Organizational Tolerance for Workplace Harassment: Development and Validation of the POT Scale

Abstract

The current investigation defines the organizational tolerance (OT) construct and statistically assesses its measurement instrument, the perceived organizational tolerance for psychological workplace harassment (POT) scale, carried out to evaluate the level of tolerance, negligence, or even connivance that can be shown by an organization when it deals with an inappropriate act occurring within its scope. Tolerance of such acts has been identified as a hindrance when trying to establish an effective and sustainable strategy for the well-being of workers. A survey that measures the construct was distributed, and 195 employed workers answered. In the first stage of analysis, a scale reduction process was applied to the obtained data using a factor extraction method, and afterward, confirmatory factor analysis was performed using structural equation models. The results validated the scale as a model of five factors: Promotion, feedback, ethics, coherence, and training. These findings indicate that this scale is acceptable as a quantifier of a key issue, namely, the diligence of the organization when dealing with psychosocial risks at work. This new construct is anticipated to be incredibly useful for measuring as much research as possible on the behavior of organizations when they deal with negative acts, with the aim of promoting sustainable healthy working environments.

1. Introduction

The significant effects of psychosocial risks in the workplace can be assessed from different perspectives, such as the consequences to both the health and performance of workers or the costs to the Public Health System, among others. For example, depression due to unhealthy labor conditions cost 9.9 million labor days in the United Kingdom between 2014 and 2015 (Health and Safety Executive, 2015) and more than 617 million euros in Europe [1]. Some authors even remark that such situations can put at risk the liability of the organization [2]. Others consider healthy organizations, regarding the well-being of workers, as a new approach of sustainability and sustainable development [3]. One of the most harmful factors at work is harassment, which the National Institute for Health and Safety in the Workplace defines as “exposition to psychological violent behavior, which takes place repetitively along the time […]” [4]. When this behavior takes place at work, it is labeled as workplace harassment, workplace abuse, labor harassment, mobbing, or bullying.

In this context, tolerance means allowing something, such as certain behaviors, to occur and continue even though it is considered illicit [5]. When an organization adopts a tolerant attitude, it is called organizational tolerance (OT). Several authors using different models have acknowledged the role of tolerance in workplace harassment. For example, Cortina [6] focused on tolerance and deemed it one of the main predictors of harassment. Similarly, Einarsen [7] suggested that a high level of tolerance is a precursor to harassment. Following this idea, Padilla’s model of the toxic triangle [8] emphasizes the role of a permissive environment in workplace harassment. This factor can be added to the traditional model, which involves only the victim and the perpetrator. This groundbreaking point of view suggests that the organization’s attitude, which shapes the working environment, is a crucial factor in workplace harassment. Most authors have recognized the weight of tolerance for workplace harassment, which can be assessed by the actions taken by the organization to prevent harassment or address it when it happens.

Despite the importance of these theoretical models, OT has not received enough attention in empirical investigations. Traditionally, investigations of workplace harassment precedents have focused on the characteristics of the workplace according to the job demand-resources model [9,10,11]; work–family conflict [12]; socioeconomic and political traits [13]; or even regional characteristics [14]. The current investigation differs from previous ones because it focuses on the operative definition of OT. To this end, a measurement tool was used to quantify the level of in/tolerance for workplace harassment and analyze its psychometric properties.

The literature reveals different instruments that have been used to evaluate OT, but such studies have focused on sexual harassment [15,16]. In an empirical investigation [17], the organizational tolerance for sexual harassment inventory (OTSHI) scale [16] for sexual harassment was adapted to study what the authors labeled “Generalized Workplace Abuse”. With this scale, participants were provided a hypothetical scenario: “One of the supervisors/bosses in your department often spreads gossip and rumors about new employees. This supervisor/boss attempts to isolate these employees from the rest of the department”. The respondents had to decide how risky it would be to make a complaint, specify whether they thought that their complaints would be taken seriously, and, finally, propose the hypothetical effects of filing the complaint. This approach provided a three-item scale for assessing tolerance, with the main focus on spreading rumors and isolating employees.

