Risk perception, barriers and working safely with silica dust in construction. A psychological network approach

doi:10.21203/rs.3.rs-4842185/v1

Download PDF

Research Article

Risk perception, barriers and working safely with silica dust in construction. A psychological network approach

https://doi.org/10.21203/rs.3.rs-4842185/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

In the construction industry workers are frequently exposed to hazardous substances. To explore and describe construction workers’ barriers and motives to (not) work safely with hazardous substances, we examined their perspectives on the health risks, perceived barriers and intention to use preventive measures with regard to silica dust. Specifically, we studied perspectives on the use of face masks, dust collection on power tools, and using a vacuum instead of a broom.

Method

Semi-structured interviews (n = 13) and a pen and paper survey (n = 187) were administered on construction and training sites. Only executive workers could participate in the study. We approached the behaviour of using specific preventive measures as an emergent property of a complex network of interacting psychological variables. To analyze the structure of these ‘behavioural decision networks’ we applied a psychological network.

Results

Construction workers were generally aware their health is at risk from working with silica dust. At the same time, they are not very concerned about that risk. Network analysis suggests that concern does play a moderate role in the behavioural decision networks, suggesting that a lack of concern may encourage unsafe behaviour. Construction workers’ level of automaticity to use specific preventive measures was relatively low. Barriers to use preventive measures such as time and effort play a relatively important role in the networks. A general intention to work safely hardly played any role in the networks, while a specific intention to use preventive measures played a more prominent role. Age and work experience did not play a role in the network. Non-parametric tests and descriptive comparison of networks suggest differences in for example the relative importance of specific variables.

Conclusions

For two preventive measures, different variables may be more successful intervention points to foster safe work. Increasing levels of concern, improving automaticity of use, addressing specific intention to use preventive measures in risk communication, and offering preventive measures at time and location where relevant tasks are performed, are discussed as possible intervention points to foster working safely with silica dust in construction. Future studies should further substantiate these findings.

Risk perception

Occupational Health and Safety

Construction

Silica dust

Preventive measures

risk communication

behaviour

Psychological Network Analysis

In the Netherlands, it is estimated that each year around 3000 people die due to exposure to hazardous substances at the workplace [1, 2]. An even greater number of people become ill. Construction is an industry in which workers are frequently exposed to hazardous substances. It is broadly acknowledged there is a need for knowledge that can help prevent and reduce occupational diseases from exposure to hazardous substances. Effective prevention partly lies with the decisions of construction workers to comply with or use preventive and protective measures. To form a better understanding of construction workers’ barriers and motives to (not) work safely with hazardous substances, we examined their perspectives on the health risks, the barriers they experience to use preventive measures, their intention to use preventive measures, and their intention to work safe in general with regard to Respirable Crystalline Silica (CAS 14808-60-7), also known as silica dust.

Several health behavioural models and theories can be used as a starting point for studying preventive behaviours by construction workers. Three prominent examples are the Theory of Planned Behaviour [TPB, 3, 4], the Health Belief Model [HBM, 5] and the Protection Motivation Theory [PMT, 6]. Many core concepts in these models are similar. For example, these models all include perceived vulnerability, severity of consequences and self-efficacy as core concepts to explain behaviour. However, some differences exist. In the current study we use the PMT as a starting point for our research because it is the only model that also accounts for perspectives on the perceived effectiveness of risk reduction measures (i.e. response efficacy). In the current study, we approach PMT as a descriptive model and extend the core concepts (perceived vulnerability, perceived severity, self-efficacy and response efficacy) with context specific factors relevant to risk perception of chemical substances in construction. We elaborate on these factors in more detail below.

Risk perception, as used here, is considered a psychological process that comprises of complex interactions among values, attitudes, (mis)beliefs and feelings associated with the risk at hand. At the same time, risk perception also comprises evaluative outcomes such as perceived severity or perceived (personal) vulnerability. Risk perception hereby holds cognitive and affective elements that may serve as drivers and barriers for working safely [7, 8]. Knowledge on key aspects in workers’ risk perception of chemical substances may thus hold key insights that can help to prevent and reduce substance-related occupational diseases. For example, research has shown that a lack of experience with negative consequences of unsafe work can cause someone to believe that they themselves are less likely to experience a negative event [9, 10, see also 11]. Additionally, studies have shown that when workers underestimate occupational risks, they are more likely to conduct risky behaviours [12, 13].

Most studies on hazard and risk perception and the use of preventive measures in the construction industry study perception of safety risks, like falling from height, cuts and wounds and handling heavy loads [9, see e.g., 14–17]. Zhang and Fang [16] for example studied the reasons why Chinese scaffolders do not use fall harnesses. They concluded that underestimating the risk of not using safety harnesses, the inconvenience and discomfort of using safety harnesses, negative pressures from people in exemplary positions (e.g. foremen and safety officers), and lack of safety lines are causes of scaffolders deciding not to use safety harnesses.

Few studies have examined workers' perceptions of occupational health risks from chemical substances in the workplace. Studies that do, provide several leads on how risk perception and use of preventive measures are associated. For example, Antonucci et al. [18] showed in a sample of Italian construction workers that risk perception is higher regarding hazards that can cause immediate injury, as opposed to those that can cause occupational illnesses. If workers underestimated the probability of experiencing adverse effects from exposure, this could lead to ineffective use of personal protective equipment (PPE) and may even foster risky behaviour [19, 20].

In a qualitative study, Hambach and colleagues [21] explored worker perceptions of chemical risks, to find leads for a workplace health program. Their sample consisted of workers who worked in paint production, cleaning and maintenance or in production of surface-active agents industries. Hambach and colleagues found that workers partly rely on their senses to assess risks [see also 22–24], often accept risks associated with their job, and mistrust prevention advisors. In another study by Pettersson-Strömbäck and colleagues [25] workers from the plastic industry were interviewed to study their mental models (i.e. their collections of beliefs, attitudes, feelings, etc. on a particular topic) of exposure to chemical substances. These mental models were compared to actual exposure data. The results showed that workers tend to estimate exposure based on their work activities, which generally resulted in an underestimation of their exposure compared to what the actual exposure data showed. In a study on the relationship between noise hazard perception of miners and their use of earplugs, Shkembi, Smith, and Neitzel [20] also found that about one in three workers underestimated their exposure to noise when their true exposure was hazardous. While not statistically significant, misperception of exposure to hazardous noise was associated with higher odds of not using earplugs.

