Causal Model Analysis of the Effect of Policy Formalism, COVID-19 Fear, Social Support and Work Stress on Construction Workers’ Anxiety during the Epidemic

Abstract

This study mainly explores the impact of policy formalism, COVID-19 fear, social support, and work stress on the anxiety of construction workers during the epidemic. The main objective of this study is to incorporate formalism variables to explore their impact on the stress and anxiety of construction workers during the epidemic. In particular, the variable of formalism is changed from the variable of government bureaucracy to the cognition of “policy formalism” of the general public. This study intends to understand whether the gap between epidemic prevention regulations and practices affects the fear of construction site workers. A total of 743 construction site worker samples in the leisure industry were collected in this study, and the established hypotheses were tested using confirmatory factor analysis and structural equation modeling. The results of the study confirmed that during the COVID-19 outbreak, social support of site workers negatively affects anxiety; COVID-19 fear positively affects anxiety and work stress; work stress mediates the relationship between COVID-19 fear and anxiety; fear of infecting family members and fear of infecting self both positively affect anxiety; policy formalism positively affects fear of infecting family members and fear of infecting self.

1. Introduction

Environment of a Construction Site

Construction sites are considered to be places where the pandemic spreads [1]. Many types of workers are found on construction sites, including painters, electricians, structural workers, engineers, surveyors, site supervisors, clerical staff, and directors. Due to the small space and poor hygiene conditions on construction sites, the construction industry is also considered to be an industry affected by COVID-19, and workers are considered to be at risk for disease transmission [2]. Many construction projects have been terminated, postponed, changed, or (in the case of new projects) have required new methods of execution. Workers have been required to comply with social distancing policies, to prepare personal protective equipment, and sometimes to perform tasks remotely. In the first two years of the pandemic, health and protective standards were not well defined, making it challenging for the various construction disciplines to work together on site [3]. Human control and intervention affect disaster prevention on construction sites [4]. During the spread of the new coronavirus, government agencies imposed many construction regulations on construction sites, and there were changes in the workflow and wearing of anti-epidemic equipment on construction sites.

Some studies have found that the critical competence of the post-COVID-19 construction industry includes managing site safety, leadership skills, technical competence, managing supply chain disruption, and financial stability [5]. Managing site safety is considered to be the most important competency of the post-COVID-19 construction industry. Workers are fearful of being exposed to the virus [6]. The construction industry worldwide has experienced business stoppages, labor shortages, and threats to health and safety due to the COVID-19 outbreak [7]. According to Assaad and El-adaway [8], the impact of the pandemic on the construction industry includes lack of labor due to infections, work stoppages, and layoffs. The supply chain for raw materials has been inadequate, due to problems in the international market. The construction industry could not change all jobs to online meetings due to the epidemic [9]. Many jobs on a construction site require workers to be on site. The large number of workers of different nationalities on the construction site have also made group virus infection serious. Many academics have begun to focus on the safety and health of construction workers, in order to reduce the risk of labor shortages in the construction industry [10]. The economic impact of the COVID-19 pandemic has also been widely discussed, with most employees worried about job insecurity [11].

COVID-19 not only impacts the physical health of construction workers, but can also have adverse effects on their mental health. Construction contractors have found that many workers suffered from anxiety symptoms during the pandemic [12]. In fact, site workers often already showed anxiety symptoms, but the outbreak of the pandemic made them worse. Many workers have been concerned about job insecurity and health threats due to the pandemic. Construction workers have gradually experienced uncertainty, frustration, anxiety, and depression [13]. If these problems are not properly addressed, they can lead to alcohol and drug abuse issues among workers.

Anxiety and fear are considered to be different emotions [14]. COVID-19 has caused fear in the public and also increased people’s anxiety. During the pandemic, numerous unsubstantiated pieces of fake news and exaggerated information have spread across social media, increasing the level of fear during the pandemic [15]. Fear is an emotion caused by danger, pain, and harm [16], and some scholars have even developed scales to measure COVID-19 fear as a result of the COVID-19 pandemic [17]. Stress is a negative emotional response to physical, psychological, and cognitive stress. One of the causes of stress is the impact of natural disasters [18]. Risk can also increase feelings of stress when the events of everyday life are uncontrollable and unpredictable [19].

In the past, there have been problems of worker pressure and even sacrifice of worker safety in order to meet deadlines on construction sites [20]. When social distancing policies were implemented, much of the design and planning work came to a halt. After colleagues were infected or isolated, workers were faced with being shorthanded and having a higher sense of work stress. Due to the shortage of labor, it was common for workers to feel higher levels of anxiety when working overtime [21]. There was also a question of whether sites could be managed effectively in response to government inspection and regulations [22].

The uncertainty and complexity created by the outbreak of the pandemic has made it imperative for construction companies to focus on the safety and well-being of their workers, and the COVID-19 outbreak left many construction workers experiencing job burnout and feeling unmotivated to devote themselves to their profession [23]. Workers saw media reports of the outbreak every day and became afraid of contracting the disease. Past studies also found that many workers continued to experience symptoms of anxiety and depression after Hurricane Katrina in the United States [24].

COVID-19 interrupted the supply of construction materials and disrupted subcontractors’ work schedules [25]. In particular, the supply of construction materials from abroad has been delayed due to social distancing and isolation policies. Construction workers on site, lacking appropriate resources to face the risk of infection, have felt the pressure of work schedules, and have worried about exposure to their family members, especially elders and young children [26]. A study has confirmed that the infection of family members and relatives caused severe depression and anxiety [27]. The trauma and anxiety caused by the death of family members and relatives due to infection runs even deeper. Site workers may feel guilty as a result of infecting their families. Many unhealthy or asymptomatic people in the workplace can easily transmit the virus to their families. During the pandemic, the lack of vaccines for children, the closure of schools due to the pandemic, and the need to take care of family members have caused a great deal of stress for site workers [28].

