Introduction

Business dynamics currently take place in a global and complex system that is the result of the socioeconomic and political changes of the last decades, characterized by volatility, uncertainty, complexity, and ambiguity (VUCA), which refers to a volatile world characterized by uncertainty and economic, political, and social complexity and the ambiguous nature of discourses, social positions, and labor relations (^{Lawrence, 2013}; ^{Terlato, 2019}). Technology, decentralization, the rise of non-state actors, the recent interdependence of organizations and sectors, and the globalization of interactions have accelerated the rise of VUCA environments in all fields, including business (^{Covarrubias-Moreno, 2020}). This leads to an increase in the presence of psychosocial risk factors (PSRFs) and psychosocial risks and the concern, at all levels, to address them in the light of the significant effects they cause on the healthillness process of the worker.

Work-related diseases and accidents resulting from these global conditions, according to the characteristics of each context, present particular features that determine the study of PSRFs, related legislation to prevent stress, violence, and harassment at work, and contractual insecurity, burnout, familywork conflict, and emotional labor (^{Peiró, 2004}; ^{Gil-Monte, 2014}; ^{Neffa, 2016}). The economic and social cost of PSRFs positions them as a public health problem and implies, as suggested by ^{Beck (2002)}, the questioning of traditional conceptions of social and labor relations in terms of a different way of understanding life and the new demands posed by the "society of risk." It is necessary to confront the parameters that govern modern society with the consequent problems derived from industrial society. The effects of modern society must be approached from new perspectives in the political, legal, cultural, economic, social, and scientific spheres concerning the prevention of risks and the factors that determine them.

An international reference for its significant work from public policies to micro levels is the consortium led by the British Institute of Work, Health and Organizations, in liaison with the University of Nottingham, the Social Dialogue department of the International Labor Organization, the Occupational Health area of the World Health Organization, and state technical agencies in occupational health in five countries (Germany, Italy, Holland, Poland, and Finland). PRIMA-EF is an acronym for Psychosocial Risk Management-Excellence Framework (^{Jain, Leka, & Zwetsloot, 2018}; ^{Vega, 2015}). This alliance aims to provide a framework for action to promote psychosocial risk management policies and practices in the European Union. It is a structure that makes it possible to apply policy and knowledge to manage effective practice in work organizations. PRIMA-EF consists of a practical, systematic, and evidencebased problem-solving strategy. The model comprises four necessary elements: scientific evidence, legal framework, guidelines, and recommendations for action and consensus based on social dialogue. It fundamentally focuses on the business field, assuming the need for learning, innovation, quality, and productivity as challenges, and determines a joint mode of action with governmental, legal, and sociallabor support to deal with PSRFs and their consequences.

According to the different political-legal contexts and regulatory frameworks of reference of the PSRF, two trends can be found that are directed from regulatory models, which can be undifferentiated and differentiated. In the case of the former, they are adopted by countries such as Spain, the United Kingdom, Ireland, Canada, and Australia; from these models, the PSRFs are legislated together with other occupational and professional risks. In the case of the differentiated regulatory models, also called specific, PSRFs are legislated in a particular way. They are adopted by countries such as Belgium, Denmark, France, Italy, the Netherlands, Sweden, Colombia, and Mexico (^{Fernández-Avilés, González, & Arague, 2019}).

PSRFs, as a scientific construct, must be differentiated from psychosocial factors and psychosocial risks since they are sources of occupational stress that threaten workers' health and have their own particular features. Among them, the most significant are: 1) they are difficult to objectify, 2) they affect other risks, 3) they have little legal visibility, 4) they are moderated by other factors, and 5) they may be difficult to modify. For their part, psychosocial risks affect the worker's fundamental rights, have global effects on the worker's health, affect the workers' mental health, and have forms of legal coverage (^{Ballester & Garcia, 2017}; ^{Moreno, 2011}; ^{Neffa, 2016}). The analysis of PSRF must be done from three levels that start from the definitions stated above. The first level corresponds to psychosocial factors present in any organization, which can be positive or negative. If they are negative, they constitute psychosocial risk factors; this is the second level of analysis. The third level corresponds to psychosocial risks, which result from psychosocial risk factors, i.e., the latter are predictors of work stress, violence, harassment, contractual insecurity, burnout, work-family conflict, and emotional labor (^{Díaz, 2011}; ^{Moreno & Báez, 2010}). Job stress is considered the most global psychosocial risk. Psychosocial risks have legal coverage in several countries (^{Hermosa & Perilla, 2015}; ^{Gil-Monte, 2014}; ^{International Labor Organization, 2016}).

