Decent work in Italy: A network analysis

Abstract

The present study investigated the Italian version of the Decent Work Scale (DWS) with a sample of 645 workers (females = 65.1%; mean age = 43.9 years; SD = 10.9) according to a network perspective. We compared factorial and network models and estimated the regularized partial correlations for the five DWS domains: physically and interpersonally safe working conditions (SC), access to healthcare (AH), adequate compensation (AC), hours that allow for free time and rest (FT) and organizational values complement family and social values (CV). The results highlighted that the network model showed the best fit to the data. Among the most central domains in the network, the high centrality of CV suggests that this domain could represent an effective target for actions addressed to fostering decent work in Italy. In contrast, the low centrality of AC and FT underlines the urgent need to advocate for more decent remunerations and working hours in Italy.

Keywords

Decent Work Decent Work Scale Italian Version Network Analysis Psychology of Working Theory

Introduction

The acceleration of technological processes (Rosa et al., 2017), the fluid characteristics of society (Bauman, 2013) and globalization (Sparks et al., 2001) describe the contemporary world of work (Guichard, 2009; Savickas, 2021; Savickas & Porfeli, 2012). This scenario has increased instability in working and living conditions (Blustein et al., 2019a, 2019b, 2019c; Cartwright & Cooper, 2014; Peiró & Tetrick, 2011). Furthermore, a series of systemic breakdowns (i.e. the 11 September 2001 attacks in the United States; the global financial crisis of 2007–2008; the COVID-19 pandemic and subsequent recession) have exacerbated this trend, increasing the prevalence of temporary, precarious and low-wage work (Allan et al., 2021; Autin et al., 2020; Blustein et al., 2020; Blustein & Guarino, 2020; Johnson et al., 2020; Kniffin et al., 2021). At the same time, these phenomena have deepened the structural inequalities of the labor market, with marginalized individuals across the socioeconomic spectrum often struggling to find jobs that secure basic human needs (Blustein et al., 2020; Duffy et al., 2019).

The Italian context is not immune to such events (Di Fabio & Kenny, 2019a). Unemployment represents an ongoing concern in Italy, with employment trends worsening more than the European Union (EU) average (EUROSTAT, 2021a). In 2021, Italy's employment rate was 62.6%, five percentage points below the Italian national employment target (EUROSTAT, 2021a). The latest available statistics on the Italian labor force have highlighted that 22.0% (13,174,000) of the resident population is inactive (ISTAT, 2020). Among them, 77.6% (10,230,000) are not seeking a job or are not available for work (ISTAT, 2020). Notably, the employment gap between men and women is high (19.9%), and the rate of employment across young people is low (16.8%), representing a further burden for the Italian world of work (EUROSTAT, 2021a). Moreover, the Italian employment structure has changed rapidly in the last 10 years (Camussi et al., 2021). Occupational trends highlight a skew toward low-quality and low-paying jobs in contrast to better trends observed in other EU countries (Camussi et al., 2021). These trends, accompanied by increasing poverty, represent another compelling concern (ISTAT, 2020). In Italy, approximately 1,674,000 families are in absolute poverty—a total of almost 4.6 million poor people (7.7% of the entire population; ISTAT, 2020). Specifically, the incidence of absolute poverty is high among young people (11.4% or 1,137,000) and has the lowest rate among people more than 60 years old (5.1%) (ISTAT, 2020). Concerning safety in working conditions, a report drawn up by Italy's National Institute for Insurance against Accidents at Work highlighted 4470 accidents with fatal outcomes focusing on the years between 2015 and 2019 (Istituto Nazionale Assicurazione Infortuni sul Lavoro; INAIL [Italy’s National Institute for Insurance against Accidents at Work]) (2019). Around 60% of workers have been exposed to risk factors that could affect their health, in line with the European average (EUROSTAT, 2021b). Those most involved in psychical accidents are foreign workers employed in low-wage and manual work (Errico et al., 2022). Regarding access to healthcare, Italian law safeguards public medical assistance for all individuals, without discrimination based on income, gender or age (Ministry of Italian Health, 2021). With respect to decent working hours in Italy, a growing number of occupations, particularly those in the service sectors, require even more work to be performed at non-traditional times (Ponzellini, 2020). The balance between organizational and family/social values remains an open issue in Italy: a work–family balance is hard to achieve for more than one-third of the employed (Landolfi et al., 2022). This trend seems to be due to a lack of family-friendly welfare policies (Marra, 2020; Poggesi et al., 2017). Against this background, the need to access decent work in Italy has become an urgent claim (Di Fabio & Kenny, 2019a).

The International Labor Organization (ILO) defines the concept of decent work as the minimum acceptable standard of working conditions with respect for human dignity and equality (ILO, 1999, 2008). Decent work encompasses full job opportunities, benefits and rights, social dialogue and social protection for all workers (ILO, 2014, 2021).

In vocational psychology, the psychology of working theory (PWT; Blustein, 2001, 2006; Duffy et al., 2016) has recently expanded the research on decent work. The PWT is a social justice-oriented framework that changes the focus in career processes from groups with power and privileges to those who face challenges in achieving rights, survival and psychosocial well-being (Blustein, 2001, 2006; Duffy et al., 2016). From this perspective, PWT researchers have outlined the necessity of acquiring workers’ direct psychological experience about the extent to which their working lives meet the ILO criteria for decent work (Blustein et al., 2016; Di Fabio & Blustein, 2016; Di Fabio & Maree, 2016; Guichard, 2013). To this end, Duffy et al. (2016) advanced an empirically testable PWT model of decent work. The centerpiece of the model is decent work itself (Duffy et al., 2016). In this view, decent work captures the ILO's guidelines, focusing on workers’ perceptions of their current job (Duffy et al., 2016). Thus, decent work captures five domains: (1) physical and interpersonally safe working conditions (e.g. an absence of physical and emotional abuse); (2) hours of free time and adequate rest; (3) organizational values that complement family and social values; (4) adequate compensation; and (5) access to adequate healthcare (Duffy et al., 2016, p. 130).

According to the PWT, a decent work model is split into two halves (Duffy et al., 2016). The first half includes predictors of decent work, that is, economic constraints, marginalization, work volition and career adaptability (Duffy et al., 2016), and four moderators, that is, proactive personality, critical consciousness, social support and economic conditions, that may strengthen or weaken work volition and career adaptability (Duffy et al., 2016). The second half of the model consists of outcomes of decent work, namely, self-determination needs, which then predict work fulfillment and well-being (Duffy et al., 2016).

Subsequently, Duffy et al. (2017) developed the Decent Work Scale (DWS), which is a 15-item self-report scale that assesses the above-mentioned five domains of decent work.

Initial findings on the psychometric properties of the DWS have demonstrated a best fit for a bifactor model that allows an overall score and five scores, one for each domain (Duffy et al., 2017). Nine follow-up studies adapting the DWS and confirming the bifactor structure of the scale with populations outside the United States have been conducted (for an overview, see Blustein & Duffy, 2020). The bifactor structure of the DWS and its stability was strongly confirmed across different countries, namely Brazil (Ribeiro et al., 2019), France (Vignoli et al., 2020), Portugal (Ferreira et al., 2019), South Korea (Nam & Kim, 2019), Switzerland (Masdonati et al., 2019), Turkey (Buyukgoze-Kavas & Autin, 2019) and the UK (Dodd et al., 2019).

Findings from the Italian population have also confirmed these results (Di Fabio & Kenny, 2019a). From this perspective, most previous research on the DWS has applied latent factor theory, such as confirmatory factor analysis, which defines psychological constructs as unobservable factors composed of single measurable items (Schmittmann et al., 2013).