[26] concluded that workers are more likely to adopt self-protective behaviours if they perceive their health has been negatively affected by the use of pesticides. In another study also providing insight in the use of protective measures, Stege and colleagues [27] applied the mental models approach [see also 25, 28]. They found that mental models of (the risks of) particulate matter of road workers differed from the mental models of scientific experts on the topic. Results further showed that while workers are aware of the existence of particulate matter and ‘reduction methods’, their knowledge is often incomplete and they do not protect themselves consistently against particulate matter.

Further insights from public risk perception of chemicals provide additional leads on how risk perception factors may serve as drivers and barriers for working safely. For example, studies have shown that non-risk experts tend to use values, norms, attitudes and psychometric aspects such as perceived familiarity with the risk, perceived controllability of exposure, and the delayed occurrence of adverse effects to evaluate risks [22, 24, 29–33].

Taken together, the above described literature suggests that a combination of many factors likely influences construction workers’ behaviour when it comes to working with chemicals. An integrative study looking at a combination of these factors can help further efforts in the prevention and reduction of substance-related occupational diseases.

Current study

The current study presents an explorative and descriptive account of the risk perception of constructions workers and the perceived drivers and barriers towards the use of preventive measures in regard to exposure to chemical substances in the construction industry. To describe construction workers’ risk perceptions and barriers to working safely we follow the Causal Attitude Network (CAN) model [34]. The CAN model conceptualizes attitudes (e.g. an overall evaluation of an event or behaviour) as psychological networks of evaluative elements (e.g. beliefs, feelings and behaviours) that interact causally and can be understood through a network topology. This means that beliefs and feelings that are more closely related will be arranged more closely in the network structure. Complex systems like the CAN have emergent properties that are not explained by understanding the individual elements of the system [35, 36]. Following this, we approach the actual use of preventive measures (i.e. safe work behaviour) as an emergent property of a system of interacting (psychological) variables [see also 37, 38]. In this conceptualization, specific patterns of interacting variables within the system (e.g risk perceptions, perceived barriers, intentions) give rise to (un)safe work behaviour. These patterns can be presented in a network topology. Figure 1 presents a schematic representation of a network topology. For this study we will refer to the networks of interaction variables as behavioural decision networks.

We investigate how risk perception, perceived drivers and barriers are related to the (intention to) use three specific types of preventive measures: using a face mask, using dust collection on power equipment, and using a vacuum instead of a broom when cleaning to prevent exposure to silica dust. Second, we investigate whether differences can observed between these three types of preventive measures.

In Europe and the Netherlands, silica dust is considered one of the most prominent health hazards for construction workers [39, 40]. Silica dust can cause severe lung problems and is classified by the IARC as a group 1 carcinogen, meaning there is definite proof that it can cause cancer in humans [41]. It is estimated that in the European Union around 5 million workers are exposed to crystalline silica dust [40]. To our knowledge the current study is one of the first studies [cf. 42] on psychological drivers and barriers and the use of protective behaviours related to silica dust. It is also the first study to focus on safe behaviour in the construction industry using a complexity perspective.

We recruited construction companies, and visited construction sites to administrate (voluntary) semi-structured interviews. Based on the findings from the interviews and from a literature review, we developed a questionnaire and administered it to construction workers. We did this face to face at construction sites of the participating companies. The procedure is described in more detail below.

Recruitment and participant criteria

Recruitment of companies was done through a non-profit knowledge and advice centre for the construction and infrastructure sector (Volandis) and through a public appeal to safety experts working at construction companies to contribute to a study about ‘risk perception in the construction industry’. After companies agreed to contribute, researchers accompanied safety experts to construction sites to recruit workers on site. Only executive workers such as carpenters, ground workers and electricians were eligible to participate in the study. Workers were informed that participation to the study was anonymous and on a voluntary basis. Workers were eligible to participate in the study if they were 16 years or older, and were exposed to silica dust through their own work activities (e.g. cutting, sawing, grinding, and/or drilling bricks, concrete, or cement) or could be exposed through the work activities of others (e.g. through silica dust that in general is present at the construction site or has become airborne due to sweeping a work site). All participants informed consent before participating in the study.

Qualitative exploratory study and questionnaire design

A qualitative exploratory study was used to identify relevant themes for the questionnaire. In March 2022, thirteen (including two pilot) semi-structured interviews on working with silica dust were administrated to construction workers. All participants were male. Mean age was 42.9 years, mean years of work experience was 23.5 years. All participants had a lower or middle vocational technical degree. All participants had a Dutch cultural background, except for one who identified as Moroccan-Dutch.

Interviews followed a protocol that was designed to cover core concepts from the Protection Motivation Theory [6, 43]. Questions aimed to elicit relevant beliefs and feelings associated with perceived vulnerability, perceived severity, self-efficacy and response efficacy. After the interview, participants were debriefed and were provided with information on who to contact in case the interview caused future health concerns. Based on interview results, several topics relevant for workers’ judgement and decision making regarding working with silica dust were identified. These will be elaborated upon more in the main study and measures section below.

Main study and measures

To ensure data quality we aimed to design a questionnaire that was as little cognitively demanding as possible. Frequently mentioned topics in the interviews were converted to corresponding psychological constructs in the questionnaire. Based on the pilot study and identified constructs from the literature, a structured questionnaire was designed, covering factors associated with the perception of the hazard, exposure, and risk, and factors associated with the use of preventive measures. An overview of constructs and (example) items is presented the supplementary materials. All items were measured using a five point Likert scale: Completely disagree (1) – Disagree (2) – Do not disagree, do not agree (3) – Agree (4) – Completely agree (5). Before answering questions, participants were asked whether they had ever heard of silica dust. Subsequently to align ideas, a short description of our understanding of silica dust was given. It said “Silica dust is the name given to very small particles of silica. Silica can be found in building materials such as bricks, plasterboards and concrete. If you work with building materials that hold silica, silica dust can come off. When we refer to silica dust in the questionnaire, we refer to these particles.”

Risk perception was measured in terms of vulnerability in light of the PMT. Both cognitive (probability of experiencing effects) and affective (worry about experiencing adverse effects) dimensions were measured. Familiarity was measured using two items; “Working with silica dust is normal to me” and “I don’t know any different than working with silica dust” (α = .60). Items were summed and averaged for the analyses. For all factors associated with preventive measures we focused on three measures that can reduce or prevent inhalation of silica dust:

using a face mask,
using dust collection on power equipment, and
using a vacuum cleaner instead of a broom.