Given the lack of an efficient and brief specific instrument to measure the tolerance of an organization’s management for psychological workplace abuse, the investigation reported herein has two purposes. The first is to develop a novel instrument in Spanish, and the second is to test its psychometric properties. In the first step toward accomplishing these goals, both the OT definition and its organizational dimensions are shown. These dimensions can be found in the literature as key indicators of organizational in/tolerance for psychological abuse at work.

1.1. Defining OT and Its Dimensions

To define the term “OT”, Bowling and Beehr [18] described the kind of organization that allows harassment. Schneider, Pryor, and Fitzgerald [19] focused on the other side of the dimension and referred to some organizations with a “zero tolerance” approach to workplace harassment.

Other authors have included aspects related to poorly designed and chaotic workplaces in the definition of OT [20] or even those involving connivance with the perpetrator instead of supporting the victim [21]. In current investigation, OT is defined as the permissiveness, either deliberate or inadvertent, of the staff responsible for handling negative acts that can take place within the organization. OT can go even further to the point of collusion, whereby the staff responsible for addressing these issues not only allows but also fosters harassment [22].

1.2. Training on Psychosocial Risks at Work

Training staff is one of the organizational strategies for addressing workplace harassment and preventing labor risks in general. Some examples [23] have shown that organizations benefit from training company executives. For example, some studies have demonstrated a positive relationship between worker training and desirable psychosocial effects, such as employee motivation [24,25], job satisfaction [26], and company profit [27]. In Spain, Escartín [28] found that training on psychosocial risks was a precursor to several quantifiable improvements in the working environment. Considering these findings, any instrument that aims to evaluate the in/tolerance of an organization for psychological workplace harassment should include some items related to training related to psychosocial risks at work.

1.3. Psychosocial Risk Assessment

Any organization that wants to prevent workplace psychosocial risks, including workplace harassment, needs to evaluate the presence of such risks, as indicated in the current Spanish law [29]. These norms require companies to take preventive actions. Moreover, they specify that the first step in prevention is the pre-evaluation of risks to the safety and health of workers. From the results of this assessment, the second step for the company is the assessment of both preventive and improvement measures. Conducting this kind of evaluation using a survey is quite common in countries of the European Union [30]. Therefore, gauging the company’s commitment to the continuous assessment of its labor conditions is an imperative part of an instrument that measures the organizational in/tolerance for workplace abuse.

1.4. Organizational Coherence

When dealing with an evident decrease in the mental health of workers, the onus is on the organization to diligently act on the situation, in which case the outcomes are positive [31]. In another example, the managers of a company were involved in a stress reduction program, which increased profit and improved labor motivation and satisfaction [32]. Biron and Karanika-Murray [33] also remarked on the need to involve as many company departments as possible (health, human resources, etc.) when establishing initiatives to improve working conditions. Thus, on the basis of the literature, the measurement scale of organizational in/tolerance for workplace psychological abuse should assess whether there is organizational coherence that links preventive and improvement actions with the support of all actors to increase the likelihood of success.

1.5. Organizational Ethics

In the last decade, ethics has played an important role in organizations. Ethics has been identified as a key characteristic of positive organizations [34] because ethical leadership promotes healthy organizations [3], and its benefits have been empirically studied [35]. Ethical behavior makes an organization’s actions to create a healthy environment credible. Other investigations have added that employees adopt ethical behavior when it is demonstrated by their leaders [36]. The ethical behavior of leaders has also been considered a prerequisite of the optimum performance of both the employees and the organization itself [37].

The studies cited herein include data that show how an organization’s ethical behavior toward psychological workplace harassment can be realized through two different actions. First, when harassment happens, the organization should responsibly promote transparency using feasible tools to listen to their workers’ opinions. This is called the organizational voice [38], and it encourages victims to report any suffered inappropriate acts. Second, when a complaint is filed by a worker, the actions that are implemented by the organization should aim to protect the victim from the perpetrator [39]. However, many workers have seen their complaints go ignored, or victims are punished for whistleblowing by reporting harassment or other abuses [40]. Thus, items regarding ethical performance on the in/tolerance scale should include queries about the style in which the management of the organization addresses complaints that are received, as well as the consequences for complainants.