The COVID-19 pandemic has also changed traditional work at the work site to working from home [29]. The isolation policy has indirectly affected workers’ mental health by eliminating opportunities to see the acquaintances they would normally see. Previous studies have pointed out that companies lacked adequate hardware should employees be required to work from home [30]. Prior to the pandemic, video conferencing software was not available or widely used, making site workers feel a different kind of work stress, and Bentley et al.’s study also noted that social isolation increases psychological stress during times of remote work [10,31]. Many site workers’ relative unfamiliarity with digital software indirectly increased the stress of working remotely. Other factors such as the quality and speed of internet connections may have also affected employee health and productivity [10].

Video conferencing technology may have also increased some workers’ work stress, which, combined with the stress of family caregiving, may have contributed to the stress felt by workers at work [32]. Previous research has confirmed that isolation policies can harm individual health [33]. Uncertain information about the pandemic and unclear isolation protocols contributed to stress disorders [33]. Recchi et al., conducted a long-term observation with a sample from France and found that a high percentage of the sample felt stress [34]. Many site office workers were unable to use certain specialized equipment during the pandemic, which affected their performance [28]. This study mainly explores whether the fear of COVID-19, personal infection, family infection, and work stress increase the anxiety of construction site workers, and whether social support of construction site workers can reduce anxiety.

One of the more significant contributions of this paper is the inclusion of formalism as a factor in on-site anti-pandemic research. Whether formalism still exists in Taiwan despite its gradual departure from the developing world has been the topic of interest of some past studies, which have confirmed that formalism still influences Taiwan’s value system [35,36]. Riggs’s use of American society as a standard for diffracted society has been discussed and questioned [37]. Riggs argues argue that although the United States is a developed and industrialized society, it is a mistake to infer that there is no formalism at all in American society. Some scholars argue that prismatic societies do not exist only in underdeveloped countries, but also exist commonly in countries with different levels of development [37]. Administrative formalism has been criticized for “absence of raw data” in the past [38]. The formalism literature in the past was a discussion of theory and qualitative data. This study is also the first to introduce policy formalism into construction site research. This study wanted to explore whether policy formalism during the epidemic affected the mental health of construction workers.

In many countries where formalism is high, civil servants are unwilling to take responsibility for the public services provided [38]. This administrative barrier to innovation and incentives to work makes civil servants intentionally evasive of responsibility. Attempts to innovate or change established systems are often obstructed by senior executives [38]. These barriers cause civil servants to become quiet, loyal, and excessively agreeable. Formalism, as proposed by Riggs, involves ritualistic methods, lack of authority, and centralization of power. As a result, there is a gap between the norms of law and its effective implementation. It also creates a gap between norms and realities. Riggs points out that, because civil servants have no pressure to implement programs and no standards for administrative performance, there is a difference between administrative ritualistic procedures and rationalistic procedures in developed countries [39]. In countries with higher levels of formalism, there is a lack of shared values at the administrative level and an inconsistency between governmental and social values [40]. The inconsistency between the regulations and the actual implementation of anti-pandemic laws can also increase the fear of infection among on-site workers. This study wants to explore whether policy formalism has affected the fear and anxiety of construction site workers during the epidemic. For the first time, policy formalism is applied to the causal model relationship of fear and anxiety among construction site workers during the epidemic.

2. Literature Review and Hypotheses Development

2.1. Policy Formalism

The administration of formalism shows that there is a gap between formal power and effective power. States with high formalism are centralized and inefficient [41]. According to Thompson, Western public administration is based on control and remains in a static state, failing to take into account rapidly changing situations [41]. Thompson also mentioned that Western administration is characterized by authority, communication, group decision making, problem solving, and high ethics [41,42]. The pandemic is an unexpected situation, and when anti-pandemic policies are not effectively implemented, it will cause site workers to become fearful when they are at work.

Burns and Stalker argue that in organizations structured under an organic model, professionals’ responsibilities and duties are not clearly defined, and they require constant interaction with colleagues, as opposed to following decisions made by supervisors [43]. In a mechanistic system, the supervisor decides whether the professionals’ work is consistent with the organization’s goals. It is difficult for a developing country to implement organizations in an organic model [44]. Organic-model organizations are designed to be more flexible, which can create anxiety for supervisors [43]. Organic systems also require a sense of trust among members, which is often lacking in developing countries [45]. Mechanistic system-oriented anti-pandemic mechanisms are undoubtedly unable to respond quickly to outbreaks of a pandemic.

Formalism first arose in government agencies; the term is primarily used to describe public officials. Formalism leaves a gap between policy implementation mandates and actual practice, and it is the focus of this paper to determine whether the public senses this during a pandemic. Past literature suggests that the public’s lack of trust in the government is mainly due to the government’s failure to meet public expectation [46]. Based on social learning theory, previous positive and negative evaluations affect future expectations. The general public’s evaluation of government trust comes from the honesty and competence of civil servants [47]. It is evident that when civil servants fail to implement policies in a pragmatic manner, site workers will also sense the formalism that permeates their governmental institutions.

The public’s evaluation of government operations reflects their evaluation of the actual administration [48,49]. Trust in the government depends on people’s confidence that civil servants will promote the public interest in the long term [50]. Numerous studies have confirmed that people’s satisfaction with government departments affects their trust [49]. When there is a gap between policy effects and policy decrees, site workers can also sense the disparity.