However, in recent years, network theory has emerged as a novel approach for examining psychological constructs (Borsboom, 2017; Borsboom & Cramer, 2013). For example, intelligence (Kan et al., 2020), emotions (Lange et al., 2020), personality (Beck & Jackson, 2021; Costantini et al., 2015), mental disorders (Contreras et al., 2019; Fried et al., 2017), well-being (Govorova et al., 2020), self-worth (Briganti et al., 2019), empathy (Briganti et al., 2018), attitudes (Dalege et al., 2017), resilience (Briganti & Linkowski, 2019), attachment (McWilliams & Fried, 2019) and parental burnout (Blanchard et al., 2021) have been studied according to the network theory. In such a theory, a psychological construct arises from the direct causal relations among single measurable items (Borsboom & Cramer, 2013; Fried, 2020). In summary, single items cause each other, and a psychological construct is an emergent property that results from these mutual relations (Fried, 2020). Nevertheless, questionnaires in psychology are often developed to assess unobservable factors (Briganti et al., 2019). Thus, scales may enclose highly similar items measuring the same factor, which has been highlighted as a challenge for network theory (Fonseca-Pedrero et al., 2018). In that occurrence, the interpretation of the links between items changes—a connection between two items simply indicates their shared variance and not an exact reciprocal relation (Fried & Cramer, 2017). This restraint also holds for the DWS, where common causes are possible: items in a given domain may assess the same construct and thus can be also investigated with factor models.

Aims of this study

Thereby, our study applied network models to the construct of decent work (Duffy et al., 2017) while tackling the challenge of items assessing the same domain using both network models (Epskamp et al., 2017; Kan et al., 2020) and the domain network approach (Briganti & Linkowski, 2019; Briganti et al., 2019). The first goal is to determine whether the network approach is a plausible theory to explain the construct of decent work, that is, whether the construct of decent work as a network of items interacting and influencing each other can be theorized. Regarding network theory, the interaction between items can reflect the judgment of workers, which evaluates whether their jobs fulfill decent work characteristics. The second goal is to explore the DWS as a network composed of its sum score domains. In this vein, we want to: (1) calculate the expected influence (EI) of domains (Robinaugh et al., 2016) and determine the most central ones in the network; and (2) measure the domain predictability (Haslbeck & Waldorp, 2018), highlighting the percentage of the shared variance of a domain with other neighboring domains. Such analysis may break a new ground to explore the most central features of decent work in Italy, allowing researchers and practitioners to detect targets for actions (Fried et al., 2018).

Methods

Participants and procedure

The current study was conducted on a dataset consisting of 645 workers from various organizations in Tuscany (females = 65.1%; males = 34.9%; mean age = 43.9 years; SD = 10.9). Participants were workers employed in different public and private organizations recruited voluntarily from their organizations, who granted permission for research in their setting. Data was collected in both paper–pencil and online formats. According to reporting standards for psychological network analyses in cross-sectional data (Burger et al., 2020), we considered only participants without missing data since the psychonetrics (Epskamp, 2021b) package applies full information maximum likelihood estimation, which requires observed data to estimate the network structure. The recruitment of participants was voluntary and before proceeding with the enrollment participants were informed about the general purposes of the study. Informed consent was obtained from all participants according to the requirements of privacy and in compliance with the ethical standards of Italian law.

Measurement

DWS—Italian version

The DWS (Duffy et al., 2017; Italian version, Di Fabio & Kenny, 2019a) is a self-report questionnaire assessing decent work in the following five domains: physically and interpersonally safe working conditions, access to healthcare, adequate compensation, hours that allow for free time and rest and organizational values complement family and social values. The DWS includes 15 items ranked on a Likert scale from 1 (strongly disagree) to 7 (strongly agree). Four reverse-scored items were included (i.e. items 7, 8, 10 and 11). The Italian version adapts the domain of access to healthcare (i.e. items 4–6) considering the Italian healthcare system (Di Fabio & Kenny, 2019a). The English version of the DWS shows excellent reliability (Cronbach's coefficients between 0.79 and 0.97; Duffy et al., 2017). The Italian version of the questionnaire has reported psychometric properties consistent with the original English version (Di Fabio & Kenny, 2019a). In the present study, Cronbach's coefficients ranged from 0.77 to 0.94.

Data analysis

Data were analyzed using R Studio version 1.3.959 for Macintosh. The packages used to perform the analysis are shown in each subsection.

Descriptive and preliminary analysis

The mean, standard deviation, kurtosis and skewness were inspected for all items and the five domains of the DWS. Values of skewness and kurtosis ranging from −1 to 1 were considered acceptable (Epskamp & Fried, 2018). The R packages readr 2.1.1 (Wickham et al., 2021b) and psych 2.1.9 (Revelle, 2021) were used. Furthermore, to check whether the summed total scores of the five DWS domains are sufficient statics, we ran the Mokken scale analysis (Sijtsma & Verweij, 1992). The scalability coefficient (H_ij) (Loevinger, 1948; Mokken, 2011) was calculated for each DWS domain. A domain was considered to constitute a strong scale for 0.5 ≤ H_ij ≤ 1.0, a medium scale for 0.4 ≤ H_ij < 0.5, an acceptable scale for 0.3 ≤ H_ij < 0.4 and an unacceptable scale for H_ij < 0.3 (Sijtsma & Verweij, 1992). The R package mokken 3.0.6 (van der Ark et al., 2021) was used. Thereafter, polychoric correlations were calculated for the DWS items, and Pearson correlations were calculated for the summed total scores of the five DWS domains. The packages qgraph 1.9 (corauto function) (Epskamp, 2021a) and corrplot 0.90 (Wei, 2021) were used.

Network analysis

We performed the network analysis following the reporting standards for psychological network analyses in cross-sectional data (Burger et al., 2020). In line with the aims of our study, we performed the analyses in two steps. In step 1, we confirmatively tested three models of the DWS and exploratively searched for a residual and a latent network structure of the DWS according to the generalized network psychometrics framework (Epskamp et al., 2017). In step 2, we estimated the network of the DWS domains with nodes displaying the sum scores of each domain (Briganti et al., 2019). These approaches were selected because the items enclosed in the DWS domains tend to compute the same construct, and it is an occurrence in which the network of all single items can be problematic (i.e. topological overlap).

Network analysis: generalized network psychometrics models of the DWS

As the first step, we performed a comparison between three factorial models and exploratively searched for a residual and a latent network structure for DWS. A comparison was run to determine which model accounted for the best theoretical explanation of the domains of the DWS. The three factorial models follow the latent factor theory. They were the correlational model (i.e. items loading on its corresponding factor and the five factors were mutually correlated), the higher-order model (i.e. items loading on its respective factor and the five factors regressed onto a higher-order factor) and the bifactor model (i.e. items are simultaneously regressed on its respective five factors and onto a decent work factor) (Di Fabio & Kenny, 2019a). According to this approach, the psychological domains of the DWS stem from latent factors. Conversely, the network psychometric models (Epskamp et al., 2017) adhere to the network theory (e.g., Cramer et al., 2010, 2012; Fried, 2020; Schmittmann et al., 2013). According to this perspective, the psychological domains of the DWS arise from the reciprocal interaction between its constituent elements (i.e. DWS items; Epskamp et al., 2017; Fried, 2020). Network psychometrics models are formal psychometric models that apply the Gaussian graphical model (GGM; Lauritzen, 1996) instead of the structural equation model (SEM; Kaplan, 2009). This framework consists of latent network modeling (LNM), which arranges network models among latent variables, and the residual network modeling (RNM), which arranges a network structure on the residuals of a SEM model. Thus, in line with Epskamp et al. (2017), we started from the correlational model, and we subsequently applied the least absolute shrinkage and selection operator (LASSO) to the RNM model to search for a residual network. We implemented 100 different tuning parameters with the extended Bayesian information criterion. After fitting a RNM model, we removed non-statistically significant edges (p > 0.05) from the latent network in the resulting model, which was labeled as the RNM + LNM model. Comparisons between network and factor models were evaluated via absolute (i.e. comparative fit index, Tucker–Lewis fit index, root mean square error of approximation, RMSEA) and relative (i.e. Bayesian information criterion, BIC, and Akaike information criterion, AIC) fit indexes (Schermelleh-Engel et al., 2003). Comparative fit index and Tucker–Lewis fit index values of > 0.97 indicated a good fit, whereas values between 0.95 and 0.97 indicated an acceptable fit. RMSEA values were evaluated as follows: good, ≤0.05; adequate, between 0.05 and 0.08; mediocre, from 0.08 to 0.10; and unacceptable, > 0.10. The model with the lowest AIC and BIC values was considered to provide the best relative fit (Schermelleh-Engel et al., 2003).