Although PSRFs are a public health, legal, and political problem, the contributions are insufficient from a scientific point of view. A study that aimed to describe through bibliometric indicators the scientific literature related to psychosocial occupational risk, in the ProQuest and PsycArticles databases, during the period 2000-2010, with a descriptive quantitative design and using bibliometrics as a method, on the state of development of the subject in Colombia, reported as the most significant results that most of the studies are based on the Karasek demand-control model and related to the social support variable. This highlights that 39% of the 52 articles reviewed by the author are not based on theoretical models, which may be related to the fact that most of the models on which these investigations are based are on occupational stress and term these psychosocial risk factors stressors. The studies focus more on individual and intra-labor risk factors, and to a lesser extent on extra-labor ones. It is concluded in the study that research on psychosocial risk factors indicates that the development of the field is at an exploratory-descriptive level. Qualitative and mixed approaches are recommended, as well as quasiexperimental and experimental designs that make possible the analysis of psychosocial risk factors and their consequences on occupational health (^{Pulido, 2015}).

In the Latin American context, the Colombian Resolution 2646 stands out, through which provisions and responsibilities are established to identify, evaluate, prevent, intervene systematically, and monitor exposure to psychosocial risk factors at work, and their incidence in the appearance of pathologies derived from occupational stress. Failure to comply with these issues has legal consequences for companies and is enforced by the territorial directorates of the Ministry of Labor in Colombia (^{López, 2015}).

In the case of Mexico, the study of psychosocial risk factors that predispose to work-related stress and other risks is urgent for guaranteeing labor rights and preserving Mexican workers' physical and mental health. According to data from the Organization for Economic Co-operation and Development (OECD), the country is in first place in number of working hours, with employees working more than 2220 hours as an annual average (^{Organization for Economic Co-operation and Development, 2020}). It is also highlighted that the country ranks first in work stress, surpassing China and the USA (^{Rodriguez & Moreno, 2015}). Although complex, the need to widen studies and measures to reduce these figures is a priority not only at the organizational level but also at the governmental level and from a legal, political, and occupational health perspective.

In Mexico, according to ^{Juarez (2018)}, the most important psychosocial factors are those concerning psychosocial risk factors, such as demands due to the workload, the pace of work, high levels of responsibility and danger, workday, shifts and schedules involving physical effort, physical environment, work harassment expressed in psychological harassment by bosses and colleagues, stressful management due to abusive supervision, work-family imbalance expressed in a double workday, family to work effects, work to family effects, inadequate rewards and recognition expressed in work injustice, dissatisfaction with remuneration, lack of recognition and feedback, limited development possibilities, and job insecurity. Regarding the effects of psychosocial risk factors in Mexico, which translate into psychosocial risks, the most studied are burnout, distress, psychological weariness, and presenteeism. Risks predispose the worker to exhaustion, depersonalization, lack of achievement, anxiety, psychosomatic disorders, indifference, disillusionment, dislike of work, presenteeism due to health, and dissatisfaction. Juarez (2018) states that, despite descriptive studies, research from a preventive approach is scarce. The programs developed are generally only accessible to large companies, representing a minority of organizations in the country.

The proposals on PSRF in Mexico understand regulations to be the most effective prevention method (^{Aldrete & Cruz, 2013}; ^{Pando et al., 2008}; ^{Uribe, 2011}). The legislation in the country aimed at preventing risk factors began with the modification to the Federal Regulation on Safety and Health at Work (2014), which included a section that defines the Psychosocial Risk Factors at Work (PSRF-T), which are defined according to their effects and consequences within the framework of non-organic anxiety disorders, effects on the sleep-wake cycle, severe and adaptive stress, derived from the nature of the job functions, the type of workday, and exposure to severe traumatic events or acts of workplace violence due to the work performed. This was the starting point for the legalization and prevention of psychosocial risk factors through the Official Mexican Standard NOM-035-STPS-2018 Psychosocial risk factors at work-Identification, analysis, and prevention. It was approved as a draft on September 26, 2016, and entered into force in 2018. The psychosocial risk factors evaluated in the standard are: a) conditions in the work environment, b) workloads, c) lack of control over work, d) working hours and shift rotation that exceed what is established in the Federal Labor Law, e) interference in the work-family relationship, f) leadership and relationships at work, g) workplace violence (^{NOM-035-STPS-2018, 2018}).