Data was collected separately (within subjects) for each of the three measures and included intention to use preventive measures; self-efficacy; response efficacy; automaticity of using protective equipment and barriers to using protective equipment. Constructs and (example) items are presented in Table 1. Automaticity of using protective equipment was initially measured using the four item automaticity index [44] based on the Self Report Habit Index [45]. We reformulated “I start doing before realizing I’m doing it” to “I have to think about before doing it” with the aim to accommodate for lower language skills. However, based on reliability analysis this item was subsequently dropped. The remaining three items were summed and averaged. Reliability was generally good (Face mask α = .73; Dust collection α = .83; Vacuuming α = .83). Through the interviews we identified five barriers to using preventive measures (PM): Not having PM at hand when needed; using PM costs too much time; using PM costs more efforts than it yields; the work environment complicates the use of PM; and the use of PM complicates the work itself. All barriers were included as items in the questionnaire (example items in Table 1)

Table 1

Questionnaire constructs, number of items and (example) items
Factor		Items	(Example) item
	General
Age		1	How old are you?
Work experience		1	How many years have you been working in construction?
	Factors associated with the perception of the hazard, exposure and risk
Familiarity		2	“Working with silica dust is normal to me”
Perceived amount of exposure		1	“During my work I inhale a lot of silica dust”
Control over exposure		1	“I have control over whether I inhale silica dust or not”
General intention to work safely		1	“I try to inhale as little silica dust as possible”
Perceived vulnerability: Cognitive		1	“There is a probability I will get sick due to exposure to silica dust”
Perceived vulnerability: Affective		1	“I worry about the possibility I can get sick due to exposure to silica dust”
Risk is part of the job		1	“The fact you can get sick from silica dust is part of my job”
	Factors associated with the use of preventive measures
Intention to use preventive measures		1	“I always use dust collection on power tools when I should” (emphasis in the questionnaire)
Self-efficacy		1	“Vacuuming instead of brooming is easy to do”^a
Response efficacy		1	“Using a face mask ensures I do not inhale silica dust”^a
Automaticity to use preventive measures		3	“[…] I do without thinking about it” ^a
Barrier: Have on hand		1	“[…] I have on hand when I need it” ^{a, b}
Barrier: Complicates work		1	“[…] sometimes complicates my work” ^{a, c}
Barrier: Time		1	“[…] takes too much time to use” ^{a, c}
Barrier: Work environment		1	“[…] I sometimes cannot use because of my work” ^{a, c}
Barrier: Effort		1	“[…] takes more effort than it yields” ^{a, c}
a. Items were asked separately for using a face mask, using dust collection on power equipment, and using a vacuum cleaner instead of a broom.
b. Note that lower score indicate experiencing the barrier.
c. Note that higher score indicate experiencing the barrier.

Data analysis

All data was analysed using R (version 4.3.1). Non-parametric comparisons for the three separate preventive measures (multiple dependent groups) were conducted using the Friedman test and Wilcoxon signed-rank (WSR) test. These tests are needed to identify possible intervention points in the behavioural decision networks. For example, if a specific potential barrier has a central position in the network, and at the same is not necessarily experienced as one (i.e. with a low mean score), it may not be the most effective intervention point to promote preventive behaviour. Furthermore, the best intervention points to promote preventive behaviour may differ between preventive measures (e.g. using a vacuum instead of a broom may require more effort than putting on a face mask).

Subsequently, network analyses were performed. We estimated behavioural decision networks of using a face mask, using dust collection on power tools, and using a vacuum instead of a broom. In network analysis, variables included in the network topology are called ‘nodes’. The nodes represent the different (psychological) factors included in the analyses (see Table 1). The undirected connections between nodes are called ‘edges’. The edges represent the mutual dependence of the nodes, given the presence of all other nodes in the network (cf. partial correlations). Thicker edges or higher ‘edge weights’ (i.e. the lines) between nodes mean they are more strongly connected (see Fig. 1).

Network estimation were conducted using mgm (version 1.2–14) [46]. We estimated three Gaussian Graphical Models from the continuous data. The mgm package was used to compute and include node predictability in the network visualisations. Respondents were removed from the data set when they did not fit our respondent inclusion criterium (i.e. only executive workers) or when there were signs of, for example, straight lining in their responses. We applied multiple imputation to missing data because data were missing at random and listwise deletion is not an option in mgm. This way, loss of data and bias (e.g. compared to applying case wise deletion) are minimized. Networks are estimated using regularization. In this procedure very small edges (i.e. very low partial correlations) are set to zero. As such, regularization ensures that the estimated network is sparse and the probability of false positive edges is low. Regularization parameters were selected using cross-validation (CV).

An important metric in network analysis is node centrality. Central nodes in the network have more and stronger relations to the other nodes and may be important intervention targets to change the network structure. Centrality of nodes was based on node strength, which is computed based on the number and thickness of edges of each node with its neighbouring nodes.

Nodewise predictability was added as a complementary analysis [47]. Nodewise predictability represents the proportion of variance in that node that can be explained by its connected neighbouring nodes. It can help identify secondary intervention targets within the network (e.g. when the primary variable of interest cannot easily be intervened on). Nodewise predictability is calculated in terms of explained variance (R²).

Analyses of stability and accuracy of the networks were performed using Bootnet [48]. To assess stability and accuracy, bootstrapped confidence intervals (CI) of edge weights were estimated. As data was regularized, non-parametric bootstrap was applied [48, 49]. The correlational stability coefficient (CSC) was calculated to assess stability of the centrality indices (strength). The CSC indicates the proportion of cases that can be dropped at random (creating a new partial network, with different centrality indices) while retaining with 95% certainty a .70 correlation with the order of the strength centralities indices in the original network. A CSC above .25 is recommended. A CSC of 0.50 or above is indicative of good robustness of the original network [48].

A Network Comparison Test (NCT) was not conducted to statistically test for differences between networks since the NCT has not been validated for within subjects data [50]. Networks are therefore only compared at a descriptive level. Networks were plotted using the qgraph (version 1.9.8) [51]. Network layout was plotted using the average layout for all three networks. A priori defined groups of variables were classified for the network visualization (see Fig. 2).

Participants

In total, 8 companies specialized in civil construction work (e.g. bridges and roadways) or housing construction agreed to contribute to the study. Questionnaire data was collected at 24 building sites and training locations. In total 187 questionnaires were included in this study. Characteristics of the sample population included in the main study are presented in Table 2.