2.2. Hypotheses Development

Anxiety is defined as the psychological feelings of tension, stress, and worry [51]. The dimensions of anxiety encompass cognitive, emotional, and behavioral aspects, and anxiety can cause individuals to become dysfunctional [52]. Anxiety occurs when one feels a sense of lack of control over an external threat or disaster [53]. The 2019 novel coronavirus created an unprecedented disaster, and it was inevitable that site workers would feel anxious. Direct exposure to hazards, and learning about them through the media, can lead to the development of anxiety and trauma disorders [54]. The spread of COVID-19 is also likely to have caused anxiety among site workers.

Social support for workers may come from fellow workers, management, family, and friends, and is considered effective in coping with work stress and anxiety [55]. Many past studies have demonstrated that social support can help people cope with disasters and infectious diseases [56]. Social support is considered to have a positive effect when individuals face health-related stress [57,58]. The interpersonal process of social support can reduce trauma symptoms, including self-blame in the face of a disaster such as the pandemic [59]. Social support allows workers to feel trusted and assisted in the face of a pandemic [52].

Social support is considered to reduce anxiety [60]. Another study confirmed that low-to-moderate levels of social support can reduce moderate-to-high levels of anxiety [61]. Another study showed that isolation during periods of lockdown caused anxiety among individuals, which could also be improved by social support [62,63]. A large sample of medical university students in China found an inverse association between social support and anxiety [64]. Many other large samples of data have also confirmed that social support can reduce anxiety [65,66].

One of the sources of anxiety for frontline workers in the past was fear of the COVID-19 outbreak [67]. Shanafelt et al., also identified sources of anxiety among frontline workers, which include fear of viruses in the workplace, lack of testing kits and protective equipment, fear of transmitting viruses at work, and fear of being assigned to units with a higher risk of being exposed to the outbreak [68]. Site workers who lacked protective gear were afraid of being assigned to sites where the risk of contracting the virus was high.

In an Israeli study, social support was found to reduce post-war anxiety [69]. After violent attacks in schools, social support can reduce the effect of post-traumatic stress disorder in students [54]. High levels of social support have also been shown to increase coping capacity among individuals in the aftermath of floods [70]. Social support can enhance self-efficacy and professional performance, which in turn can reduce anxiety [71,72]. The availability of social support for site workers working during the pandemic can certainly reduce anxiety.

Hypothesis H1.

Social support reduces anxiety.

Early in the development of the COVID-19 outbreak, scholars began to focus on the factor of fear of infection with COVID-19 [33]. Several studies confirmed that fear of infection by COVID-19 is associated with feelings of depression and anxiety [17,73]. Studies over the past two years have also developed the Fear of COVID-19 Scale [17]. The “fear of COVID-19” here refers to a general fear, which is not the same as fear of infection in oneself or in one’s family.

The fear of COVID-19 also triggers anxiety and stress in individuals [74]. The COVID-19 pandemic has caused people to experience a number of discomforts, including working for extended periods of time at home and being unable to move about freely [75]. Some studies have also found that COVID-19 can contribute to psychological disorders [75]. Other studies suggest that daily viewing of news regarding the pandemic on television, the internet, and social media at home also contributes to COVID-19-related phobias [76,77]. Many countries adopted different bans and isolation policies during the pandemic, increasing individuals’ feelings of stress [78].

Anxiety is a natural cautionary response to immediate danger [79]. Anxiety is also a reaction to the future when one is faced with new situations and significant changes. When an individual feels an increase in heart rate and sweating, this is a common sign of anxiety. When an individual faces fearful anxiety, a discordant avoidance model develops [80,81,82]. COVID-19 is a contagious disease that has, in recent years, become relatively life-threatening to individuals, and is thus also likely to cause anxiety and depression [81,82]. The novel coronavirus began circulating in 2019, and when faced with this new, unknown disease, with an unknown transmission pathway and a lack of effective vaccines and medications, site workers were naturally anxious due to fear.

Previous studies have shown that the COVID-19 outbreak has negatively affected workers’ mental health, resulting in stress, depression, and post-traumatic stress disorder [18,83]. Work anxiety has also been produced by the fear of losing family members and colleagues [84].

Hypothesis H2.

COVID-19 Fear increases anxiety.

As shown by some scholars, the risk of infection is a source of stress for hospital workers [85,86,87]. Because of the high risk of infection, work stress and turnover rates are high, and because of the rapid transmission of the COVID-19 virus, frontline workers are often considered a high-risk group. Hotel receptionists are considered to be one of the groups with higher fear of infection, because of their role in face-to-face customer service [85]. Wong et al., found that many media reports on the rapid transmission of the virus have contributed to the fear of COVID-19 among frontline workers [88].

Work stress has also been suggested as a response to environmental threats [88]. Chen and Eyoun claim that fear of COVID-19 can contribute to emotional exhaustion among workers [89]. The increasing number of patients testing positive, inadequate personal protective equipment, increased workloads, and increased risk of disease transmission during the outbreak have all contributed to work stress [71,90,91]. Site workers have also experienced an increase in work stress due to COVID-19 fears because of the frontline nature of their work.

Previous studies on COVID-19 have shown that the negative psychological effects of infection can last for an extended period of time and even cause post-traumatic syndrome and stress [83,84,92]. Therefore, this paper proposes the following hypothesis.

Hypothesis H3.

COVID-19 Fear increases work stress.

Fear itself is a warning signal that causes behavioral changes in people to adapt to the environment [93], and in the early stages of the Ebola and MERS pandemics, fear led to an increase in anti-pandemic measures and vaccination administration [94]. However, high levels of fear can create psychological burdens, including anxiety, depression, post-traumatic stress disorder, and insomnia [93,94,95,96].