NOM-035-STPS-2018 proposes three questionnaires. The first is an instrument that surveys severe traumatic events, and the other two aim to identify the presence of psychosocial risk factors in organizations. The first can be applied to any company, and the second and third will depend on the number of workers. The second, described in Reference Guide II, applies to companies with fewer than 50 workers. The third applies to companies with more than 50 workers, and a dimension is added to evaluate the organizational environment. This last version is described in Reference Guide III (GRIII) and was the one selected to conduct this research (^{NOM-035-STPS-2018, 2018}).

According to NOM-035-STPS-2018, the GRIII is not mandatory and makes it possible for workplaces to design and apply their instruments, depending on the characteristics of both the work environment and working conditions. However, NOM-035, numeral 7.5, establishes the requirements that must be observed:

The recent legislation on psychosocial risk factors in the workplace is extremely important to begin a process of promotion, prevention, and intervention in occupational health in Mexico; however, there are not enough studies reported that account for these instruments' psychometric properties. There are only two that serve as direct antecedents to this research. The first one had the purpose of presenting in a critical-creative way the psychometric properties of the scale Identification of Psychosocial Risk Factors and Evaluation of the Organizational Environment with a sample stratified by gender of workers of a service company (n=114) from Mexico City and the interior of the Republic. According to the results of this study, the scale shows reliability, but the results of the Confirmatory Factor Analysis (CFA) do not validate by adjustments and qualifications. Therefore, it was recommended to the Secretariat of Labor and Social Security of the Mexican government (Spanish: Secretaría del Trabajo y Previsión Social, STPS) to consider the discrimination of items, to add items, to evaluate a proposal of a sample size of 10 subjects per item, and to analyze by principal components and with second-order factor analysis (Uribe, Gutiérrez, & Amézquita, 2019).

In the case of the second study, its aim was to evaluate the reliability and validity of the 5 categories of the GRIII. This questionnaire was applied to a sample of 1,247 full-time workers working in a coffee and food company in Mexico in September and October 2019. The research results confirm that the categories proposed by NOM-035-STPS-2018 comply with a minimum Cronbach's Alpha of .70. However, the construct validity criteria derived from an Exploratory Factor Analysis and a Confirmatory Factor Analysis are not confirmed. It is recommended to review the theoretical conceptualization of the PSRF involved and the design of the questionnaire (^{Littlewood-Zimmerman, Uribe, & Gurrola, 2020}). It is necessary to highlight that the legal implications and the processes derived from PSRF assessments in Mexico are still in the process of development and delimitation from the legal framework; although there is legislation, the STPS states that:

"The standard does not contain any instrument for measuring stress, so it is false that workplaces will have to pay fines due to worker stress. It is not an instrument to perform a psychological evaluation of workers, nor does it oblige workplaces to perform it, and it does not contain a questionnaire to identify mental disorders or a method to identify internal psychological variables of the individual, such as attitudes, values, personality, among others." (^{Secretaría del Trabajo y Previsión Social del Gobierno de México, 2019}).

This study aimed to evaluate the validity of the psychometric properties of the GRIII proposed in 2018 by the Secretariat of Labor and Social Welfare (STPS) in the NOM-035-STPS-2018 in Mexico. The construct validation with confirmatory factor analysis (CFA), Structural Equation Modeling, and reliability levels through Cronbach's Alpha and Omega were then proposed. Determining the psychometric properties of the instruments proposed by NOM-035-STPS-2018 will make possible a more rigorous process of identifying and managing psychosocial risk factors and provide the STPS and researchers with scientific results that validate, confirm, and improve the legislated proposal since they are applied according to regulatory principles. It is necessary to shed light on the validity of the proposed instruments.

Method

Results

Construct validity analysis

When the reliability and confirmatory analyses were performed on the originally proposed structure, it was determined that Cronbach's alpha coefficient is valid above .70 and the overall coefficient .90, except in the categories of work environment and organization of working time. However, they are close to the desired value.