The RNM model run with extended Bayesian information criterion LASSO was first estimated via lvnet 0.3.5 (Epskamp et al., 2019). Thereafter, all of the models were calculated and compared via the R package psychonetrics 0.10 (Epskamp, 2021b), explicitly developed to model and compare latent factor and psychometric network models. The R packages qgraph 1.9 (Epskamp, 2021a), lavaan 0.6-9 (Rosseel et al., 2021), SemPlot 1.1.2 (Epskamp et al., 2019) and dplyr 1.0.7 (Wickham, 2021) were also used.

Network analysis: sum score domains of the DWS

As the second step, we investigated the network structure of the five domains of the DWS, following the procedure of Briganti et al. (2019). This latter procedure considers the sum scores of the domains as variables in the estimation of the network structure (Epskamp & Fried, 2018). Thus, first, the correlation matrix of the sum scores of the five DWS domains was calculated. Second, partial correlation was estimated via the GGM, and the obtained edge weight parameters were regularized by applying the graphical LASSO (Epskamp & Fried, 2018). The network model (out of 100) with the lowest lambda (tuning = 0.001) was selected (Briganti et al., 2019). In this model, five nodes were present, one for each domain of the DWS. Edges display the regularized partial correlation between the two domains controlled by all further domains. Blue and red edges indicate positive and negative relations with corresponding thicknesses displaying their weight, respectively. The Fruchterman–Reingold algorithm (Fruchterman & Reingold, 1991) was applied to locate the nodes in the network. The R packages qgraph 1.9 (Epskamp, 2021a) and glasso 1.11 (Friedman et al., 2019) were used.

The network structure of the five domains of the DWS was evaluated via two measures of different local inference: the EI index (Robinaugh et al., 2016) and node predictability (Haslbeck & Waldorp, 2018). The EI index is the sum of node connections and identifies the most central and influential nodes across a network (Robinaugh et al., 2016). Node predictability represents an absolute measure of the interconnectedness of a particular node in the network in terms of the percentage of shared variance with related nodes (Fried et al., 2018; Haslbeck & Waldorp, 2018). The higher the percentage of node predictability is, the higher the connection is with all of the surrounding nodes. In the graphical representation of the network, node predictability for each node is shown via the pie chart surrounding the node. The R packages mgm 1.2-9 (Haslbeck, 2021) and qgraph (Epskamp, 2021a) were used.

Network stability was calculated via the correlation stability (CS) coefficient, non-parametric bootstrapped difference test for EI, non-parametric bootstrapped difference test for edge weights and bootstrap tests of edge weight accuracy (Epskamp et al., 2018). A CS coefficient of > 0.50 indicates a stable strength (Epskamp et al., 2018). The non-parametric bootstrapped difference test for edge weights (2000 bootstraps on 95% confidence intervals, CIs, of estimated differences) was employed to determine the statistically significant differences among edges. Central boxes display thicknesses of edges, gray boxes indicate no statistically significant differences and black boxes indicate statistically significant differences. The non-parametric bootstrapped difference test for EI (2000 bootstraps on 95% CIs of estimated differences) was employed to determine the statistically significant difference between the EI calculated for each node. Central boxes display the row value of EI, gray boxes indicate no statistically significant differences and black boxes indicate statistically significant differences (Epskamp et al., 2018). Finally, the bootstrap tests of the edge weight accuracy furnish a plotted curve of the 95% CIs, where larger CIs indicate lower precision and narrower CIs indicate higher precision (Epskamp et al., 2018). The R packages igraph (Nepusz, 2021) and bootnet (Epskamp, 2021c) were used.

The following R packages were also used: lavaan 0.6-9 (Rosseel et al., 2021), dplyr 1.0.7 (Wickham, 2021), bnlearn 4.7 (Scutari et al., 2021), reshape2 1.4.4 (Wickham, 2020), ggplot2 3.3.5 (Wickham et al., 2021a) and Hmisc 3.3.5 (Harrell, 2021).

Results

Table 1 shows the descriptive statistics of the DWS. All of the items and the five domains showed acceptable skewness and kurtosis values (Table 1). Thus, they were retained for the network analyses. Table 1 also shows the Mokken homogeneity coefficients (H_ij) for the items and domains of DWS. All of the domains showed adequate homogeneity ranging from strong (access to healthcare, H_ij = 0.84) to acceptable (free time and rest, H_ij = 0.33) coefficients. Thus, the summed total scores of the DWS domains were considered sufficient statistics.

Table 1.

The Decent Work Scale: means, standard deviations, skewness, and kurtosis (n = 645).

Italian version of the Decent Work Scale domain names and items	Mean	SD	Skewness	Kurtosis	H_ij
Physically and interpersonally safe working conditions (SC)
1. I feel emotionally safe interacting with people at work	5.05	1.54	−0.61	−0.44	0.62
2. At work, I feel safe from emotional or verbal abuse of any kind	4.65	1.77	−0.42	−0.88	0.64
3. I feel physically safe interacting with people at work	5.43	1.50	−0.83	0.01	0.62
SC total score	15.13	4.12	−0.57	−0.33	0.63
Access to healthcare (AH)
4. I get good healthcare benefits from my country/government	4.48	1.64	−0.38	−0.74	0.84
5. I have a good healthcare plan provided by my country/government	4.47	1.62	−0.34	−0.69	0.86
6. My country/government provides acceptable options for healthcare	4.54	1.60	−0.45	−0.58	0.81
AH total score	13.50	4.58	−0.42	−0.49	0.84
Adequate compensation (AC)
7. I am [NOT] properly paid for my work	3.24	1.89	0.51	−0.87	0.55
8. I do [NOT] feel I am paid enough based on my qualifications and experience	3.26	1.87	0.52	−0.82	0.57
9. I am rewarded adequately for my work	3.91	1.77	0.00	−0.99	0.34
AC total score	10.40	4.46	0.38	−0.26	0.49
Hours that allow for free time and rest (FT)
10. I do [NOT] have enough time for non-work activities	3.86	1.70	0.09	−0.83	0.36
11. I have [NO] time to rest during the work week.	4.02	1.73	0.00	−0.89	0.39
12. I have free time during the work week	4.14	1.73	−0.12	−0.89	0.23
FT total score	12.02	3.81	−0.01	−0.04	0.33
Organizational values complement family and social values (CV)
13. The values of my organization match my family values	4.73	1.57	−0.34	−0.60	0.78
14. My organization's values align with my family values	4.76	1.52	−0.34	−0.59	0.81
15. The values of my organization match the values within my community	4.56	1.44	−0.24	−0.49	0.66
CV total score	14.05	4.10	−0.32	−0.40	0.75

Note: [NOT] and [NO] display previously reversed items. Higher mean scores indicate higher decent work among all the items and domains. H_ij = Mokken Scalability coefficient.

Figure 1 shows zero-order polychoric correlations of the DWS items. As expected, the items comprised in each domain correlate highly with each other (Figure 1). Figure 2 shows zero-order Pearson correlations of the five domains of the DWS. The correlation between safe working conditions and access to healthcare was the strongest. In contrast, correlations between free time and rest and both dimensions of safe working conditions and adequate compensation are the weakest.