The occurrence of an emergency can alter an individual’s emotional state [97]. The COVID-19 pandemic has been shown to have negative psychological effects on many individuals, including feelings of anxiety and depression [68,98]. The pandemic has already changed workflows and working environments, including increased hours, tasks, and risks, which in turn have also increased work stress. Frontline workers are often at risk of infection and increased anxiety [99,100]. Site workers’ work stress can easily turn into anxiety as well.

The COVID-19 pandemic has caused changes in wages, shift schedules, workloads, and work stress among site workers, which has in turn affected their physical and mental health [101]. The increase in night shifts and flexible working hours has also affected the productivity of site workers [10]. Many workers at construction sites have also found that they are experiencing increased work stress [10].

Several past studies of infectious diseases have found that fear of infection increases anxiety and stress [102]. Fear of unknown viruses predisposes people to stress and anxiety, and Wang et al., found that many people in China had anxiety and stress problems at the beginning of the outbreak [103]. Anxiety and stress syndromes were also found in the general population during the earlier SARS pandemic [104]. On the one hand, workers’ fear of COVID-19 may affect anxiety directly, but on the other hand, it may also affect anxiety through the mediation of stress.

Hypothesis H4.

Work stress increases anxiety.

Hypothesis H5.

Work stress mediates the relationship between COVID-19 fear and anxiety.

For humans, fear is a natural and protective mechanism [105]. However, fear can also result in a post-stress syndrome [106]. Fear of COVID-19 infection has also created a sense of stress in many students [73]. Many frontline personnel in the pandemic have been found to have a fear of infection [107], which in turn causes stress and anxiety [108]. Psychological problems arising from COVID-19 are thought to include increased work stress, fear of infection for themselves and their families, and fear of self-infection and subsequent transmission to family members [68,109].

COVID-19 is a virus with rapid transmission and a high mortality rate, especially in the elderly and those with pre-existing medical conditions [107,110]. There is a fear of being infected and of being in contact with an infected individual. Fear of COVID-19 infection is a concern not only in areas with severe outbreaks, but also in areas with low numbers of infections [111]. There have been many studies that confirm that COVID-19 causes anxiety and depression among members of a population [106,112]. The fear of COVID-19, including illness, infection, and death, has a negative psychological impact.

In the early stages of the outbreak, a great deal of focus was on drug and vaccine development, with relatively little emphasis on the psychological and social aspects [113]. People were not only worried about being infected themselves, but were also worried about the wellbeing of their family members and friends. Frontline workers have a relatively elevated sense of fear regarding COVID-19, and are more likely to feel anxious [74]. Anxiety is also more likely to develop when family members or relatives are infected with COVID-19 [114,115]. Site workers need to work during the pandemic, and the fear of infection of self and family members in the face of the complex workforce at the site also increases anxiety.

Yıldırım et al., suggested that fear of COVID-19 is a cause of depression and anxiety [116]. Rovai et al., also pointed out that fear of COVID-19 can contribute to emotional fluctuations in humans [117]. The uncertainty of the COVID-19 pandemic has affected personal psychological and physiological health. The uncertainty of the outbreak has increased the fear of infection, and fear increases as one continues to think about negative events. Research by Ouellet et al., has pointed out that those with a lower tolerance for uncertainty are more prone to fear when faced with environmental threats. Fear includes fear of infection for oneself as well as fear of infection for family members. Increased fear of COVID-19 can lead to more information-seeking and psychological anxiety [118]. Some studies suggest that the COVID-19 pandemic has caused many people to search for additional information online, which in turn makes anxiety worse [119].

Hypothesis H6.

Fear of infecting family members increases anxiety.

Hypothesis H7.

Fear of infecting self increases anxiety.

Many civil servants in developing countries consider themselves to be independent groups, separate from the general public. Civil servants are primarily concerned with their own interests, rather than the interests of the public [38]. Formalist civil servants are less sympathetic to the people [38]. Political affiliation and seniority are the main factors for the promotion of senior civil servants [38]. Civil servants are relatively reluctant to take responsibility, mainly because of the privileged and authoritarian leadership of government departments. Bureaucracies are full of formal and legalistic structures and civil servants are reluctant to make risky decisions [38]. The arbitrary and erratic nature of political leadership leaves civil servants with a lack of self-efficacy and sense of commitment [38]. During a pandemic, when civil servants lack compassion and are unable to make effective decisions, the public naturally worries that they and their families will become infected because of the lack of implementation of anti-pandemic measures. Does formalism, which first originated in government agencies, affect public perception regarding government performance? Many studies in the past have shown that government decisions affect public opinion [48,49].

Thompson also found that developing countries have greater numbers of generalists, who place greater emphasis on hierarchy and process, forgetting their purely instrumental origin [41]. Generalist civil servants incorporate program procedures into law, but forget that they are only tools [41]. Compared to developed countries, developing countries’ actual capacity to develop professionals is relatively low [44]. Many developing countries have generalists who override specialists [120]. The average civil servant does not always have the necessary expertise, and this may cause site workers to worry about poor anti-pandemic measures and about infection of themselves and their families.

As Fox and Joiner argue, the reason for the reluctance of civil servants with high degrees of formalism to take responsibility is a lack of devotion to those they serve; Fox and Joiner also point out that civil servants may be relatively apathetic, lack clear goals, and be unwilling to make sacrifices for society [38]. Civil servants tend to put their own interests ahead of the interests of the state and society. In countries where formalism is low, innovative practices are encouraged to motivate civil servants to achieve their goals. Countries with high formalism, however, are authoritarian and work in ways that inhibit creativity [38]. Faced with an unprecedented novel coronavirus, and sensing that civil servants are unwilling to sacrifice for the common good and adopt innovative methods to achieve their goals, site workers fear for themselves and their families.