Concerning the CFA, the absolute fit indices are not the desired ones (Table 3), according to ^{Jöreskog and Sörbom (1981)}. Several items were obtained as non-significant--values lower than 0.1--and are shown below:

Item 25: During the workday, you can take breaks when you need them, from the dimension of the activity factors, from the domain, workload, and the dimension of lack of control and autonomy in the work. The way it is worded means it can be answered according to the workload and working time, which determines that it is valued, if it measures what is proposed, according to the domain and the category where it is located (the worker can take breaks or not and this does not necessarily mean that they have autonomy and control or not in their work). Something similar happens with item 30: When changes occur in my work, my ideas or contributions are considered in the category of factors specific to the activity, in the domain of lack of control over the work and in the dimension of insufficient participation and management of change. It should be investigated whether how items 25 and 30 are posed is accurate according to what is to be measured or if it generates some kind of conflict and doubt in the workers when answering them.

Table 3 Cronbach's alpha values by GRIII factor and CFA values in model 1 

Model 1	Reliability values	Values in the factorial model fit
Category	Alpha	CMIN/DF	GFI	CFI	RMSA
Model total	0.924	4.069	0.485	.454	0.08
Work environment	0.624
Factors specific to the activity	0.759
Organization of working time	0.669
Leadership and relationships at work	0.913
Organizational environment	0.774

Source: created by the author

The following items are also non-significant:

Item 65: Do you deal with very angry clients or users? This question focuses on the relationships established with clients. It should be reviewed because, in this case, it is not significant in the dimension of emotional and psychological burdens of the workload domain of the activity factors category. This is also the case with item 66: The work requires dealing with people in need of help or sick people; item 67: To do the work, I must show feelings different from my own; and item 68: The work requires addressing situations of violence of this dimension. Item 66, due to the mission and vision of the company, may not apply. In item 68, the term violence may be ambiguous in this context. Theoretically, several positions define it, and it is also polysemic in colloquial language. It is recommended to review these items more precisely according to what they seek to measure. It is also suggested to revise item 54 because it is not significant. Item 54: There is continuous turnover in my work in the organizational environment category, of the insufficient sense of belonging and instability domain, of the labor instability dimension is more focused on or understood as personnel turnover, and it should be reviewed if it should go in this category of organizational environment concerning labor instability and according to the characteristics of the company.

Item 69: Late communication of work issues; 70: They hinder the achievement of work results; 71: Little cooperation when needed; and 72: They ignore suggestions to improve their work in the leadership and work relations category of the social relations at work dimension, and in the deficient relationship with the employees they supervise domain, were also not significant. This dimension seeks to measure the relations of managers and supervisors with subordinates. In the test items, given what is asked, it would seem that the aim is to identify bad practices in the processes of organizational management and direction rather than the leadership practices in the company, as stated in the category, the domain, and the dimension if the starting point is the theoretical bases of the leadership construct from a scientific point of view.

In this model 1 (Table 3), the reliability value Omega was not included as an additional analysis to the calculation of Alpha since the original model of the GRIII was replicated. Furthermore, since it does not present concatenations (links between errors), the Omega is not relevant (^{Viladrich et al., 2017}). However, the model proposed in the present work (Table 4) eliminated some items and dimensions that were not significant, in addition to incorporating concatenations (Figure 1), so the Omega was integrated. The latter model showed adequate absolute goodness-of-fit indices, whose values ≤ 3 are considered a good fit (^{Jöreskog & Sörbom, 1981}) and Cronbach's alpha and Omega coefficients, as seen in Table 4.