Figure 1.

Zero-order polychoric correlations among the Decent Work Scale items (n = 645). Note: Negatively worded items (i.e. 7, 8, 10, 11) were reversed before the calculation of the polychoric correlations. Among all items, higher scores indicate higher decent work.

Figure 2.

Zero-order Pearson correlations among the Decent Work Scale domains (n = 645).

Table 2 reports the results of the relative and absolute fit measures of the five tested models. The RNM and RNM + RLM models showed a good absolute fit, the bifactor model showed an adequate fit and the higher-order and correlational models reported an unacceptable fit (Table 2). Similar results were observed in the RMSEA values: the fit of the RNM and RNM + RLM models was good, whereas the fits of all the factorial models were only adequate. Finally, the RNM + RLM model showed better AIC and BIC values than the other models. Thus, the results were favorable for the network approach over the factorial approach, indicating that the network realm is the best framework for explaining the psychological self-perception of decent work. It has, however, to be noted that across the factorial models, the bifactor model showed the best fit.

Table 2.

Fit measures for the four models estimated for the Decent Work Scale (n = 645).

Model	(d.f.)	p-Value	TLI	CFI	RMSEA [CI₉₀]	BIC	AIC
RNM + LNM	58 (98.14)	<0.001	0.99	0.99	0.033 [0.021−0.044]	32,506.82	32,162.69
RNM	56 (95.55)	<0.001	0.99	0.99	0.033 [0.021−0.044]	32,517.16	32,164.09
Bifactor model	75 (214.40)	<0.001	0.96	0.97	0.054 [0.045−0.062]	32,513.09	32,244.94
Correlational model	80 (314.10)	<0.001	0.93	0.95	0.067 [0.060−0.075]	32,580.45	32,334.64
Higher order model	85 (342.77)	<0.001	0.93	0.95	0.069 [0.061−0.076]	32,576.77	32,353.31

Note: CFI, Comparative fit index; TLI, Tucker–Lewis fit index; RMSEA, Root mean square error of approximation; AIC, Akaike information criterion; BIC, Bayesian Information Criterion. RNM is the unvarying model as the correlational model with a residual network. RNM + LNM designates the unvarying model as the RNM model in which two edges of the latent network have been deleted.

Figure 3 displays the factor structure, the latent network and the residual network of the DWS inherent in the RNM + RLM model. The latent network of DWS showed that the most connected domains were complementary values, safe working conditions and access to healthcare (Figure 3). Notably, even though statistically significant, nodes between adequate compensation and access to healthcare, as well as between complementary values and free time and rest, are very thick, showing very low values at 0.14 and 0.08, respectively (Figure 3). Furthermore, the residual network of the DWS showed that the strongest edges were between items of the same domains, and only item 9 was not included in the network of residuals (Figure 3).

Figure 3.

Graphical representation of the factor structure and latent network (left) and the residual network (right) of the Decent Work Scale (DWS; n = 645).

Figure 4 illustrates the network for the sum scores of the five domains of the DWS. Safe working conditions and access to healthcare share the strongest connection in the network. Moreover, safe working conditions also shares a strong connection with organizational values. Lastly, access to healthcare is also connected with safe working conditions. In contrast, adequate compensation is connected only with organizational values and, similarly, free time and rest is only connected with access to healthcare.

Figure 4.

Graphical representation of the network model for five domains of the decent work scale (n = 645).

Figure 5 illustrates the EI as Z-scores for the DWS domain network. Safe working conditions, access to healthcare and organizational values have the highest EI values, showing that these domains are the most connected in the network and have a higher influence on the rest of the domains in the network. In contrast, free time and rest and adequate compensation have the lowest EI values, showing that these domains are the least connected in the network and have less influence on the rest of the domains in the network.

Figure 5.

Expected influence centrality estimates for the five-domain network of DWS (n = 645).

Mean node predictability ranged from 0.02 to 0.20, with a mean value of 0.11. This value implies that, on average, 11% of the variance of the network's node can be accounted for by its neighbors. Free time and rest showed the lowest node predictability, sharing only 1% of the variance with its neighboring nodes. Conversely, safe working conditions shows the highest node predictability, sharing 20% of its variance with its neighboring nodes. Access to healthcare has the second largest node predictability (0.17). The correlation between the EI and predictability was 0.98, and the CS coefficient was 0.67, indicating a stable strength of the network structure.

Figure 6 shows the non-parametric bootstrapped difference test for EI. Nodes with higher EI (access to healthcare, EI = 0.57; organizational values, EI = 0.56; safe working conditions, EI = 0.53) are statistically different from nodes with lower EI (free time and rest, EI = 0.15 and adequate compensation, EI = 0.11) (Figure 6).

Figure 6.

Non-parametric bootstrapped difference test for expected influence for the five-domain network of DWS (n = 645).

Figure 7 reports the non-parametric bootstrapped difference test for edge weight. The results showed that stronger edges in the network are significantly stronger than the other edges; thus, the edges displayed in Figure 4 can be interpreted as stronger than the weaker edges.

Figure 7.

Non-parametric bootstrapped difference test for the edge weight five-domain network of DWS (n = 645).

Figure 8 reports the bootstrap tests of edge weight accuracy. As presumed in a network with 645 participants and only five nodes, the CIs are relatively small, showing that the edge weight estimation was reasonably precise.

Figure 8.

Bootstrap tests of the edge weight accuracy (95% confidence intervals) for the DWS domains network (n = 645).

Discussion

To our knowledge, the current work is the first to study a network analysis on the recently developed DWS (Di Fabio & Kenny, 2019a; Duffy et al., 2017), as conceived in the PWT perspective (Blustein, 2001, 2006; Duffy et al., 2016). The analyses showed that psychometrics network models fit better than factor models, with the RNM + RLM model demonstrating the best fit. However, the bifactor model showed the best fit across the factorial models in line with Duffy et al. (2017). This indicates that the measurement model of the construct allows both total and domains scores. Furthermore, the excellent fit observed in the network models provides additional information on the construct. According to network interpretation, the workers’ psychological perception of decent work (Duffy et al., 2017) arises from the interaction of single constitutive elements (i.e. items) of the construct. This finding confirms our hypothesis on the construct of decent work, and it could reflect the workers’ judgment of whether their job fulfills decent work characteristics.

Regarding this perspective, both LNM + RLM and sum score domain approaches provided similar findings, identifying the most connected domains of the DWS. The results showed that the most connected nodes are those with higher EI, that is, safe working conditions, access to healthcare and organizational values. In contrast, adequate compensation and free time and rest are relatively disconnected from the network and display a low EI. Similarly, safe working conditions, access to healthcare and organizational values showed a moderate variance in node predictability (i.e. nodes are moderately explained by other surrounding nodes), whereas adequate compensation and free time and rest displayed a minimal variance in node predictability (i.e. nodes are not explained by other surrounding nodes).

These findings could be explained regarding the Italian context showing Tuscany as a region with an advanced healthcare system and welfare service safeguarding workers (e.g., Cervia, 2012; Di Fabio et al., 2021; ISTAT, 2020). Thus, hypothetically, the most central nodes reflect the workers’ self-perception of the welfare service and high standards of policy that warrant security in the workplace (both physically and ethically; ISTAT, 2020).

Most interestingly, organizational values was also found to be among the most central domains of decent work. This finding could suggest that even though addressing most central nodes may not change the network structure (Fried et al., 2018), organizational values could be an effective target for action aimed at fostering decent work in Italy. To this end, positive strength-based primary preventive actions (Di Fabio et al., 2018; Di Fabio & Kenny, 2019b; Di Fabio & Peiró, 2018; Di Fabio & Saklofske, 2021) could be used.