1.6. Implementing a Healthy Workplace Environment

A healthy work environment promotes both the well-being of employees and organizational goals [41]. When building a healthy work environment, a key issue is the prevention of psychosocial risks, especially harassment. Thus, a positive environment requires that the company be proactive in the prevention and handling of harassment. However, this is not enough: Specific actions must be launched by the company to reach the goal of building a healthy work environment [42].

In essence, according to the reviewed literature so far, some indicators from the inventory should include information about the performance of a firm concerning workplace harassment.

2. Materials and Methods

In the first stage, the items of the perceived organizational tolerance for psychological workplace harassment (POT) scale were selected. First, the above-mentioned studies were examined to identify the dimensions that are conducive to an operative definition of OT. The authors who designed the scale established the items in the first version of the questionnaire as a result of those dimensions. After this, some professionals from the fields of labor psychology and medicine reviewed this first version. The experts were encouraged to determine whether the meaning of each item was clear and to test the degree of adequacy of each item as an indicator of OT. The output of the task was an 18-item OT scale, which is the POT scale. The level of agreement with each item is measured using a five-element Likert-type scale that ranges from 1 (fully disagree) to 5 (fully agree).

Given that the present research did not involve any interventional studies involving humans and that the collected data were both anonymous and untraceable, ethical approval was not required. The first version of the survey was completed by the respondents of the empirical study, and other sociodemographic information was collected. This step of the study was carried out by distributing questionnaires to different organizations by collaborators of the research team, and they performed the task after receiving precise instructions to ensure that the administration procedure was implemented consistently. Participants were informed of the goal of the study and of the anonymity of the data collected, and they expressed their consent verbally. Then, they completed a booklet containing the diverse scales of the study.

2.1. Participants

The survey was answered by 195 active workers who voluntarily took part in the investigation. The sample consisted of 44.1% women and 55.9% men. The participants belonged to different companies located in Spain. The data collection was finished at the end of 2015.

Of all the respondents, 75.9% had completed university studies, 12.8% had completed secondary education, and 8.7% had received some professional training. The kind of enterprise that employed these workers was included in the questionnaire: 61% worked in a public organization, 17.4% worked in the private sector, and 21% worked in a private company that was supported by public funding. The respondents who had no other employees in their charge amounted to 66.7%, middle-level managers formed 28.2% of the sample, and the rest were owners or senior managers.

The economic sector of the respondents’ companies was 33.8% education/research, 30.8% services, 15.9% technology, and 7.7% health, with the remaining participants belonging to other fields. Most of the companies had a workforce of more than 200 workers, while 12.8% had 50–199 workers. The rest of the companies had fewer than 50 workers.

2.2. Data Analysis

The first step of data analysis was to perform both descriptive statistical analysis and exploratory factor analysis (EFA) [43] using SPSS 20. The EFA grouped the scale into five factors. Afterward, the behavior of the extracted factors was confirmed by confirmatory factor analysis (CFA) [44]. The size of the sample was suitable because [45] a minimum of 150 answers, or at least five answers per item of the scale, is required.

3. Results

Table 1. Sociodemographic analysis.

Parameter	Minimum	Maximum	Average	DT
Age	22	65	47.75	8.72
Working life	0	42	20.09	9.62
Workers in their charge	0	120	2.04	12.25

summarizes the sociodemographic data on the participants.

As can be verified from the

Table 2. Indicators on the perceived organizational tolerance for psychological workplace harassment (POT) scale and their descriptive statistics.