Table 2

Characteristics of the sample population in the main study
Variable
Sample size	187
Gender (% male)	100
Highest finished education (N, %)
1 (low)	79 (42.2%)
2 (medium)	87 (46.5%)
3 (high)	12 (6.4%)
Age range (years)	17–63
Age mean (standard deviation)	46.2 (11.4)
Experience in construction work (years)	23.7 (12.5)
Note. Education level according to CBS classification.

Percentages highest finished education do not add up to 100 due to missing values (9).

Table 3

Means (M), standard deviations (SD) and non-parametric test results.
	Factors associated with the perception of the hazard, exposure and risk
Item		Valid responses	M (SD)
Familiarity		187	2.95 (.97)
Perceived amount of exposure		185	2.89 (1.01)
Control over exposure		184	3.23 (1.19)
General intention to work safely		186	4.02 (0.90)
Perceived vulnerability (cognitive)		186	3.51 (.90)
Perceived vulnerability (affective)		186	2.89 (1.07)
Risk is part of the job		186	2.11 (1.04)
	Factors associated with the use of preventive measures
		Valid responses	M(SD)
			Face mask	Dust Collection	Vacuuming	Friedman test
Behavioural intention
[…] I always use when I should		187; 187; 186	3.22 (1.08)	3.33 (1.06)	3.22 (1.01)	χ2(2) = 3.653, p = 0.161
Response efficacy
[…] makes sure I don’t breath in silica dust		186; 187; 187	3.51 (1.00)	3.36 (1.05)	3.63 (1.00)	χ2(2) = 9.966, p = 0.007
Self-efficacy
[…] is easy to use		186; 187; 187	3.33 (1.05)	3.38 (1.04)	3.57 (1.01)	χ2(2) = 10.151, p = 0.006
Automaticity of use (scale)
Scale total average		186;183;186	2.82 (0.81)	3.11 (.86)	2.97 (.86)	χ2(2) = 18.719, p < 0.001
Barriers
[…] I have at hand when I need it (on hand)		187; 187; 187	2.73 (1.08)	2.46 (1.02)	2.93 (1.16)	χ2(2) = 23.451, p < 0.001
[…] Costs too much time to use (time)		187; 187; 187	2.34 (.97)	2.55 (1.05)	2.58 (.96)	χ2(2) = 6.843, p = 0.033
[…] Takes more effort than it yields (effort)		186; 186; 187	2.20 (.93)	2.37 (.97)	2.40 (.93)	χ2(2) = 6.497, p = 0.039
[…] I sometimes cannot use due to my work (environment)		187; 185; 187	2.31 (1.05)	2.79 (1.20)	2.75 (1.14)	χ2(2) = 20.728, p < 0.001
[…] Sometimes complicates my work (complicates)		187; 187; 187	2.97 (1.18)	3.11 (1.20)	2.87 (1.09)	χ2(2) = 9.375, p = 0.009
Note. Higher scores indicate stronger intentions, stronger beliefs on the respective items (e.g. stronger beliefs of vulnerability or stronger perceived barriers)

Means, standard deviations and non-parametric test results

Table 3 presents the means and standard deviations and non-parametric test results on drivers and barriers of using a face mask, dust collection on power tools, and using a vacuum instead of a broom (vacuuming).. The table also includes an overview of non-parametric test results (Friedman test) of comparisons between drivers and barriers of using a face mask; dust collection on power tools; and using a vacuum instead of a broom (vacuuming).

The Friedman test showed no significant differences in intention to use face masks, dust collection or vacuuming. Face masks are not found to be significantly more easy to use than vacuuming (χ2(2) = 3.653, p = 0.161). For all other drivers and barriers, statistically significant differences between preventive measures were found (see Table X). These differences are further outlined below.

Response efficacy. The use of face masks is not considered more effective in preventing inhalation of silica dust than using dust collection on power tools (Z = -1.804, p = .071) or vacuuming (Z = -1.485, p = .137). Vacuuming is considered more effective than dust collection on power tools (Z = -3.086, p = .002)

Self-efficacy. Using a face mask is considered more easy to do than vacuuming (Z= -2.781, p = .005) but not statistically significantly more easy to do than using dust collection on power tools (Z = − .852, p = .417). Using dust collection on power tools is considered more easy to do than vacuuming (Z = -2.050, p = .040)

Automaticity of behaviour. Using dust collection on power tools is more automatically done than using face masks (Z= -3.394, p < .001) and vacuuming (Z = -2.555, p = .011). Vacuuming is more automatically done than using face masks (Z= -2.454, p = .014).

Barrier: On hand. Dust collection on power tools is more often on hand than a face mask (Z= -3.321, p < .001), and more often on hand than a vacuum (Z= -4.949, p < .001). A face mask is not more often on hand than a vacuum (Z= -1.867, p = .062).

Barrier: Time. Using dust collection on power tools is more strongly considered to take too much time than using a face mask (Z= -2.109, p = .035), but not more than vacuuming (Z = − .487, p = .627). Vacuuming is more strongly considered to take too much time to use than using a face mask (Z= -3.043, p = .002).

Barrier: Effort benefit trade-off. Using dust collection on power tools is considered more purposeful than using a face mask (Z= -2.100, p = .036), but not less purposeful than vacuuming (Z = − .558 p = .577). Vacuuming is considered more purposeful than using a face mask (Z= -2.363, p = .018).

Barrier: Work environment. Using dust collection on power tools is more affected by the job that needs to be done than using a face mask (Z= -5.008, p < .001), but not more than vacuuming (Z = − .569, p = .569). Vacuuming is more affected by the job that needs to be done than using a face (Z= -4.418, p < .001).

Barrier: Complicates work. Using dust collection on power tools complicates work more than vacuuming (Z= -2.417, p < .016), but not more than using a face mask (Z= -1.627, p = .423). Vacuuming does not complicate work more than using a face mask (Z = − .802, p = .423).

Network analyses

Figure 2 shows the behavioural decision networks of using a face mask (2a), using dust collection on power tools (2b), and using a vacuum instead of a broom (2c). Recall that the actual use of face mask; dust collection on power tools, and a vacuum instead of a broom are considered behaviours that are emergent of a (favourable) network topology. Strength centrality for each node (i.e. an estimation of how strongly a given node is connected to other nodes in the network), is presented in Fig. 3. Stability analysis of the strength centrality hierarchy showed good robustness for all networks (face mask CS-coefficient = .44; dust collection CS-coefficient = .52; vacuuming CS-coefficient = .67).