In contrast, network analysis showed that adequate compensation and free time and rest are relatively disconnected nodes. A plausible interpretation of this finding is statistical, that is, a floor effect that determines the few connections with other domains (Briganti et al., 2019). Moreover, this finding could also be explained by considering the decent–indecent work continuum. In this case, they are disconnected because they are judged as indecent by workers—by being excluded from decent work characteristics. This finding is consistent with statistics that illustrate the ongoing critical concern among workers in Italy regarding the poor quality of remuneration and jobs, as well as the growing level of poverty (Camussi et al., 2021; EUROSTAT, 2021a; ISTAT, 2020). Furthermore, this finding highlights the compelling necessity to advocate for more decent remunerations and working hours in the Italian context. The PWT-informed decent work agenda (Blustein et al., 2019a) can offer a background to stimulate new initiatives and solutions for multidisciplinary and institutional actions.

Our findings should be considered in view of the limitations and strengths. First, our sample consisted of workers from Tuscany, which may potentially limit the generalization of our results to the Italian context. Second, the current study is cross-sectional, and we cannot infer whether a given node (or domain) causes or is caused by another node to which it is linked. Thus, future studies must analyze the temporal–causal relationship of the DWS domain by implementing longitudinal methods. Third, future research must also inspect the different decent work networks associated with different cross-cultural contexts (Di Fabio & Maree, 2016; Duffy et al., 2020), different decent work profiles (Kim et al., 2021) and different groups of vulnerable workers and challenges (Di Fabio et al., 2021; Di Fabio & Svicher, 2021; Svicher & Di Fabio, 2021).

The main strengths are that the network analysis was carried out for the first time to study the construct of decent work (Duffy et al., 2017), following state-of-the-art guidelines (Epskamp et al., 2017, 2018; Epskamp & Fried, 2018), and that our results are stable and trustworthy.

In brief, our results revealed that a network model could provide a novel perspective for understanding the construct of decent work (Duffy et al., 2017). In this line, the workers’ psychological perception of decent work arises from the interaction of single constitutive elements of the construct. The most central aspects of decent work in Italy include safe working conditions, access to healthcare and organizational values. The results also indicate that organizational values could represent a potential target for positive strength-based primary preventive actions. In contrast, adequate compensation and free time and rest are not connected to the main pathway of the network, highlighting that these job characteristics are not considered decent by Italian workers. This latter finding further underlines the urgent need for institutions, policy makers, trade unions, and practitioners to advocate for more decent remunerations and working hours in Italy. That said, the robustness, trustworthiness and stability of the network are highly promising for the replicability of the results.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Annamaria Di Fabio

Andrea Svicher

Alessio Gori

References

Allan

B. A.

Autin

K. L.

Wilkins-Yel

K. G

. (2021). Precarious work in the 21st century: A psychological perspective. Journal of Vocational Behavior, 126, Article 103491. https://doi.org/10.1016/j.jvb.2020.103491

Autin

K. L.

Blustein

D. L.

Ali

S. R.

Garriott

P. O

. (2020). Career development impacts of COVID-19: Practice and policy recommendations. Journal of Career Development, 47(5), 487–494. https://doi.org/10.1177/0894845320944486

Bauman

(2013). Liquid modernity. Polity Press.

Beck

E. D.

Jackson

J. J.

(2021). Network approaches to representing and understanding personality dynamics. In Wood

Read

S. J.

Harms

P. D.

Slaughter

(Eds.), Measuring and modeling persons and situations (pp. 465–497). Academic Press. https://doi.org/10.1016/B978-0-12-819200-9.00003-X

Blanchard

M. A.

Roskam

Mikolajczak

Heeren

. (2021). A network approach to parental burnout. Child Abuse & Neglect, 111, 104826. https://doi.org/10.1016/j.chiabu.2020.104826

Blustein

D. L

. (2001). Extending the reach of vocational psychology: Toward an inclusive and integrated psychology of working. Journal of Vocational Behavior, 59(2), 171–182. https://doi.org/10.1006/jvbe.2001.1823

Blustein

D. L.

(2006). The psychology of working: A new perspective for counseling, career development, and public policy. Lawrence Erlbaum Associates.

Blustein

D. L.

Duffy

(2020). Psychology of working theory. In Brown

S. D.

Lent

R. W.

(Eds.), Career development and counseling: putting theory and research to work. 3rd Edition (pp. 577–671). Wiley.

Blustein

D. L.

Guarino

P. A

. (2020). Work and unemployment in the time of COVID-19: The existential experience of loss and fear. Journal of Humanistic Psychology, 60(5), 702–709. https://doi.org/10.1177/0022167820934229

10.

Blustein

D. L.

Olle

Connors-Kellgren

Diamonti

A. J

. (2016). Decent work: A psychological perspective. Frontiers in Psychology, 7(407). https://doi.org/10.3389/fpsyg.2016.00407

11.

Blustein

D. L.

Kenny

M. E.

Di Fabio

Guichard

. (2019a). Expanding the impact of the psychology of working: Engaging psychology in the struggle for decent work and human rights. Journal of Career Assessment, 27(1), 3–28. https://doi.org/10.1177/1069072718774002

12.

Blustein

D. L.

Kenny

M. E.

Autin

Duffy

. (2019b). The psychology of working in practice: A theory of change for a new era. The Career Development Quarterly, 67(3), 236–254. https://doi.org/10.1002/cdq.12193

13.

Blustein

D. L.

Ali

S. R.

Flores

L. Y

. (2019c). Vocational psychology: Expanding the vision and enhancing the impact. The Counseling Psychologist, 47(2), 166–221. https://doi.org/10.1177/0011000019861213

14.

Blustein

D. L.

Duffy

Ferreira

J. A.

Cohen-Scali

Cinamon

R. G.

Allan

B. A

. (2020). Unemployment in the time of COVID-19: A research agenda. Journal of Vocational Behavior, 119, 103436. https://doi.org/10.1016/j.jvb.2020.103436

15.

Borsboom

. (2017). A network theory of mental disorders. World Psychiatry, 16(1), 5–13. https://doi.org/10.1002/wps.20375

16.

Borsboom

Cramer

A. O

. (2013). Network analysis: An integrative approach to the structure of psychopathology. Annual Review Of Clinical Psychology, 9, 91–121. https://doi.org/10.1146/annurev-clinpsy-050212-185608

17.

Briganti

Linkowski

. (2020). Item and domain network structures of the resilience scale for adults in 675 university students. Epidemiology and Psychiatric Sciences, 29, e33. https://doi.org/10.1017/S2045796019000222

18.

Briganti

Kempenaers

Braun

Fried

E. I.

Linkowski

. (2018). Network analysis of empathy items from the interpersonal reactivity index in 1973 young adults. Psychiatry Research, 265, 87–92. https://doi.org/10.1016/j.psychres.2018.03.082

19.

Briganti

Fried

E. I.

Linkowski

. (2019). Network analysis of contingencies of self-worth scale in 680 university students. Psychiatry Research, 272, 252–257. https://doi.org/10.1016/j.psychres.2018.12.080

20.

Burger

., Isvoranu

., Lunansky

., Haslbeck

J. M. B

., Epskamp

., Hoekstra

R. H. A

., Eiko

., Fried

., Borsboom

Blanken

. (2020). Reporting standards for psychological network analyses in cross-sectional data. Unpublished article. https://doi.org/10.31234/osf.io/4y9nz

21.

Buyukgoze-Kavas

Autin

K. L

. (2019). Decent work in Turkey: Context, conceptualization, and assessment. Journal of Vocational Behavior, 112, 64–76. https://doi.org/10.1016/j.jvb.2019.01.006

22.