A highly formalist bureaucracy is full of pathological behavior, including lack of authority, excessive emphasis on regulation, lengthy paperwork, indifference, and fear of innovation. Due to a lack of communication, civil servants wait for orders from the higher levels and lack a sense of personal security [121]. Transitional societies are less efficient because they exaggerate the formal legal functions and in fact rely on relationships in order to function. Many informal administrative behaviors in developing countries impede the achievement of organizational goals [39].

Administrations in developing countries are full of “irrational management” [38]. Many civil servants are unwilling to make decisions, because they are not empowered to do so, and many minor issues are decided by supervisors. Administrative procedures in developing countries are complicated, and civil servants are required to administer the procedures in accordance with the law, even to the point of ignoring the spirit of the law [38]. The formalism of these legal procedures is seen as a form of self-protection for civil servants. The lack of communication, authorization, and legal administration after an outbreak are sensed by site workers as a lack of ability to respond to the crisis, and workers worry about contracting the disease themselves at the site and then passing it on to their families. Policy formalism creates a sense of anxiety about pandemic preparedness, which leads to fear of infection for themselves and their families.

Hypothesis H8.

Policy formalism increases fear of infecting family members.

Hypothesis H9.

Policy formalism increases fear of infecting self.

The research framework according to the hypothesis argument is shown in Figure 1. Policy formalism reduces fear of infecting self and family members. COVID-19 fear, fear of infecting self and family members, and work stress increase the anxiety of construction site workers. Social support reduces anxiety for construction site workers.

3. Materials and Methods

3.1. Sample, Tools, and Procedure

Taiwan’s leisure industry has suffered from the impact of the pandemic, and leisure industry operators have taken the opportunity to think about transforming their businesses. The government hopes to enhance the depth and quality of the tourism industry, and also hopes that operators will learn new digital skills to create win–win–win tourism benefits. In this study, workers at construction sites in the leisure industry were used as the subject population; 743 valid samples were obtained through a convenience sampling method, with a recovery rate of 49.53%. Regarding interviewees’ basic information, 52.2% of the respondents were male. In terms of occupation, directors accounted for 3.0%, engineers accounted for 3.5%, supervisors accounted for 3.2%, administrators accounted for 18.0%, technical staff accounted for 14.7%, and workers accounted for 57.6% (see

Table 1. Sample basic information.

Gender	Percentage (%)	Seniority	Percentage (%)
Male	52.2%	1–3 years	33.9%
Female	47.8%	4–7 years	15.9%
Age	Percentage (%)	8–11 years	13.3%
20–29 years old	24.9%	12–15 years	5.7%
30–39 years old	42.7%	16 years or above	31.2%
40–49 years old	9.0%	Marriage	Percentage (%)
50 years old or older	23.4%	Unmarried	31.1%
Occupation	Percentage (%)	Married	63.8%
Director	3.0%	other	5.1%
Engineer	3.5%	Number of children	Percentage (%)
Supervisor	3.2%	1	12.5%
Administrator	18.0%	2	35.1%
Technical staff	14.7%	3	15.3%
Worker	57.6%	4 or more	5.8%
		none	31.2%

). This study regards construction site workers as people who participate in construction site work, so it includes director, engineer, supervisor, administrator, technical staff, and worker. All participants working on the construction site were affected by the epidemic regulations. In terms of marriage status, 63.8% were married. In regard to the number of children, 12.5% had one child, and respondents with more than 2 children accounted for 40.9% of all samples. Almost all unmarried construction site workers had no children.

3.2. Measures

The sources for the scales used in this study were as follows: “Work stress”, modified from the scale developed by Shukla and Srivastava [122]. Item examples: I have a lot of work and I’m afraid I won’t have time to do it because of the pandemic. I feel overwhelmed with work and it seems like a bad day not to be at work. I feel that the epidemic has accumulated a lot of work, and there is almost no time to ask for leave. “COVID-19 fear”, modified from the scale used by Ahorsu et al. [17]. Item examples: I am afraid of running into someone infected with the new coronavirus at work. Thinking about workers contracting the new coronavirus makes me uncomfortable. My hands get wet when I think about the new coronavirus. “Anxiety”, referenced from the scale provided by Hamilton [123]. Item examples: I make worst-case predictions about future jobsites. I feel nervous and tired easily when working on the construction site. When I’m working on a construction site, I feel uneasy. “Fear of infecting self” and “fear of infecting family members”, modified from those used in Tasso et al., and Spatafora et al. [124,125]. Item examples: I am worried that my family member will be infected with the new crown virus and become a severe patient. I am worried that my family members will be infected. I am worried that I have contracted the new coronavirus. I am worried that I will be infected with the new crown virus and become a severe patient.

Based on the formalism definitions and previous scales [35,36], this study developed the following items: I believe that the anti-pandemic regulations and the actual enforcement at the site will not be exactly the same. I think it is sometimes difficult to enforce the anti-pandemic laws on the site. I think that many anti-pandemic programs are not easily implemented. I think there are differences between the anti-pandemic regulations and the current state of on-site practice. The Cronbach’s alpha values for the reliability of all factors in the study ranged from 0.90 to 0.96 (

Table 2. Item loading and model fits.