Items	M	DT	Skewness	Kurtosis
POT1. My organization gives high priority to dealing with harassment complaints	3.46	1.04	−0.22	−0.53
POT2. If I report a harassment incident, my organization will try to deal with it diligently.	3.37	1.14	−0.17	−0.85
POT3. My organization offers us training in communication skills or conflict resolution.	3.67	0.99	−0.37	−0.48
POT4. My organization has clearly explained to us how to act if we suffer workplace harassment.	4.07	0.82	−0.66	0.28
POT5. My organization has clearly explained to us how to act if we are not sure whether we are suffering workplace harassment.	4.09	0.79	−0.74	0.62
POT6. The workers’ representatives have collaborated with my organization in the design of a harassment prevention plan.	3.85	0.90	−0.61	0.14
POT7. My organization encourages us to report any potential harassment incident we may suffer as soon as possible.	3.79	0.89	−0.40	−0.07
POT8. My organization encourages us to report any potential harassment incident suffered by a colleague as soon as possible.	3.86	0.87	−0.51	0.17
POT9. My organization often updates the information we receive on workplace harassment actions.	4.11	0.74	−0.54	−0.07
POT10. My organization will deal with any harassment incident in the same way if either a manager or an employee is accused of workplace harassment.	3.78	1.01	−0.63	−0.02
POT11. I have answered the psychosocial risk survey carried out in my organization.	3.67	1.34	−0.65	−0.87
POT12. My organization has clearly explained to me the results of the psychosocial risk survey.	3.87	1.13	−0.78	−0.34
POT13. My organization has launched actions based on the results of the psychosocial risk survey.	4.23	0.89	−1.32	1.80
POT14. The decisions taken by my company based on the results of the psychosocial risk survey have improved working conditions.	4.23	0.76	−1.02	1.31
POT15. In my organization, the workers who have reported workplace harassment have suffered reprisals.	3.51	1.18	−0.68	−0.28
POT16. My organization has publicly explained lessons learned when they have dealt with cases of workplace harassment.	4.18	0.75	−0.77	0.70
POT17. My organization has respected the anonymity of the people involved in reported incidents of workplace harassment.	3.50	1.00	−0.62	0.51
POT18. My organization will not try to stop me from reporting a workplace harassment incident.	3.57	1.23	−0.76	−0.23

, the resulting values of the descriptive statistics were reliable [46]; that is, less than 2 for skewness and 7 for kurtosis.

Once the statistics describing the scale responses were analyzed, the factor extraction was tackled, as represents

Table 3. Factorial analysis: Saturation matrix for the POT scale.

Item	1	2	3	4	5
POT 1	0.03	0.16	0.72	0.31	0.26
POT 2	0.05	0.21	0.86	0.35	0.10
POT 10	0.25	0.23	0.54	0.09	0.18
POT 3	0.35	0.28	0.28	0.20	0.21
POT 4	0.90	0.18	0.18	−0.01	0.13
POT 5	0.87	0.13	0.15	0.01	0.22
POT 6	0.55	0.30	0.20	0.27	0.17
POT 9	0.70	0.12	−0.10	0.08	0.27
POT 11	0.07	0.66	0.20	0.33	0.01
POT 12	0.09	0.87	0.20	0.26	0.02
POT 13	0.23	0.77	0.18	0.07	0.14
POT 14	0.35	0.62	0.07	0.06	0.16
POT 15	0.07	0.24	0.29	0.68	−0.02
POT 17	0.07	0.22	0.33	0.41	0.15
POT 18	0.05	0.154	0.16	0.89	0.00
POT 7	0.34	0.07	0.34	0.03	0.84
POT 8	0.39	0.14	0.30	0.04	0.75
POT 16	0.24	0.06	0.01	0.01	0.25

. Before the analysis, the reliability of the data was tested. The Kaiser–Meyer–Olkin (KMO) index (0.842) and Bartlett’s sphericity test (χ² (153, N = 195) = 2317.904, p < 0.000) supported the use of EFA. Maximum likelihood was chosen as the factor extraction method. This method determines the weighted estimations of the parameters that maximize the probability of producing the observed data [47]. Varimax with Kaiser normalization was selected as the factor rotation method. This method minimizes the number of variables that have high loadings on each factor.

The matrix determines the five factors that segment the 18 items of the POT scale. The bold content in Table 3 highlights the higher loadings. Applying the cut-off point reported in most studies [48] leads to items 3 and 16 being not significant because of their low loadings. However, other authors [49] have considered the possibility of using more stringent cut-offs, as was done in this investigation because of the theoretical interest of these items. Although all of the items’ loads seem to be distributed among different factors, each item was considered to belong to the factor on which it had the highest load.

As a result of the grouping obtained by EFA and considering how many items belonged to each factor, the following factors were determined. The first factor, training, was formed by items related to general training, specific training on workplace harassment, and periodical retraining. Some of these items were “(It) offers us training […]”, “My organization has clearly explained to us how to act […]”, and “(It) updates the information […]”.