Overall, results of the network analysis suggest that risk perception, in terms of perceived vulnerability, plays a moderately central role (compared to other nodes) in the network. The affective evaluation (worry), as compared to the cognitive evaluation (probability), held a more central position (see Fig. 3). Perceived vulnerability (both cognitive and affective) is mostly indirectly and distantly associated with the intention to use preventive measures (see Fig. 2). Perceived controllability of exposure is the third most central node in the networks of using a face mask and vacuuming. It is a moderately central node in the network of using dust collection (see Fig. 3). Perceived familiarity with the hazard was the least central node in all networks.

Perceived barriers to using preventive measures play a relatively central role in all networks. The order of the most central barriers and the relative centrality (compared to other nodes) differ between networks (see Fig. 3). In all networks the barriers effort and time have a direct positive association. Meaning that when a preventive measure is perceived as ‘takes more effort that it yields’ it is generally also perceived as taking too much time. Work environment complicating the use of preventive measures and using preventive measures complicating work itself also share a direct positive relationship in all networks. In all networks not having the preventive measure on hand has a direct negative association with the intention to use the preventive measure, and the automaticity to use preventive measure (see Fig. 2). Also in all networks, automaticity (to use preventive measures) has a direct positive association with the intention to use the preventive measures. Furthermore, intention to use preventive measures, was the 10th, 1st and 3rd most central node for using a face mask, dust collection or a vacuum respectively. In contrast, the general intention to work safely is one of the least central nodes in the networks, being second-last in dust collection and face mask networks, and third-last in vacuuming network. In all networks the nodes age and work experience are strongly associated with each other, but disconnected from other nodes in the networks (see Fig. 2). Because age and work experience are logically dependent [47] we will not further elaborate on these variables .

Table 4 present the results of the predictability analysis in terms of explained variance (R²). The predictability is depicted in Fig. 2 as a circle around each node. The degree to which this circle is filled (grey) represents the proportion of variance in that node that can be explained by its connected neighbouring nodes. Predictability of nodes varies between networks.

Table 4

Explained variance (R²) of nodes in the face mask, dust collection and vacuuming networks.
		R²			Range (R²)
		Face mask	Dust collection	Vacuuming
	Node
General
1	Age	.71	.72	.72	.71 –. 72
2	Work experience	.71	.71	.71	.71
Factors associated with the perception of the hazard, exposure and risk
9	Familiarity	.00	.06	.00	.00 –.06
6	Perceived exposure	.18	.24	.14	.14 – .24
8	Perceived control over exposure	.33	.23	.28	.23 – .33
5	General intention to work safely	.10	.11	.08	.08 – .10
3	Perceived vulnerability (cognitive)	.23	.26	.25	.23 –.26
4	Perceived vulnerability (affective)	.33	.33	.28	.28 – .33
7	Risk part of job	.21	.26	.21	.21 – .26
Factors associated with the use of face mask, dust collection or a vacuum
13	Intention to use preventive measure	.41	.40	.67	.40 –.67
10	Response efficacy	.22	.18	.10	.10 – .22
11	Self-efficacy	.37	.40	.27	.27 – .40
18	Automaticity to use preventive measure	.38	.47	.55	.38 – .55
12	B: On hand	.41	.38	.50	.38 – .50
14	B: Time	.48	.39	.41	.39 – .48
15	B: Effort	.41	.44	.40	.40 – .44
16	B: Environment	.50	.48	.43	.43 – .50
17	B: Complicates work	.30	.48	.39	.30 – .48

The aim of this study was to present an explorative and descriptive account of risk perception, perceived barriers, and the use of three preventive measures aimed at reducing or preventing exposure to chemical substances in the construction industry. To do so, we outlined how construction workers perceive the hazard, exposure and risk (e.g. familiarity or perceived exposure and worry) of silica dust. Also, we outlined perceived barriers, efficacy beliefs, and intentions regarding the use of face mask, dust collection or a vacuum (e.g. perceived effort, response efficacy and intention to use the preventive measure) to prevent exposure to silica dust. Finally, we applied a psychological network approach and studied the interdependence of all these factors to study potential intervention points to promote preventive behaviour with regard to working with silica dust. Applying this approach provided several insights that can help foster the prevention and reduction of substance-related occupational diseases in construction. These insights are discussed in further detail below.

It is important to note that in this study and approach, safe behaviour is understood as an emergent property of a complex system of interacting variables (e.g. risk perceptions, efficacy beliefs, perceived barriers). Factors that are not directly associated with the intention to use preventive measures can thus still have an important role in the decision to use preventive measures. Our study hereby also provides an illustrative account of intention-behaviour gaps in the context of risk behaviours in construction [52].

Risk perception of (working with) silica dust

The current study showed that construction workers are generally inclined to belief there is a chance they will experience adverse health effects from exposure to silica dust during their work (perceived vulnerability – cognitive). At the same time they are less likely to worry about this probability (perceived vulnerability – affective). In the behavioural decision networks this perceived vulnerability played a moderate role. Worry, as compared to the perceived chance, had a more central role, which indicates that concerns about the health risks likely play a more important role in the decision making network than the perceived chance of experiencing health effects. Given that a lack of worry about the risks of exposure to silica dust can contribute to more unsafe behaviours [12, 13], these findings suggest that the current lack of concern about the personal risk of silica exposure may encourage unsafe behaviour.

In a qualitative study, Hambach and colleagues [21] concluded that workers often accept chemical risks associated with the job. However, in the current study construction workers mostly did not agree that the risks associated with exposure to silica dust are to be considered “part of the job”. While we did not specifically ask for acceptance (the item in the questionnaire read “The fact you can get sick from silica dust is part of my job”) this finding seems to contrast the qualitative findings by Hambach and colleagues [21]. Furthermore, in the different decision networks the importance of this belief varied. That is, the belief played the most central role (based on many weaker edges to other factors in the network) in the decision network of using dust collection on power tools. In this network, only the perceived barrier that the use of dust collection on power tools complicates work itself held a more central role (having a strong edge with the perceived barrier that using dust collection on power tools complicates the work itself). In the other networks it held a more moderate role.

Psychometric studies of risk perceptions have shown that familiarity with a hazard and controllability of exposure can affect perceptions of severity [30] and may therefore affect motivation for preventive behaviour [6]. Results of the current study also suggest that familiarity with silica dust hardly plays any role in the decision networks of using preventive equipment and therefore likely do not affect the use of preventive behaviour. In line with the psychometric studies, the current study showed that constructions workers are likely to agree they can control exposure to silica dust. Furthermore, the network analyses showed that in the decision networks of using a face mask and using a vacuum instead of a broom, controllability was one of the most central factors. Therefore, the current study suggests that emphasizing control construction workers have on exposure to silica dust may indeed positively affect the use of preventive measures [cf. 6].