Variables	Items	Lambda	z Values	Composite Reliability	Cronbach’s Alpha
Policy formalism	Policy formalism 1	0.88	--	0.95	0.95
	Policy formalism 2	0.91	98.9
	Policy formalism 3	0.96	99.3
	Policy formalism 4	0.90	95.2
Social support	Social support 1	0.82	--	0.94	0.94
	Social support 2	0.93	51
	Social support 3	0.86	51.1
	Social support 4	0.92	51.6
	Social support 5	0.81	48.8
Work stress	Work stress 1	0.92	--	0.95	0.94
	Work stress 2	0.91	181.7
	Work stress 3	0.88	186.8
	Work stress 4	0.85	172.9
	Work stress 5	0.93	188.6
	Work stress 6	0.67	132.6
Anxiety	Anxiety 1	0.78	--	0.91	0.90
	Anxiety 2	0.84	154.0
	Anxiety 3	0.89	169.1
	Anxiety4	0.86	170.1
COVID-19 fear	COVID-19 fear 1	0.79	--	0.92	0.92
	COVID-19 fear 2	0.80	150.9
	COVID-19 fear 3	0.80	153.7
	COVID-19 fear 4	0.90	162.5
	COVID-19 fear 5	0.88	165.7
	COVID-19 fear 6	0.70	152.3
Fear of infecting family members	Fear of infecting family members 1	0.80	--	0.92	0.91
	Fear of infecting family members 2	0.88	134.2
	Fear of infecting family members 3	0.92	138.8
	Fear of infecting family members 4	0.81	131.4
	Fear of infecting family members 5	0.69	111.8
Fear of infecting self	Fear of infecting self 1	0.92	--	0.96	0.96
	Fear of infecting self 2	0.92	201.5
	Fear of infecting self 3	0.93	201.1
	Fear of infecting self 4	0.94	199.6
	Fear of infecting self 5	0.83	166.8

Note: The first item of each variable is set to 1, so there is no Z value.

), which is higher than the minimum standard of reliability set by Nunnally [126], i.e., 0.60.

3.3. Controlling for Common Method Variance (CMV)

Common method variance (CMV) is considered to be a kind of variance that results from the measurement method; it causes an internal consistency error and must be controlled for [127,128].

The questionnaires in this study were all self-administered, which may have created common method variance problems. To ensure that the results of the study were not affected too much by common method variance, the questionnaires were administered anonymously, using a mixed Likert 5–7-point scale to reduce the problem of common method variance [127]. In addition, the questionnaire was designed to meet certain operational procedures and standards, and the item design was based on the principles of simplicity and ease of understanding; any questions that were confusing to the respondents, would lead to different interpretations, or were difficult to answer were avoided.

For post-testing, Harman’s one-factor test was used [129]. The explained variance of the exploratory factor analysis’ first principal component without rotation was only 45.73%, which was not excessively high, confirming that the common method variance problem was not significant in this study.

4. Results

4.1. Validity and Reliability Analysis

The reliability and validity of the constructs and items were examined by confirmatory factor analysis (CFA) using SEM (structural equation modelling) software. SEM provides a method to deal with measurement errors, using multiple indicators to reflect potential variables, and also makes estimating the relationship between the factors of the entire model more accurate and reasonable compared to traditional regression methods. In terms of the model overall fit measures, the SRMR of the conceptual model was 0.069; this was slightly higher than the judgment criterion of 0.05, yet nevertheless, it was within an acceptable range. The GFI was 0.99, which was higher than 0.90. The NNFI was 0.98, NFI was 0.98, CFI was 0.98, IFI was 0.98, and RFI was 0.98, all of which were higher than the judgment criterion of 0.90, indicating that the hypothetical model was acceptable. In terms of model parsimonious fit measures, PNFI was 0.89, and PGFI was 0.80, both higher than the standard of 0.50. These all show that the conceptual model of this study is appropriate and the model fits the empirical data, which also confirms the overall construct validity of this study.

Secondly, the factor loading λ values of all the constructs ranged from 0.67 to 0.96, all of which were higher than 0.5; this is in line with the recommendations of Hair, Anderson, Tatham, and Black (>0.5) [130], indicating that the individual items in this study had an acceptable level of reliability. All of the item loading t-values in this study reached statistically significant levels, partially confirming the construct validity and convergent validity of the constructs in this study.

The composite reliability (CR) of the latent constructs measures the consistency of the variables within the constructs. According to Hair et al., the CR value should be greater than 0.7 [130]. The CR values of the latent constructs in this study ranged from 0.91 to 0.96, all greater than 0.7, indicating that the latent constructs in this study had favorable internal consistency.

The average variance extraction (AVE) represents the percentage of potential constructs that can be measured for the observed questions, and can be used to determine not only reliability, but also discriminant validity and convergent validity. According to Fornell and Larcker [131], an AVE value greater than 0.5 indicates constructs with convergent validity. The AVE values of the potential variables in this study ranged from 0.66 to 0.84, all greater than 0.5, indicating that the potential variables in this study had good discriminant and convergent validity.

In general, the square root of the average variance extracted (AVE) of individual constructs should be greater than the correlation coefficient between the construct and other constructs in the model, indicating discriminant validity [132]. The table below presents the matrix of correlation coefficients between the constructs, and the diagonal line is the square root of the constructs’ AVE. The square roots of the constructs’ AVEs in this study ranged from 0.81 to 0.92(see

Table 3. Square root of AVE and intercorrelations.

	1	2	3	4	5	6	7	ASV	MSV	AVE
Policy formalism (1)	(0.91)							0.11	0.27	0.83
COVID-19 fear (2)	0.34	(0.81)						0.26	0.49	0.66
Fear of infecting family members (3)	0.52	0.54	(0.84)					0.21	0.33	0.70
Fear of infecting self (4)	0.35	0.70	0.57	(0.92)				0.26	0.49	0.84
Anxiety (5)	0.22	0.59	0.42	0.61	(0.85)			0.23	0.40	0.72
Social support (6)	0.13	0.05	0.26	0.07	−0.15	(0.87)		0.02	0.07	0.75
Work stress (7)	0.30	0.54	0.35	0.52	0.63	0.01	(0.88)	0.20	0.40	0.77

Note: The figures in parentheses indicate the square root of AVE of the study constructs. MSV = maximum share variance; ASV = average share variance.