Assessment was the second factor, and it queried whether the companies ask their workers for their feedback about the quality of the working environment and assess any evidence of psychosocial risks at work. The items included in this factor ask about the psychosocial risk survey and how the company has tried to improve working conditions according to the results of this tool.

The third factor, coherence, was determined by three items that assessed whether the organization’s performance is coherent, not only in theory but also in practice, and prioritizes solving issues related to workplace harassment. Coherence-related questions included “[…] (It) gives high priority to dealing […]”, “[…] my organization will try to deal with it diligently […]”, and “[…] (It) will deal with any harassment incident in the same way regardless of who is accused of such a behavior, be it an executive or a mere employee […]”.

The indicators that belonged to the fourth factor, ethics, analyzed the potential support that a company provides perpetrators. The statements in these items question the likelihood that victims will be punished after reporting abuses, whether their anonymity would not be respected, or whether the company has tried to deter them from using the complaint process.

The fifth factor, promotion, had three items that analyzed the real interest of the company in the worker’s commitment so as to resolve negative situations. Its indicators included sentences such as “My organization encourages us to report any potential harassment episode […]”, among similar statements.

The confirmatory analysis was carried out from the factors previously mentioned using the structural equation method. The model was respecified to improve its adjustment, thus allowing the correlation of some measured errors within the same factor [50].

The goodness-of-fit index (GFI), the adjusted goodness-of-fit index (AGFI), the comparative fit index (CFI), the incremental fit index (IFI), and the Tucker–Lewis index (TLI) must all be over 0.90. Additionally, the standardized root-mean-square residual (SRMR) and the root-mean-square error of approximation (RMSEA) must be below 0.80, and the χ²/df value must be lower than 3.0. The adjusted model was considered suitable [51,52,53], given that the following indexes were obtained: χ² (119) = 320.784, GFI = 0.842, CFI = 0.910, and RMSEA = 0.093. The results of the CFA can be seen in Figure 1.

The global scale had a good reliability value (Cronbach’s Alpha = 0.901). Moreover, the limitations of Cronbach’s alpha coefficient can be addressed by using CFA factor loadings to provide a more accurate estimation of reliability through composite reliability (CR), developed by Werts, Linn, and Joreskög [54] and recommended by Rios and Wells [55]. Scores should be higher than 0.70.

In order to show that the five factors did not converge to a degree that would render them redundant, we tested convergent and divergent validity with CFA, following Fornell and Larcker’s [56] recommendation. The analysis of the average variance extracted (AVE) reflects the total quantity of the variance of the indicators captured by the latent construct, and AVE values are recommended to be higher than 0.50.

Table 4. Results of the reliability study.

Factor	Item	λ	ε	AVE	CR
Training	POT3	0.62	0.65	0.92	0.98
	POT4	0.90	0.12
	POT5	0.93	0.08
	POT6	0.67	0.45
	POT9	0.70	0.26
Assessment	POT11	0.77	0.74	0.61	0.86
	POT12	0.92	0.16
	POT13	0.80	0.27
	POT14	0.60	0.36
Coherence	POT1	0.84	0.31	0.64	0.84
	POT2	0.92	0.19
	POT10	0.65	0.61
Ethic	POT15	0.88	0.30	0.54	0.78
	POT17	0.65	0.56
	POT18	0.76	0.63
Promotion	POT7	0.94	0.08	0.72	0.87
	POT8	0.92	0.11
	POT16	0.30	0.51

shows CR and AVE for each factor or construct. For each item or indicator, its factorial loading or correlation with the factor and its measure error ε are reported.

As was previously mentioned, indicator 16 has a poor loading, which is just at the limit of the acceptance level [45]. This fact was acknowledged, but, for this analysis, the indicator was not removed because of its relevance in light of the final results and its theoretical merit.

In the last step, a discriminant validity analysis was performed. It is represented in

Table 5. Discriminant validity assessment.

Factor	Training	Assessment	Coherence	Ethics	Promotion
Training	0.96	0.39	0.35	0.27	0.65
Assessment		0.78	0.53	0.55	0.29
Coherence			0.80	0.68	0.53
Ethic				0.74	0.25
Promotion					0.85

. Its value is accepted [56] when the square root of the average variance extracted is higher than any of its correlations with other constructs. The result of this calculation is shown in the following table. The square root of the variance is the element on the main diagonal of the matrix presented in the table. The relations among constructs are shown by the elements that are not in the diagonal. The interpretation of the table proves that discriminant validity is satisfied in this study.