Perceived barriers to working safely with silica dust

The barriers to working safely included in this study were brought forward by construction workers in the qualitative pre study and therefore hold strong ecological validity. Study results indicate that similar barriers (e.g. time vs. effort, and working environment complicating work vs. preventive measure complicating work) are linked most strongly. Partial (but incomplete) explained variance indicates that these perceived barriers overlap but also play a unique role in the decision network. For example, barriers time and effort are relatively strongly linked in the behavioural decision networks, but variance in each variable is only partly explained by making it separate possible points for intervention. Importantly, all barriers (while in different orders of relative importance) play an important role in the behavioural decision networks for using preventive measures. Therefore, in order to foster the use of preventive measures, efforts should be taken to lower these perceived barriers.

Some barriers may not easily be addressed by behavioural interventions. For example, carpenters may not be able to use dust collection on power tools when they have to drill holes in tight spots (i.e. the work environment complicates work). Others do present promising points of departure. For example, in all behavioural decision networks, not having the preventive measures at hand, is directly negatively associated with automaticity to use preventive measures, and the intention to use preventive measures. To overcome this barrier, efforts can be made to offer the measure close to the location where relevant tasks (e.g. drilling in concrete) are performed. For example, construction workers could be encouraged to carry a face mask in high exposure environments or for high exposure tasks (provided this does not obstruct them in any way). Similarly, dust collection for power tools, when not installed by default (or taken off by workers), should preferably be installed by default [see e.g., 53], and should otherwise be offered at specific times and places when and where construction workers frequently work with silica containing materials. In the context of using a vacuum instead of a broom, efforts could be made to offer a vacuum at specific times when construction workers clean work sites. This will reduce time and effort in terms of walking to a specific distribution point. It will of course not be able to reduce the extra effort (and extra time) of using the vacuum. Offering the preventive measures at times and places when it is relevant for choice and behaviour of the construction worker likely also lowers the barriers ‘time’ and ‘effort’.

Intention to work safely with silica dust

The current study provided several other interesting insights into intentions to work safely. First, results showed that constructions workers generally (completely) agreed they try to avoid exposure as much as possible (i.e. general intention to work safely). In contrast, they did not necessarily agree that they always use preventive measures when they should. This conflicting result holds for all three preventive measures. Interestingly, further analyses showed that general intention to work safely hardly played any role in the behavioural decision networks, while specific intention to use preventive measures played a far more prominent role.

Construction companies frequently use motto’s like “work safely or do not work at all” to emphasize the importance of working safely. While this motto establishes an important overarching value and long term goal (also important in promoting work safety), findings of the current study suggest that promoting this general intention in risk communication strategies is unlikely to contribute to construction workers’ use of preventive measures. To promote specific behaviours that contribute to these long term goals, risk communication efforts should therefore also focus on addressing the specific intention to use preventive measures. These findings and interpretations are in line with Goal Setting Theory and the Construal Level Theory (CLT) of psychological distance, which state that setting specific near goals (e.g. I am taking a face mask with me tomorrow) are more effective for behavioural change than vague, abstract distant goals (e.g. I am going to avoid silica dust as much as possible) [see e.g., 54, 55].

Results of the current study also showed that fostering automaticity with which construction workers use preventive measures can also contribute to safe behaviour by construction workers [cf. 56]. That is, construction workers’ level of automaticity to use face masks, dust collection and a vacuum instead of a broom was relatively low (all below the scale average). At the same time network analyses showed that automaticity to use preventive measures played an important role in the behavioural decision networks of using a vacuum instead of a broom and using dust collection on power tools. In all cases, automaticity had a strong direct positive link with the intention to use the preventive measure (see also explained variance). This suggest that improving automaticity may contribute to the use of preventive measures, both directly through increasing intention to use the measure and as part of the behavioural decision network as a whole. Improving automaticity to use preventive measures, as part van (new) habit formation, is a slow process [57]. Frequency of behaviour, satisfaction, monitoring and rewarding can be important factors in habit formation [56, 57]. Reducing time and effort to use preventive measures (where possible) may boost the frequency with which workers preventive measures. Secondly, assuring positive affect or satisfaction with using preventive measures can help. For example, making salient what amount of dust has not been inhaled due to using preventive measures such as face masks. Thirdly, monitoring and rewarding (i.e. positive reinforcement) consistent use of specific preventive measures over time. For example, site supervisors (or other designated persons) could randomly walk around a building site at times during the day and monitor use of preventive measures among individuals. Consistent use could be rewarded in some forms.

Finally, results from all network analyses showed that neither age nor work experience played a prominent role in the behavioural decision networks. While the centrality measures for both variables were high, this can be explained because the nodes are logically dependent [47] and therefore strongly connected with one another. Relationships with other nodes of the networks were mostly missing. This suggests that age or work experience is not likely to be an effective intervention point for to foster safe work behaviour when working with silica dust.

Strengths and limitations

A strength of this study is its complex systems (psychological network) approach. To our knowledge this study is the first application of the psychological network to study safe behaviour in the construction industry using a complexity perspective. This approach makes it possible to illustrate that (and: how) construction workers risk perceptions, decisions, and behaviours regarding working with silica dust are part of a complex system of interacting (psychological) variables. Another strength of this study is that it builds on a qualitative pre study, in which relevant themes in construction workers decisions and behaviour were brought forward by the construction workers themselves. This qualitative foundation of a psychological network analysis strengthens the ecological validity and relevance of the quantitative results of the study.

A limitation of this study is the lack of a statistical test to compare the (within subjects) behavioural decision networks. However, the Network Comparison Test [58], used to compare psychological networks, has not been validated for within-subjects data. For this reason we chose to omit these test results. Instead we relied on a descriptive comparison. It should be noted that the application of psychological network analysis to the domain of risk perception research is still under development. Methodology needs to be developed further to substantiate findings using this approach.

The relatively small sample size for network analysis could be considered another limitation of the study. However, data met the analysis requirements and stability analysis indicated sufficiently robust results. Finally, what may be considered a limitation of the study are the undirected associations between study variables. The explorative nature of this study should be taken into account in the interpretation of the results, especially in terms of possible target factors (i.e. nodes) for interventions to improve safe working practices.