), which were larger than the correlation coefficients between any two constructs. In addition, the AVEs were also larger than the MSV and ASV, which also confirms the discriminant validity of this paper [130].

The following correlation matrix shows the preliminary relationship between the constructs in this study. Policy formalism and fear of infecting family members and fear of infecting self were positively correlated, with coefficients of 0.52 and 0.35. Anxiety was positively correlated with COVID-19 fear, fear of infecting family members, fear of infecting self, and work stress, with coefficients of 0.59, 0.42, 0.61, and 0.63, respectively. The constructs of this study were all in line with initial expectations.

4.2. Path Coefficients

In this study, the structural equation modeling (SEM) path coefficient analysis was used to test the developed hypotheses [133].

Table 4. Path coefficients (coefficients, STDEV, Z Statistics).

Hypotheses	Causal Path	Coefficients	Standard Deviation	z Statistics	Accept or Reject
H1	Social support->Anxiety	−0.28 ***	0.03	−11.37	accepted
H2	COVID-19 fear->Anxiety	0.18 ***	0.04	4.65	accepted
H3	COVID-19 fear->Work stress	0.80 ***	0.06	19.88	accepted
H4	Work stress ->Anxiety	0.60 ***	0.03	13.72	accepted
H6	Fear of infecting family members->Anxiety	0.07 **	0.02	2.92	accepted
H7	Fear of infecting self->Anxiety	0.26 ***	0.02	10.88	accepted
H8	Policy formalism->Fear of infecting family members	0.60 ***	0.04	16.27	accepted
H9	Policy formalism->Fear of infecting self	0.50 ***	0.05	13.70	accepted

Note: ** and *** represent statistical significance at p < 0.01 and p < 0.001, respectively.

shows that social support of site workers negatively affects anxiety, with a path coefficient of −0.28; this validates H1. This means that the more social support felt by site workers, the lower their anxiety will be. When site workers feel COVID-19 fear, their work stress and anxiety increase. COVID-19 fear positively affects anxiety and work stress with coefficients of 0.18 and 0.80. Such coefficients validate H2 and H3. When site workers experience COVID-19 fear, their work stress and anxiety also increase. Many studies in the past have confirmed that social support can reduce anxiety [60,61,62,63]. This study also confirmed that the social support of construction site workers will reduce their anxiety. The fear of COVID-19 is due to its rapid spread and high mortality rate. This study also confirmed that personal and family infections increase the anxiety of construction site workers [114,115].

The relationships between COVID-19 fear and work stress, work stress and anxiety, and COVID-19 fear and anxiety all reached statistically significant levels, and met the mediation criterion proposed by Baron and Kenny [134]. The indirect effect of COVID-19 fear on anxiety was 0.28, the direct effect was 0.28, and the mediation percentages were 50.3 and 49.7 (see

Table 5. Mediation estimates.

Path Estimates	Label	Estimate	SE	z	p	% Mediation
COVID-19 fear->Work stress	a	0.79	0.03	28.8	<0.001
Work stress->Anxiety	b	0.36	0.02	17.9	<0.001
COVID-19 fear->Anxiety	c	0.28	0.02	12.8	<0.001
Mediation Estimates
Indirect	a × b	0.28	0.02	15.2	<0.001	50.3
Direct	c	0.28	0.02	12.8	<0.001	49.7
Total	c + a × b	0.56	0.02	31.4	<0.001	100.0

), respectively; this verified H4. This demonstrates that COVID-19 fear among site workers, mediated by work stress, affects anxiety. The risk of infection has been a predictor of work stress [85,86,87]. Fear of COVID-19 and work stress will increase anxiety [10,103]. This study confirmed that the work stress of construction site workers mediates the relationship between COVID-19 fear and anxiety.

Fear of infecting family members and fear of infecting self both positively affected anxiety, with coefficients of 0.07 and 0.26, respectively, verifying H6 and H7. Site workers’ fear of infecting self and fear of infecting family members increases their anxiety, i.e., workers working at the site will see increased anxiety if they are worried about infecting themselves and their family members as a result of the virus. Fear of personal and family infection increases anxiety [68,109], especially worry about relatives in the family with old age and pre-existing diseases [107,110].

Policy formalism positively affected fear of infecting family members and fear of infecting self, with coefficients of 0.60 and 0.50; this verifies H8 and H9. Site workers feel that anti-pandemic measures are difficult to execute at a work site; many anti-pandemic systems are not easily implemented. When there is a discrepancy between the anti-pandemic regulations and the status quo in site implementation, it makes them worry about infecting themselves and their family members. Civil servants with high formalism have low self-efficacy and decision-making ability [38]. The inefficiency of civil servants’ decision-making and implementation increases the gap between regulations and practices of epidemic prevention policies. During the COVID-19 epidemic, the implementation of epidemic prevention policies by generalists will increase workers’ fear of infection [41]. Generalists are not as specialized in epidemic prevention policies as specialists. The gap between regulations and practices on epidemic prevention makes workers feel the fear of being infected.

5. Discussion

The present study validated all the established hypotheses, using structural equation modeling (SEM) path coefficient analysis. First, this study confirmed that social support negatively affects anxiety, which is similar to many previous studies [62,63,66]. Site workers who have the support of their fellow workers, management, family members, and friends can effectively cope with work stress and anxiety, making workers more capable of coping with disasters and disease infections. A study has pointed out that social support can help people cope with disasters and infectious diseases [56]. The rapid spread of COVID-19 is terrifying construction workers. Taking Israel as an example, previous studies have confirmed that social support reduces post-war anxiety [69]. Social support can enhance coping skills and reduce anxiety caused by floods [70].