Table 4 GRIII alpha and omega by factor and CFA values in Model 2 

Model 2	Reliability values		Values in the factorial model fit
Category	Alpha	Omega	CMIN/DF	GFI	CFI	RMSA
Model total	0.924	0.926	2.68	0.654	0.722	0.066
Work environment	0.624	0.771
Factors specific to the activity	0.759	0.819
Organization of working time	0.669	0.788
Leadership and relationships at work	0.921	0.929
Organizational environment	0.8204	0.856

Source: created by the author

Note: Factor 1 = Work environment; Factor 2 = Activity-specific factors; Factor 3 = Organization of working time; Factor 4 = Leadership and work relationships; Factor 5 = Organizational environment

Source: created by the author

Figure 1. CFA of model 2 of NOM-035-STPS-2018 

It is recommended that the company use other instruments and methodologies to identify and evaluate the PSRFs, considering the items and dimensions which were excluded according to the statistical results. The results on the construct validity of the GRIII of the NOM-035-STPS-2018 indicate that the instrument alone is not sufficient to address all the Psychosocial Risk Factors proposed by the standard; therefore, its application must be evaluated with ethical and scientific rigor, particularly the results of the dimension of a poor relationship with supervised employees, of the work relationships domain belonging to the category Leadership and work relationships (although they are kept in the model for a better fit), and the dimension of emotional and psychological burdens of the workload domain of the category activity-specific factors.

Evaluating these dimensions with other instruments and the rest of the items that were not significant is suggested. These results put the scientific discussion on the validity of the instruments proposed by NOM-035-STPS-2018 in context.

Discussion

According to ^{Arias-Galicia (2018)}), psychological instruments in the organizational area in Mexico should be improved. The rigorousness protocols regarding the theoretical approaches that support the instruments, the evidence of reliability, validity, biases, and shortcomings, and the interpretation of results in a contextualized manner should constitute principles to develop an effective use with scientific rigor of the techniques being proposed. This study aims to evaluate the validity of the scale's psychometric properties for the Identification of Psychosocial Risk Factors and Evaluation of the Organizational Environment proposed by NOM-035-STPS-2018 in Mexico.

Currently, there are not enough studies to compare these results and triangulate the information obtained in order to generate final proposals on the instrument. Only two studies (which constitute the most important direct antecedent of this research) highlight that reliability is relevant for instrument measurement and that the STPS has proposed NOM-035-STPS-2018 without providing the instrument's psychometric properties and suggests its application throughout the country. The questionnaire was analyzed psychometrically to propose improvements to the STPS, and it is recognized in both research reports that there are no valid instruments to measure psychosocial factors jointly in Mexico. Therefore, there is no Mexican theoretical framework to support the NOM-035-STPS-2018 proposed by the STPS. The proposal is perceived as strictly pragmatic, as research intended to criticize and propose improvements with an exploratory scope (Uribe et al., 2019; ^{Littlewood-Zimmerman et al., 2020}).

Given these results, it is considered pertinent to present, in the theoretical-methodological aspect, evaluations regarding the application of the GRIII in this specific company:

It is assumed that the results obtained are part of a complex process of research and measures that are under construction to identify, control, intervene, and reduce PSRFs in this particular company in northern Mexico, considering the validity criteria of the instrument applied critically and proactively in terms of scope and limitations.

The need for studies in other contexts is recommended, as well as to assess replicability in terms of improving the application processes of the instrument and the possible results that should be subject to criteria of reliability, trustworthiness, and validity in order to address PSRFs in Mexican companies. The authors of this study recognize the difficulty of finding national scientific reports that make it possible to contrast results and propose, if necessary, improvements to the necessary and useful proposal of the STPS in NOM-035-STPS-2018.

Conclusions

The results on the construct validity GRIII of the NOM-035-STPS-2018 determine that the instrument alone is not sufficient to address all the Psychosocial Risk Factors that NOM-035-STPS-2018 proposes; therefore, at least in this company, the dimensions of a poor relationship with supervised collaborators, of the work relationships domain of the category Leadership and work relationships, and the dimension emotional and psychological burdens of the workload domain of the category activity factors should be evaluated with ethical and scientific rigor. It is suggested that the dimensions of a poor relationship with supervised collaborators of the category Leadership and work relationships and the dimension emotional and psychological burdens of the workload domain of the category activity factors, and non-significant items be evaluated with other instruments or methodologies, such as focus groups and psychological interviews, to gain clarity regarding what is to be measured by the evaluators and what is understood by the workers in the companies. It is necessary to triangulate quantitative and qualitative results and to evaluate the possible biases of the GRIII according to authentic situations and business contexts. It is considered pertinent to state that similar studies should become more prominent from a political and scientific point of view to contrast and discuss contributions in order to improve the instruments proposed by the STPS in NOM-035-STPS-2018.