The current study addressed a need for knowledge that can help prevent and reduce occupational diseases from exposure to hazardous substances in practice. We did so by providing several possible points of departure to foster the use of preventive equipment by construction workers when working with silica dust. We investigated how risk perception, perceived barriers to working safely, and intentions to work safely were related to the use of specific preventive measures. Secondly, we aimed to investigate whether differences can observed between different types of preventive measures. We described several differences in the behavioural decision networks for using a face mask, using dust collection on power tools and using a vacuum instead of a broom. Increasing levels of concern, improving automaticity of use, addressing specific intention to use preventive measures in risk communication, and offering preventive measures at time and location where relevant tasks are performed, were discussed as possible points of departure to foster working safely with silica dust in construction. Future studies should further substantiate current explorative findings.

Ethics approval and consent to participate

This study was carried out in accordance with the principles of the Declaration of Helsinki. The study was registered (VLH-544) at the study register for human-related research at the Dutch National Institute for Public Health and the Environment (RIVM). The study design was assessed and it was judged that the research was exempted for further approval by the ethical research committee. An additional ethicist was consulted for the drafting of information letters and post-research consultation for participants regarding possible health effects of long term work with silica dust. All participants received information on the study goals and participants’ rights before providing written consent to participate in the study.

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request and will be made publicly available after finalizing a second manuscript related to the dataset.

Competing interests

The authors declare no competing interests.

Funding

This research was carried out in the framework of RIVM Strategic Programme [grant S/080004/01/PR], in which expertise and innovative projects prepare RIVM to respond to future issues in health and sustainability.

Authors’ contributions

TJ and ML developed the theoretical and conceptual framework of the research project. TJ, MvdB, EvM, ML collected and reviewed empirical literature for the study. TJ, ML, MvdB and EvM conceptualized and developed the interview protocol and survey instrument. TJ and MvdB collected and analyzed the qualitative data from the pilot study. MvdB and TJ collected the data from the main study. TJ and GL conceptualized the data analysis methodology. TJ and GL analyzed quantitative data from the main study. TJ conceptualized and wrote the manuscript. ML, MvdB, GL and EvM reviewed the manuscript.

Acknowledgements

We would like to thank prof. dr. Frenk van Harreveld and dr. Sanne de Wit for their contribution to the conceptualization of the research project. We would like to thank all participating construction companies, safety experts, site supervisors and executive workers for their time and effort. Without them, this study was not possible.