This study also confirmed that fear of COVID-19 can cause anxiety when felt by construction workers. The construction industry is an industry considered to be affected by COVID-19, and construction workers are susceptible to COVID-19 [2]. Because of the limited space and poor hygiene conditions at construction sites, the virus is easily transmitted. Work stress is also increased by the fear of exposure to viruses [71,90,91]. Fear of infection with COVID-19 has become a variable in post-epidemic research. [33]. Past studies have found that fear of COVID-19 can cause personal anxiety and stress [74]. The COVID-19 pandemic has changed the shifts, workload, and workflow of workers on the construction site, thus increasing their work stress [10]. This study further examined whether work stress mediates the relationship between COVID-19 fear and anxiety. COVID-19 fear affects anxiety through work stress, and can also affect anxiety directly.

Site workers working at construction sites are not only fearful of contracting the virus themselves, but also of the same happening to their family members, which leads to anxiety [74,115]. The uncertainty of the pandemic gives rise to considerable fear in workers, which in turn generates anxiety. The spread of the new coronavirus has caused construction site workers to fear that they and their families will be infected [68,109]. The transmission rate of COVID-19 is fast, and the mortality rate of the elderly and those with pre-existing diseases is high [107,110]. Fear of infection among individuals and family members increases anxiety [114,115].

More specifically, this paper confirms that policy formalism positively affects fear of infecting family members and fear of infecting self. Site workers find many anti-pandemic systems difficult to implement. When there is a discrepancy between the anti-pandemic regulations and the status quo in site implementation, site workers worry about infecting themselves and their family members. Construction site workers with high awareness of formalism believe that the epidemic prevention policy will not be effectively implemented [41]. They believe that civil servants only pay attention to compliance with laws and regulations, and do not care whether the epidemic prevention is effective. The government’s epidemic prevention is full of irrational management [38], causing construction site workers to worry that they and their families will be infected.

6. Conclusions

6.1. Theoretical Implications

The major theoretical contribution of this study is the application of formalism to the pandemic, verifying that construction site workers’ cognition of policy formalism affects their fear of infection. Formalism, as mentioned by Riggs, is primarily used with reference to government agencies; however, on the basis of Mizrahi, Vigoda-Gadot, and Cohen’s suggestion that people can sense public policies [49], this study applies formalism to site workers’ perceptions of pandemic preparedness during an outbreak. In the past, formalism was always considered to occur in underdeveloped societies, but scholars have suggested that formalism can also occur in developed societies, only with the development of formalism varying in its extent [37]. In Taiwan, which has gradually moved toward being a developed society, but with an absence of performance appraisals or market competition, formalism still exists for civil servants; e.g., civil servants’ service performance evaluations still rotate between grade A and B, more or less regardless of actual performance.

This study examines the feelings of site workers towards the formalism in anti-disease policies during the critical time of the spread of the pandemic. When workers believe that the anti-pandemic policies will not be implemented and there is a discrepancy between the anti-pandemic regulations on paper and their actual implementation, this increases workers’ fear of infection for themselves and their family members, which leads to increased anxiety. This study confirms that workers’ perceptions of formalism based on their past perceptions of public policy will increase their doubts about the effective implementation of anti-pandemic policies, leading to fear and anxiety.

Government agencies with high formalism have many formal and legalistic structures, causing civil servants to adopt conservative decision-making [38]. When civil servants lack self-efficacy and a sense of commitment, policy formalism is more likely to occur [38]. The formalism of the epidemic prevention policy cause construction site workers to worry about themselves and family members being infected. Anti-epidemic work requires professionals, but government agencies are full of generalists. The epidemic prevention of generalists causes fear in construction site workers that they and their families will be infected. Civil servants with high formalism are apathetic, lack clear goals, and are unwilling to make sacrifices for society [38]. The pathological behavior of civil servant including excessive emphasis on regulation, lengthy paperwork, indifference, and fear of innovation [121], as well as the formalistic attributes of civil servants have made construction site workers fearful of COVID-19.

6.2. Practical Implications

For site workers, social distancing policies should be implemented in a way that maintains connections with supervisors and co-workers, so as to enhance social support. It is important to communicate properly with supervisors and co-workers regarding the difficulties of working during the pandemic, in order to find solutions. Such social support can reduce anxiety.

It is important to properly identify the sources of workplace stress, and find solutions with work stakeholders; it is important to discover how to handle the increasing work demands during the pandemic with increasingly limited resources. We must access credible information about the pandemic, stay away from untruthful social networks, and understand how to protect ourselves and our families from infection.

For construction companies, it is important to establish two-way communication channels with employees and build worker trust by letting site workers know that the company cares about their situation. By building resilience skills, workers can reduce fear and stress relating to the outbreak, which in turn reduces anxiety. Mindfulness programs can be developed for workers to help them cope with the stress of the outbreak. Site supervisors need to be more empathetic and increase their interactions with and understanding of their workers. Organizations must strengthen workers’ career wellbeing and help workers understand the long-term value of the construction industry. Through the outbreak, organizations need to learn how to re-establish mentor systems, so that workers can continue to communicate with their mentors.

7. Further Study

In the past, the literature on formalism has been qualitative in nature, and little quantitative research has been conducted. There is limited literature to refer to, and more research is needed to verify these findings in the future. In this study, we could not conduct a longitudinal study because of time and cost considerations; we were only able to use a single sample to investigate our hypothesis. The sample size of 743 in this study may not be very small, but it is expected to be further validated by more studies in the future. This study used self-reported data, which may have issues such as selective memory, telescoping effects, and attribution, and we hope that additional research designs will be added in the future. This study focuses on Taiwan, and there may be cultural differences and biases, so we hope that there will be additional regional studies in the future that can support the findings of this study.