4. Discussion

The current investigation has achieved its goals. The first is to completely define OT because it is deemed necessary to clearly differentiate it from other terms. For example, corporate social responsibility [57], perceived organizational support [58], civic organizational behavior, culture, and climate [59], and organizational commitment [60] share some common features with OT; however, at their core, they are different constructs. A conclusion drawn from examining the literature is that assessing the diligence of a company in its prevention and fight against workplace harassment is a rarely used approach compared with other traditional constructs used to investigate the above-discussed problems. This observation is paradoxical, considering that some investigations [6] take OT to be the single most critical term for predicting abuse.

As the other main objective in the present investigation, the psychometric properties of the measurement scale of the construct were also analyzed. The results show that the scale has satisfactory psychometric properties. The loadings of the items and the factor structure are adequate. The reliability values for each factor are high, despite the limited number of items. Moreover, the results support the validity and discriminant criteria for the five factors of the POT scale.

From the literature review, few to no studies were found that applied similar scales. The POT scale could be considered similar to the negative acts questionnaire (NAQ) [61,62], which has been frequently used to assess the intensity or frequency of inappropriate acts; however, the POT scale is not intended to be such an instrument. Instead, its purpose is to identify whether the management of an organization is negligent or even complicit when negative acts happen.

4.1. Limitations and Future Research Areas

This study had several weaknesses and limitations that should be mentioned. Issues related to the size and representativeness of the samples were that the sample was non-random and non-representative; thus, the results could not be generalized to other populations. Moreover, all the data were obtained from self-reports, which could include a source of uncontrolled error from the common variance. Validation using independent criteria, such as health records or responses from relatives and friends, will be explored in future studies with this measure.

Among the limitations of this study, the low loading of item 16 must be underlined. Further research to improve the scale will have to decide on its suitability, although, because of its theoretical relevance, it is suggested that it be rewritten. Thus, instead of “My organization has publicly explained how they have dealt with the cases of workplace harassment”, a new item, which must be validated, is proposed, namely, “My organization has publicly explained positive lessons learned from solving reported harassment cases”.

Another potential improvement is the use of two completely different samples: One for EFA and another for CFA. This has been suggested by several authors [63]. According to their arguments, it is better to first study the structure of a dataset and then try to confirm this structure with a different source.

According to several experts [64], the values obtained to fit the model using structural equations are acceptable but not exceptional. For a well-fitted model, a CFI index closer to 0.9 and an RMSEA under 0.09 are recommended.

This investigation suggests a future line of research that focuses on OT. This proposal could help to not only improve knowledge but also prevent harassment. The developed POT scale can be used to generate different models that establish relationships between variables and aim to determine the influence of OT on organizational psychology, something that the authors encourage.

Another useful prospective strategy is to use the POT scale in different countries to assess whether culture is a factor that affects perceived tolerance.

4.2. Implications

The current law on the prevention of risk in the workplace [65] claims that it is compulsory for companies to evaluate and prevent any risk to workers’ health posed by their daily work. Society has traditionally established that mitigating workers’ physical risks, such as strokes, falls, and poisoning, is of the utmost importance.

However, it must not be forgotten that the complete definition of health [66] includes “[…] physical, mental and social well-being […]”. Thus, the current investigation suggests updating the most common surveys (methods istas21, CoPsoQ, FPSICO, MC-UB, etc.) that are used to assess psychosocial risks at work and adding issues similar to those dealt with in this study, that is, those requiring the diligence of the company to solve.

The different factors of the POT scale detect the lack of several management skills (training, exemplary behavior, etc.) that should be mandatory for any leader who wants to manage a team in a healthy manner. The present investigation bridges the gap between those factors and two current trends—authentic leadership and the psychology of sustainability. If the factors of the POT scale are matched with the characteristics of authentic leadership [67] and sustainability [68], then some overlapping aspects could be identified, especially those related to ethical or honest behavior.