Camussi

S. A.

Maccarrone

Aimone Gigio

. (2021). Changes in the employment structure and in job quality in Italy: A national and regional analysis. Bank of Italy Occasional Paper no. 603. https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.3828111

23.

Cartwright

Cooper

C. L

. (2014). Towards organizational health: Stress, positive organizational behavior, and employee well-being. In Bauer

G. F.

Hämmig

(Eds.), Bridging occupational, organizational and public health: A transdisciplinary approach (pp. 29–42). Springer Science + Business Media. https://doi.org/10.1007/978-94-007-5640-3_3

24.

Cervia

. (2012). Paradigma democratico e partecipazione dei cittadini in sanità: Il caso della toscana [Democratic paradigm and citizen participation in healthcare: The case of Tuscany]. Salute e Società, 9, 151–164. https://doi.org/10.3280/SES2012-001010

25.

Contreras

Nieto

Valiente

Espinosa

Vazquez

. (2019). The study of psychopathology from the network analysis perspective: A systematic review. Psychotherapy and Psychosomatics, 88(2), 71–83. https://doi.org/10.1159/000497425

26.

Costantini

Richetin

Borsboom

Fried

E. I.

Rhemtulla

Perugini

. (2015). Development of indirect measures of conscientiousness: combining a facets approach and network analysis. European Journal of Personality, 29(5), 548–567. https://doi.org/10.1002/per.2014

27.

Cramer

A. O. J.

Waldorp

L. J.

van der Maas

H. L. J.

Borsboom

. (2010). Comorbidity: A network perspective. Behavioral and Brain Sciences, 33(2-3), 137–150. https://doi.org/10.1017/S0140525X09991567

28.

Cramer

A. O. J.

van der Sluis

Noordhof

Wichers

Geschwind

Aggen

S. H.

Kendler

K. S.

Borsboom

. (2012). Dimensions of normal personality as networks in search of equilibrium: You can’t like parties if you don’t like people. European Journal of Personality, 26(4), 414–431. https://doi.org/10.1002/per.1866

29.

Dalege

Borsboom

van Harreveld

van der Maas

H. L. J

. (2017). Network analysis on attitudes: A brief tutorial. Social Psychological and Personality Science, 8(5), 528–537. https://doi.org/10.1177/1948550617709827

30.

Deranty

J.-P.

MacMillan

. (2012). The ILO’s decent work initiative: Suggestions for an extension of the notion of “decent work”. Journal of Social Philosophy, 43(4), 386–405. https://doi.org/10.1111/josp.12003

31.

Di Fabio

Blustein

D. L.

(2016). Editorial: From meaning of working to meaningful lives: The challenges of expanding decent work. Frontiers in Psychology, 7(1119). https://doi.org/10.3389/fpsyg.2016.01119

32.

Di Fabio

Kenny

M. E

. (2019a). Decent work in Italy: Context, conceptualization, and assessment. Journal of Vocational Behavior, 110(Part A), 131–143. https://doi.org/10.1016/j.jvb.2018.10.014

33.

Di Fabio

Kenny

M. E.

(2019). Resources for enhancing employee and organizational well-being beyond personality traits: The promise of emotional intelligence and Positive Relational Management. Personality and Individual Differences, 151, Article 109278. https://doi.org/10.1016/j.paid.2019.02.022

34.

Di Fabio

Maree

J. G

. (2016). Using a transdisciplinary interpretive lens to broaden reflections on alleviating poverty and promoting decent work. Frontiers in Psychology, 7(503). https://doi.org/10.3389/fpsyg.2016.00503

35.

Di Fabio

Peiró

J. M.

(2018). Human capital sustainability leadership to promote sustainable development and healthy organizations: A new scale. Psychology of Sustainability and Sustainable Development. Special Issue in Sustainability MDPI, 10(7), 2413. https://doi.org/10.3390/su10072413

36.

Di Fabio

Saklofske

D. H

. (2021). The relationship of compassion and self-compassion with personality and emotional intelligence. Personality And Individual Differences, 169, 110109. https://doi.org/10.1016/j.paid.2020.110109

37.

Di Fabio

Svicher

. (2021). The psychology of sustainability and sustainable development: Advancing decent work, inclusivity, and positive strengths-based primary preventive interventions for vulnerable workers. Frontiers in Psychology, 12, 718354. https://doi.org/10.3389/fpsyg.2021.718354

38.

Di Fabio

Maree

J. G.

Kenny

M. E

. (2019). Development of the life project reflexivity scale: A new career intervention inventory. Journal of Career Assessment, 27(2), 358–370. https://doi.org/10.1177/1069072718758065

39.

Di Fabio

Svicher

Gori

. (2021). Occupational fatigue: Relationship with personality traits and decent work. Frontiers in Psychology, 12(3782). https://doi.org/10.3389/fpsyg.2016.00503

40.

Dodd

Hooley

Burke

. (2019). Decent work in the UK: Context, conceptualization, and assessment. Journal of Vocational Behavior, 112, 270–281. https://doi.org/10.1016/j.jvb.2019.04.002

41.

Duffy

R. D.

Blustein

D. L.

Diemer

M. A.

Autin

K. L

. (2016). The psychology of working theory. Journal of Counseling Psychology, 63(2), 127–148. https://doi.org/10.1037/cou0000140

42.

Duffy

R. D.

Allan

B. A.

England

J. W.

Blustein

D. L.

Autin

K. L.

Douglass

R. P.

Ferreira

Santos

E. J. R

. (2017). The development and initial validation of the Decent Work Scale. Journal of Counseling Psychology, 64(2), 206–221. https://doi.org/10.1037/cou0000191

43.

Duffy

R. D.

Gensmer

Allan

B. A.

Kim

H. J.

Douglass

R. P.

England

J. W.

Autin

K. L.

Blustein

D. L

. (2019). Developing, validating, and testing improved measures within the psychology of working theory. Journal of Vocational Behavior, 112, 199–215. https://doi.org/10.1016/j.jvb.2019.02.012

44.

Duffy

R. D.

Blustein

D. L.

Allan

B. A.

Diemer

M. A.

Cinamon

R. G

. (2020). Introduction to the special issue: A cross-cultural exploration of decent work. Journal of Vocational Behavior, 116, 103351. https://doi.org/10.1016/j.jvb.2019.103351

45.

Epskamp

(2019). “lvnet”. Latent Variable Network Modeling (R package version 0.3.5). [Computer software]. Retrieved from https://cran.r-project.org/web/packages/lvnet/lvnet.pdf

46.

Epskamp

. (2021a). Qgraph. Graph plotting methods, psychometric data visualization and graphical model estimation. R package version 1.9 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/qgraph/qgraph.pdf

47.

Epskamp

. (2021b). Psychonetrics. R package version 0.10 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/psychonetrics/psychonetrics.pdf

48.

Epskamp

. (2021c). Bootnet. Bootstrap methods for Various network estimation routines. R package version 1.5 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/bootnet/bootnet.pdf

49.

Epskamp

Fried

E. I

. (2018). A tutorial on regularized partial correlation networks. Psychological Methods, 23(4), 617–634. https://doi.org/10.1037/met0000167

50.

Epskamp

Rhemtulla

Borsboom

. (2017). Generalized network psychometrics: Combining network and latent Variable models. Psychometrika, 82(4), 904–927. https://doi.org/10.1007/s11336-017-9557-x

51.

Epskamp

Borsboom

Fried

E. I

. (2018). Estimating psychological networks and their accuracy: A tutorial paper. Behavior Research Methods, 50(1), 195–212. https://doi.org/10.3758/s13428-017-0862-1

52.

Epskamp

Stuber

Nak

Veenman

Jorgensen

, D. (2019). Semplot. Path Diagrams and Visual Analysis of Various SEM Packages’ Output. R package version 1.1.2 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/semPlot/semPlot.pdf

53.

