Moreover, when promoting these structures to the public, it may be more effective to emphasize their functionality rather than their underground location, to help challenge and reduce biases against such spaces. One way to achieve this is by renaming the floor levels of the building, for example, re-labeling B2 based on its relative position to the deepest level (e.g., B5 could become “Ground Floor,” and B2 could be renamed as Level 4 to remove “B” designation). By not mentioning “basement,” this approach could help dissociate the space from the idea of being underground, potentially reducing negative perceptions and feelings of confinement21.

Alternatively, it may be beneficial to enhance existing underground workspaces by addressing architectural elements that contribute to feelings of confinement and loss of control, thereby improving overall perceptions of underground spaces. For instance, better lighting could be installed to expand the view of workers49; landmark-like features as well as maps that could better assist one’s wayfinding within the building should be easily accessible to enhance one’s sense of direction; greenery can be added to the indoor design to give more access to nature so that it reduces a sense of segregation from the rest of the world50; more spacious transitional spaces (e.g., corridors) should be adopted to decrease a sense of confinement51, building of more exit routes, light wells, and artificial windows within the structure could enhance openness and provide a greater sense of safety and connection to the outside environment21,52.

Moreover, to attract more people to jobs or offices already located underground, companies can focus on enhancing employees’ control over their environment. Offering larger personalized spaces, allowing control over local ventilation, heating/cooling, and lighting intensity, and providing flexible working hours or remote work options can counter perceptions of underground spaces as restrictive and confining. These measures not only improve the experience for current employees but also make these roles more appealing to potential candidates.

Limitations and future directions

As with most research that examines relatively new phenomena in real-life working populations, there are some limitations with the present research. While our hypotheses assume that LoC influences perception toward UOW, and possibly impact potential employees’ adoption of UOW, both our studies were observational studies that cannot confirm the causal link. Building on our empirical findings, we encourage future research to explore the causal relationship between LoC and workspace choice. This could begin with a lab-based experiment that examines how a manipulated sense of control influences workspace preferences while controlling for other unknown variables. However, recognizing the limitations of lab studies in ecological validity, this field would greatly benefit from a large-scale longitudinal study that observes individuals over an extended period to examine whether LoC genuinely influences their choice of physical workplace environments, providing deeper insights into its long-term effects on job settings.

This paper examined two distinct populations across studies. Study 1 focused on prospective occupants (the general public) with limited or no experience in underground spaces, while Study 2 investigated actual occupants of such spaces. While the discrepancy between these populations for testing the link between LoC and the adoption of UOW could be considered a limitation—potentially affecting the results—it also serves as a strength. Testing the proposed mechanism in both groups enhances the generalizability of our findings by demonstrating how the same mechanism underlies both pre-beliefs about the space and the actual experience of the space.

As with all field studies, which inherently lack the precision and control achievable in laboratory settings, Study 2 faced challenges in achieving complete parity or strict control between the two groups (underground vs. aboveground). This introduces the possibility that external variables beyond our control may have influenced the results. Despite this, we made significant efforts to ensure the two populations were as comparable as possible. Both groups are employed within the same organization, share similar demographic characteristics such as age and gender, and work in office environments designed according to standardized organizational guidelines. These measures were implemented to mitigate the influence of uncontrolled factors and to emphasize the role of vertical location (underground vs. aboveground) in shaping the findings.

Lastly, the unique and rare sample of current underground and comparable aboveground office workers makes our Study 2 very special and ecologically valid. However, the participants were mostly males due to the type of industry they were in. While we did not find any gender effect in explaining the differences in LoC of the underground and aboveground office workers, it could still restrict our conclusions to a certain extent. Thus, we encourage future studies to explore participants of different industries where they have more balanced demographics to generalize the unique effect of LoC on the adoption of UOW.

Conclusion

This research aimed to explore a pressing and emerging urban development trend: the expansion of UOW, employing scientific methods to pursue a human-centered approach to development. Previous research identified a negative psychological link with UOW, but the precise mechanisms remained unclear. This is a theoretical but also practical problem, as architectural and engineering solutions were mostly intuitive and opportunistic. We give solid evidence that a central psychological mechanism that underlies the negative association of UOW is the perception of control. With a human-centered approach in focus, the design of UOW can now target more specific solutions, emphasizing autonomy, personal control, and choice for occupants. Consequently, the scientific insights presented in this paper are not only valuable to the academic community but also hold significance for urban planners, designers, engineers, and policymakers.