LEXCES. Why are we here? Landelijk Expertise Centrum Stoffengerelateerde Beroepsziekten. 2021. www.lexces.nl/en/about-lexces. Accessed 30 Jul 2024.
Netherlands Labour Authority. Annual report 2021. 2022.
Bosnjak M, Ajzen I, Schmidt P. The Theory of Planned Behavior: Selected Recent Advances and Applications. Eur J Psychol Assess. 2020;16:352–6.
Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:179–211.
Janz NK, Becker MH. The Health Belief Model: a decade later. Health Educ Q. 1984;11:1–47.
Floyd DL, Prentice-Dunn S, Rogers RW. A meta-analysis of research on protection motivation theory. J Appl Soc Psychol. 2000;30:407–29.
Loewenstein GF, Weber EU, Hsee CK, Welch N. Risk as feelings. Psychol Bull. 2001;127:267–86.
Slovic P, Finucane ML, Peters E, MacGregor DG. Risk as analysis and risk as feelings. Decision Research. 2002.
Namian M, Albert A, Zuluaga CM, Behm M. Role of safety training: Impact on hazard recognition and safety risk perception. J Constr Eng Manag. 2016;142:04016073.
Geller ES. Working safe: How to help people actively care for health and safety. CRC Press; 2017.
Tversky A, Kahneman D. Judgment under Uncertainty: Heuristics and Biases. Science. 1974;185:1124–31.
Tixier AJ-P, Hallowell MR, Albert A, van Boven L, Kleiner BM. Psychological antecedents of risk-taking behavior in construction. J Constr Eng Manage. 2014;140:04014052.
Patel DA, Jha KN. Neural Network Model for the Prediction of Safe Work Behavior in Construction Projects. J Constr Eng Manage. 2015;141:04014066.
Chaswa EN, Kosamu IBM, Kumwenda S, Utembe W. Risk Perception and Its Influencing Factors among Construction Workers in Malawi. Safety. 2020;6:33.
Hallowell M. Safety risk perception in construction companies in the Pacific Northwest of the USA. Constr Manage Econ. 2010;28:403–13.
Zhang M, Fang D. A cognitive analysis of why Chinese scaffolders do not use safety harnesses in construction. Constr Manage Econ. 2013;31:207–22.
Xiang Q, Ye G, Liu Y, Miang Goh Y, Wang D, He T. Cognitive mechanism of construction workers’ unsafe behavior: A systematic review. Saf Sci. 2023;159:106037.
Antonucci A, Siciliano E, Ladiana D, Boscolo P, Di Sivo M. Perception of occupational risk by rural workers in an area of central Italy. J Biol Regul Homeost Agents. 2012;26:439–45.
Bye R, Lamvik GM. Professional culture and risk perception: Coping with danger on board small fishing boats and offshore service vessels. Reliab Eng Syst Saf. 2007;92:1756–63.
Shkembi A, Smith LM, Neitzel RL. Risk perception or hazard perception? Examining misperceptions of miners’ personal exposures to noise. Int J Hyg Environ Health. 2023;254:114263.
Hambach R, Mairiaux P, François G, Braeckman L, Balsat A, Van Hal G, et al. Workers’ perception of chemical risks: a focus group study. Risk Anal. 2011;31:335–42.
Kraus N, Malmfors T, Slovic P. Intuitive Toxicology: Expert and Lay Judgments of Chemical Risks. Risk Anal. 1992;12:215–32.
Slovic P, Malmfors T, Krewski D, Mertz CK, Neil N, Bartlett S. Intuitive Toxicology. II. Expert and Lay Judgments of Chemical Risks in Canada. Risk Anal. 1995;15:661–75.
Jansen T, Claassen L, van Kamp I, Timmermans DR. “It is not entirely healthy.” A qualitative study into public appraisals of uncertain risks of chemical substances in food. Public Underst Sci. 2020;29:139–56.
Pettersson-Strömbäck AE, Liljelind IE, Nordin S, Järvholm B. Workers’ mental models of chemical exposure in the workplace. Risk Anal. 2010;30:488–500.
Remoundou K, Brennan M, Sacchettini G, Panzone L, Butler-Ellis MC, Capri E, et al. Perceptions of pesticides exposure risks by operators, workers, residents and bystanders in Greece, Italy and the UK. Sci Total Environ. 2015;505:1082–92.
Stege TAM, Bolte JFB, Claassen L, Timmermans DRM. Particulate matter exposure in roadwork companies: A mental models study on work safety. Saf Sci. 2019;120:137–45.
Granger Morgan M, Fischhoff B, Bostrom A, Atman CJ. Risk Communication: A Mental Models Approach. Cambridge: Cambridge University Press; 2002.
Jansen T, Claassen L, van Kamp I, Timmermans DRM. “All chemical substances are harmful.” public appraisal of uncertain risks of food additives and contaminants. Food Chem Toxicol. 2020;136:110959.
Slovic P. The Perception of Risk. Earthscan; 2000.
Siegrist M, Árvai J. Risk Perception: Reflections on 40 Years of Research. Risk Anal. 2020;40:2191–206.
Bearth A, Saleh R, Siegrist M. Lay-people’s knowledge about toxicology and its principles in eight European countries. Food Chem Toxicol. 2019;131:110560.
Bearth A, Kwon S, Siegrist M. Chemophobia and knowledge of toxicological principles in South-Korea: perceptions of trace chemicals in consumer products. J Toxicol Environ Health A. 2021;84:183–95.
Dalege J, Borsboom D, van Harreveld F, van den Berg H, Conner M, van der Maas HLJ. Toward a formalized account of attitudes: The Causal Attitude Network (CAN) model. Psychol Rev. 2016;123:2–22.
Luke DA, Stamatakis KA. Systems science methods in public health: dynamics, networks, and agents. Annu Rev Public Health. 2012;33:357–76.
Mitchell M. Complexity: A Guided Tour. Oxford University Press; 2009.
Mkhitaryan S, Crutzen R, Steenaart E, de Vries NK. Network approach in health behavior research: how can we explore new questions? Health Psychol Behav Med. 2019;7:362–84.
Yang J, Ye G, Xiang Q, Kim M, Liu Q, Yue H. Insights into the mechanism of construction workers’ unsafe behaviors from an individual perspective. Saf Sci. 2021.
Ministry of Social Affairs and Employment. Factsheet Kwartsstof. 2018. https://www.arboportaal.nl/onderwerpen/kwartsstof/wat-zijn-de-risicos-van-werken-met-kwartsstof. Accessed 19 Oct 2023.
EU-OSHA. The Facts on Silica dust. Roadmap on carcinogens. 2017. https://roadmaponcarcinogens.eu/silicadust. Accessed 19 Oct 2023.
IARC. List of classifications. 2016. http://monographs.iarc.fr/ENG/Classification/index.php. Accessed 24 Jun 2016.
Falk L, Bozek P, Ceolin L, Levitsky M, Malik O, Patel J, et al. Reducing agate dust exposure in Khambhat, India: Protective practices, barriers, and opportunities. J Occup Health. 2019;61:442–52.
Rogers RW. A Protection Motivation Theory of Fear Appeals and Attitude Change1. J Psychol. 1975;91:93–114.
Gardner B, Abraham C, Lally P, de Bruijn G-J. Towards parsimony in habit measurement: testing the convergent and predictive validity of an automaticity subscale of the Self-Report Habit Index. Int J Behav Nutr Phys Act. 2012;9:102.
Verplanken B, Orbell S. Reflections on past behavior: A self-report index of habit Strength1. J Appl Soc Psychol. 2003;33:1313–30.
Haslbeck JMB, Waldorp LJ. mgm: Estimating Time-Varying Mixed Graphical Models in High-Dimensional Data. J Stat Softw. 2020;93:1–46.
Haslbeck JMB, Waldorp LJ. How well do network models predict observations? On the importance of predictability in network models. Behav Res Methods. 2018;50:853–61.
Epskamp S, Borsboom D, Fried EI. Estimating psychological networks and their accuracy: A tutorial paper. Behav Res Methods. 2018;50:195–212.
Fried IE, Epskamp S, Veenman M, van Borkulo CD. Network stability, Comparison, and Replicability. In: Isvoranu A, Epskamp S, Waldorp LJ, Borsboom D, editors. Network psychometrics with R. New York, NY: Routledge Member of the Taylor and Francis Group; 2022. p. 133–54.
van Borkulo C. A tutorial on R package NetworkComparisonTest (NCT). In: Symptom network models in depression research: From methodological exploration to clinical application. unknown; 2018. p. 249–57.
Epskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D, Others. qgraph: Network visualizations of relationships in psychometric data. J Stat Softw. 2012;48:1–18.
Sheeran P, Webb TL. The intention-behavior gap. Soc Personal Psychol Compass. 2016;10:503–18.
Johnson EJ, Shu SB, Dellaert BGC, Fox C, Goldstein DG, Häubl G, et al. Beyond nudges: Tools of a choice architecture. Mark Lett. 2012;23:487–504.
Locke EA, Latham GP. New Directions in Goal-Setting Theory. Curr Dir Psychol Sci. 2006;15:265–8.
Trope Y, Liberman N. Construal-level theory of psychological distance. Psychol Rev. 2010;117:440–63.
Verplanken B, Orbell S. Attitudes, Habits, and Behavior Change. Annu Rev Psychol. 2022;73:327–52.
Lally P, Gardner B. Promoting habit formation. Health Psychol Rev. 2013;7:S137–58.
Van Borkulo CD, Epskamp S, Milner A. Network comparison test: permutation-based test of differences in strength of networks. 2015.

No competing interests reported.

SilicadustRCode.r

Download PDF

Editorial decision: Revision requested
05 Aug, 2024
Editor assigned by journal
05 Aug, 2024
Submission checks completed at journal
02 Aug, 2024
First submitted to journal
01 Aug, 2024

You are reading this latest preprint version

Risk perception, barriers and working safely with silica dust in construction. A psychological network approach

Status:

Version 1

Abstract

Background

Method

Results

Conclusions

Figures

Background

Current study

Methods

Recruitment and participant criteria

Qualitative exploratory study and questionnaire design

Main study and measures

Data analysis

Results

Participants

Means, standard deviations and non-parametric test results

Network analyses

Discussion

Risk perception of (working with) silica dust

Perceived barriers to working safely with silica dust

Intention to work safely with silica dust

Strengths and limitations

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1