Errico

Drommi

Calamano

Barranco

Molinari

Ventura

. (2022). Fatal work-related injuries in the Genoa district (north-western Italy): Forensic analysis of the 10-year period between 2011 and 2020. Journal of Forensic and Legal Medicine, 85, 102294. https://doi.org/10.1016/j.jflm.2021.102294

54.

Eurostat. (2021a, April). Employment—annual statistics. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Employment_-_annual_statistics

55.

Eurostat. (2021b, September). Translate self-reported work-related health problems and risk factors—key statistics. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Self-reported_work-related_health_problems_and_risk_factors_-_key_statistics#Exposure_to_physical_risk_factors_at_work

56.

Ferreira

J. A.

Haase

R. F.

Santos

E. R.

Rabaça

J. A.

Figueiredo

Hemami

H. G.

Almeida

L. M

. (2019). Decent work in Portugal: Context, conceptualization, and assessment. Journal of Vocational Behavior, 112, 77–91. https://doi.org/10.1016/j.jvb.2019.01.009

57.

Fonseca-Pedrero

Ortuño

Debbané

Chan

R. C. K.

Cicero

Zhang

L. C.

Brenner

Barkus

Linscott

R. J.

Kwapil

Barrantes-Vidal

Cohen

Raine

Compton

M. T.

Tone

E. B.

Suhr

Inchausti

Bobes

Fumero

, …, Fried

E. I

. (2018). The network structure of schizotypal personality traits. Schizophrenia Bulletin, 44(suppl_2), S468−SS479. https://doi.org/10.1093/schbul/sby044

58.

Fried

E. I

. (2020). Lack of theory building and testing impedes progress in the factor and network literature. Psychological Inquiry, 31(4), 271–288. https://doi.org/10.1080/1047840X.2020.1853461

59.

Fried

E. I.

Cramer

A. O. J

. (2017). Moving forward: Challenges and directions for psychopathological network theory and methodology. Perspectives on Psychological Science, 12(6), 999–1020. https://doi.org/10.1177/1745691617705892

60.

Fried

E. I.

van Borkulo

C. D.

Cramer

A. O. J.

Boschloo

Schoevers

R. A.

Borsboom

. (2017). Mental disorders as networks of problems: A review of recent insights. Social Psychiatry and Psychiatric Epidemiology, 52(1), 1–10. https://doi.org/10.1007/s00127-016-1319-z

61.

Fried

E. I.

Eidhof

M. B.

Palic

Costantini

Huisman-van Dijk

H. M.

Bockting

C. L. H.

Engelhard

Armour

Nielsen

A. B. S.

Karstoft

K.-I

. (2018). Replicability and generalizability of posttraumatic stress disorder (PTSD) networks: A cross-cultural multisite study of PTSD symptoms in four trauma patient samples. Clinical Psychological Science, 6(3), 335–351. https://doi.org/10.1177/2167702617745092

62.

Friedman

Hastie

Tibshirani

(2019). Glasso. Graphical Lasso: Estimation of Gaussian Graphical Models. R package version 1.11 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/glasso/glasso.pdf

63.

Fruchterman

T. M. J.

Reingold

E. M.

(1991). Graph drawing by force-directed placement. Software: Practice and Experience, 21(11), 1129−1164. https://doi.org/10.1002/spe.4380211102

64.

Govorova

Benítez

Muñiz

. (2020). Predicting student well-being: Network analysis based on PISA 2018. International Journal of Environmental Research and Public Health, 17(11), 4014. https://doi.org/10.3390/ijerph17114014

65.

Guichard

. (2009). Self-constructing. Journal of Vocational Behavior, 75(3), 251–258. https://doi.org/10.1016/j.jvb.2009.03.004

66.

Guichard

. (2013, November 26). Career guidance, education, and dialogues for a fair and sustainable human development. Inaugural conference of the UNESCO chair of Lifelong guidance and counselling. University of Wroclaw, Wroclaw, Poland. https://hal-cnam.archives-ouvertes.fr/hal-03240556/document

67.

Harrell

F. E.

Jr (2021). Hmisc. Harrell Miscellaneous. R package version 4.6-0 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/ggplot2/ggplot2.pdf

68.

Haslbeck

. (2021). Mgm. Estimating time-varying k-order mixed graphical models. R package version 1.2-12 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/mgm/mgm.pdf

69.

Haslbeck

J. M. B.

Waldorp

L. J

. (2018). How well do network models predict observations? On the importance of predictability in network models. Behavior Research Methods, 50(2), 853–861. https://doi.org/10.3758/s13428-017-0910-x

70.

International Labor Organization. (1999). Report of the Director-General: Decent work. https://www.ilo.org/public/english/standards/relm/ilc/ilc87/rep-i.htm

71.

International Labor Organization. (2008). World of work report 2008: Income inequalities in the age of financial globalization. https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/@publ/documents/publication/wcms_100354.pdf

72.

International Labor Organization. (2014). Stock-taking of the ILO’s programme on measuring decent work (2014). Decent Work Indicators and Decent Work Country Profiles. https://www.ilo.org/wcmsp5/groups/public/---dgreports/---integration/documents/publication/wcms_312643.pdf

73.

International Labor Organization. (2021). Decent work. https://www.ilo.org/global/topics/decent-work/lang--en/index.htm

74.

Istituto Nazionale Assicurazione Infortuni sul Lavoro. (2019). Andamento degli infortuni sul lavoro e delle malattie professionali [Italy’s National Institute for Insurance against Accidents at Work - Trend of accidents at work and occupational disease]. https://www.inail.it/cs/internet/docs/alg-dati-inail-2019-giugno-pdf.pdf

75.

Istituto Nazionale di Statistica. ISTAT. (2020). Annuario statistico italiano 2020 [Italian Statistical Yearbook]. https://www.istat.it/storage/ASI/AnnuarioStatistico_2020/Asi_2020.pdf

76.

Johnson

M. T.

Johnson

E. A.

Webber

Nettle

. (2020). Mitigating social and economic sources of trauma: The need for universal basic income during the coronavirus pandemic. Psychological Trauma: Theory, Research, Practice, and Policy, 12(S1), S191–S192. https://doi.org/10.1037/tra0000739 https://doi.org/10.1037/tra0000739

77.

Kan

K.-J.

de Jonge

van der Maas

H. L. J.

Levine

S. Z.

Epskamp

. (2020). How to compare psychometric factor and network models. Journal of Intelligence, 8(4), 35. https://www.mdpi.com/2079-3200/8/4/35 https://doi.org/10.3390/jintelligence8040035

78.

Kaplan

. (Ed.) (2009). Structural equation modeling: Foundations and extensions (2nd edn) (vols 1–10). Sage. https://www.https://doi.org/10.4135/9781452226576

79.

Kim

H. J.

Duffy

R. D.

Allan

B. A

. (2021). Profiles of decent work: General trends and group differences. Journal of Counseling Psychology, 68(1), 54–66. https://doi.org/10.1037/cou0000434

80.

Kniffin

K. M.

Narayanan

Anseel

Antonakis

Ashford

S. P.

Bakker

A. B.

Bamberger

Bapuji

Bhave

D. P.

Choi

V. K.

Creary

S. J.

Demerouti

Flynn

F. J.

Gelfand

M. J.

Greer

L. L.

Johns

Kesebir

Klein

P. G.

Lee

S. Y.

, …, Vugt

M. V

. (2021). COVID-19 and the workplace: Implications, issues, and insights for future research and action. American Psychologist, 76(1), 63–77. https://doi.org/10.1037/amp0000716

81.

Landolfi

Brondino

Molino

Lo Presti

. (2022). Don’t worry, be happy! Positive affect at work, greater balance at home. A daily diary study on work-family balance. European Review of Applied Psychology, 72(1), 100715. https://doi.org/10.1016/j.erap.2021.100715

82.

Lange

Dalege

Borsboom

van Kleef

G. A.

Fischer

A. H

. (2020). Toward an integrative psychometric model of emotions. Perspectives on Psychological Science, 15(2), 444–468. https://doi.org/10.1177/1745691619895057

83.

Lauritzen

S. L.

(1996). Graphical models. Clarendon Press.

84.

Loevinger

. (1948). The technic of homogeneous tests compared with some aspects of “scale analysis” and factor analysis. Psychological Bulletin, 45(6), 507–529. https://doi.org/10.1037/h0055827

85.

Marra

. (2020). Intersectionality and emergence in time use analysis: possible pathways for gender research and work–family reconciliation policy in Italy. Community, Work & Family, 23(1), 80–106. https://doi.org/10.1080/13668803.2018.1462146

86.

Masdonati

Schreiber

Marcionetti

Rossier

. (2019). Decent work in Switzerland: Context, conceptualization, and assessment. Journal of Vocational Behavior, 110, 12–27. https://doi.org/10.1016/j.jvb.2018.11.004

87.

McWilliams

L. A.

Fried

E. I

. (2019). Reconceptualizing adult attachment relationships: A network perspective. Personal Relationships, 26(1), 21–41. https://doi.org/10.1111/pere.12263

88.

Ministry of Italian Health. (2021). Strengths of the Italian national health service. https://www.salute.gov.it/portale/cureUE/dettaglioContenutiCureUE.jsp?lingua=english&id=3879&area=healcareUE&menu=vuoto

89.

Mokken

(2011). A theory and procedure of scale analysis: with applications in political research. De Gruyter Mouton. https://doi.org/10.1515/9783110813203

90.

Nam

J. S.

Kim

S. Y

. (2019). Decent work in South Korea: context, conceptualization, and assessment. Journal of Vocational Behavior, 115, 103309. https://doi.org/10.1016/j.jvb.2019.05.006

91.

Nepusz

. (2021). Igraph. Network Analysis and Visualization. R package version 1.2.7 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/igraph/igraph.pdf

92.

Peiró

J. M.

Tetrick

. (2011). Occupational health psychology. In Martin

P. R.

Cheung

F. M.

Knowles

M. C.

Kyrios

Littlefield

Overmier

J. B.

Prieto

J. M.

(Eds.), IAAP handbook of applied psychology (pp. 292–315). Wiley Blackwell. https://doi.org/10.1002/9781444395150.ch12

93.

Poggesi

Mari

De Vita

(2019). Women entrepreneurs and work-family conflict: an analysis of the antecedents. International Entrepreneurship and Management Journal 15(2), 431–454. https://doi.org/10.1007/s11365-017-0484-1

94.

Ponzellini

A. M.

(2020). Worker participation in negotiating working time in Italy. In Gold Gold

(Ed.), New frontiers of democratic participation at work (pp. 249–272). Routledge. https://doi.org/10.4324/9781003073567-14

95.

Revelle

. (2021). Psych. Procedures for Psychological, Psychometric, and Personality Research. R package version 2.1.9 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/psych/psych.pdf

96.

Ribeiro

M. A.

Teixeira

M. A. P.

Ambiel

R. A. M

. (2019). Decent work in Brazil: Context, conceptualization, and assessment. Journal of Vocational Behavior, 112, 229–240. https://doi.org/10.1016/j.jvb.2019.03.006

97.

Robinaugh

D. J.

Millner

A. J.

McNally

R. J

. (2016). Identifying highly influential nodes in the complicated grief network. Journal of Abnormal Psychology, 125(6), 747–757. https://doi.org/10.1037/abn0000181

98.

Rosa

Dörre

Lessenich

. (2017). Appropriation, activation and acceleration: The escalatory logics of capitalist modernity and the crises of dynamic stabilization. Theory, Culture & Society, 34(1), 53–73. https://doi.org/10.1177/0263276416657600

99.

Rosseel

Jorgensen

T.D.

Rockwood

Oberski

Byrnes

Vanbrabant

Savalei

Merkle

Hallquist

Rhemtulla

Katsikatsou

Barendse

Scharf

. (2021). Lavaan. Latent Variable Analysis. R package version 0.6-9 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/lavaan/lavaan.pdf

100.

Savickas

M. L

. (2021). Journal of vocational behavior’s 50th anniversary: Looking back and going forward. Journal of Vocational Behavior, 125, 103540. https://doi.org/10.1016/j.jvb.2021.103540

101.

Savickas

M. L.

Porfeli

E. J

. (2012). Career adapt-abilities scale: Construction, reliability, and measurement equivalence across 13 countries. Journal of Vocational Behavior, 80, 661–673. https://doi.org/10.1016/j.jvb.2012.01.011

102.

Schermelleh-Engel

Moosbrugger

Müller

. (2003). Evaluating the Fit of Structural Equation Models: Tests of Significance and Descriptive Goodness-of-Fit Measures. Methods of Psychological Research, 8(2), 23–74. https://www.stats.ox.ac.uk/∼snijders/mpr_Schermelleh.pdf

103.

Schmittmann

V. D.

Cramer

A. O. J.

Waldorp

L. J.

Epskamp

Kievit

R. A.

Borsboom

. (2013). Deconstructing the construct: A network perspective on psychological phenomena. New Ideas in Psychology, 31(1), 43–53. https://doi.org/10.1016/j.newideapsych.2011.02.007

104.

Scutari

Silander

Ness

. (2021). Bnlearn. Bayesian Network Structure Learning, Parameter Learning and Inference. R package version 4.7 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/bnlearn/bnlearn.pdf

105.

Sijtsma

Verweij

A. C

. (1992). Mokken scale analysis: Theoretical considerations and an application to transitivity tasks. Applied Measurement in Education, 5(4), 355–373. https://doi.org/10.1207/s15324818ame0504_5

106.

Sparks

Faragher

Cooper

C. L

. (2001). Well-being and occupational health in the 21st century workplace. Journal of Occupational and Organizational Psychology, 74(4), 489–509. https://doi.org/10.1348/096317901167497

107.

Svicher

Di Fabio

. (2021). Job crafting: A challenge to promote decent work for vulnerable workers. Frontiers in Psychology, 12, 681022. https://doi.org/10.3389/fpsyg.2021.681022

108.

van der Ark

Koopman

Straat

J. H.

van den Bergh

. (2021). Mokken. Conducts Mokken Scale Analysis. R package version 3.0.6 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/mokken/mokken.pdf

109.

Vignoli

Prudhomme

Terriot

Cohen-Scali

Arnoux-Nicolas

Bernaud

J.-L.

Lallemand

. (2020). Decent work in France: Context, conceptualization, and assessment. Journal of Vocational Behavior, 116, 103345. https://doi.org/10.1016/j.jvb.2019.103345

110.

Wickham

. (2020). Reshape2. Flexibly Reshape Data: A Reboot of the Reshape Package. R package version 1.4.4 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/reshape2/reshape2.pdf

111.

Wickham

. (2021). Dplyr. A Grammar of Data Manipulation. R package version 1.9 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/dplyr/dplyr.pdf

112.

Wickham

Chang

Henry

Pedersen

T. L.

Takahashi

Wilke

Woo

Yutani

Dunnington

. (2021a). ggplot2. Create elegant data visualisations using the grammar of graphics (R package version 3.3.5 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/ggplot2/ggplot2.pdf

113.

Wickham

Hester

Bryan

. (2021b). Readr. Read Rectangular Text Data. R package version 2.1.1. [Computer software]. Retrieved from https://cran.r-project.org/web/packages/readr/readr.pdf

114.

Wei

. (2021). Corplot. Visualization of a Correlation Matrix. R package version 0.90 [Computer software]. Retrieved from https://cran.r-project.org/web/packages/corrplot/corrplot.pdf