The Impact of the Quality of Interpersonal Relationships between Employees on Counterproductive Work Behavior: A Study of Employees in Poland

: The purpose of the article is to determine how the quality of interpersonal relationships at work (QIRW) a ﬀ ects the extent of counterproductive work behavior (CWB), and whether this impact is moderated by employees’ demographic features (education, age, sex, length of service and type of work). These questions are particularly important for organizations that want to function sustainably, because counterproductive behavior also includes wasting resources, polluting the environment and using environmentally unfriendly products. The research objectives were met using a survey conducted in 2018 among 1488 professionally active people in Poland. Structural equation modeling (SEM) was used to analyze the empirical data. The proposed theoretical model was intended to determine how particular categories of relationship quality a ﬀ ect dimensions of CWB (which included taking into account employees’ aforementioned demographic features). I determined that relationship quality has an inverse relationship with counterproductive behavior of employees (the higher the quality, the lower the propensity for CWB), but there are also many paradoxes that I discuss in detail. Moreover, this impact is signiﬁcantly moderated by employees’ demographic features (mainly education, type of work, length of service and sex). I also discuss the theoretical contributions, practical implications and limitations of this study, and directions for future research. and threatens the wellbeing of an organization, its members, or both.”


Introduction
The quality of interpersonal work relationships determines the behaviors that employees engage in both at work and in their private lives [1,2]. As a rule, high quality relationships translate beneficially into, among others: commitment, performance, motivation, innovation, error detection, OHS, employee green behaviors (EGB), teamwork, helping others, internal and external organizational communication, absenteeism, conflict and resilience to negative events. Conversely, low quality relationships between employees have a detrimental impact on these aspects of an organization's operation [3][4][5][6][7][8][9].
Therefore, counterproductive work behaviors (CWB) must also be influenced by the quality of interpersonal relationships at work. This is because these behaviors are influenced by the group [10,11] and are a behavioral (adaptive) response to certain workplace conditions [12,13].
This interrelationship is quite apparent, so, if these relationships are of a high quality, employees will tend less towards CWB. By the same reasoning, if relationship quality is low, counterproductive behaviors will be engaged in more often. In practice, however, there are many paradoxes, which is to say that negative relationships can have beneficial effects, while positive relationships can have detrimental effects (for the organization and its stakeholders) [14,15]. One such example is the case in 1. to determine how the quality of interpersonal work relationships affects the extent of counterproductive work behavior, 2.
to determine whether the impact of interpersonal work relationship quality on the extent of counterproductive work behavior is moderated by employees' demographic features (education, age, sex, length of service and type of work).
The goals will be achieved using a survey conducted in March 2018 on a sample of 1488 professionally active people in Poland. To achieve the objectives, two research hypotheses were adopted (see the research model in Figure 1): environment are against every organization's legitimate interests and should be considered counterproductive [5].
This article is also part of the dominant and important discourse in the literature, and at the same time empirically verifies and significantly furthers the investigations to date by other authors into the impact that quality of interpersonal work relationships has on counterproductive work behavior. Accordingly, the following objectives were set: 1. to determine how the quality of interpersonal work relationships affects the extent of counterproductive work behavior, 2. to determine whether the impact of interpersonal work relationship quality on the extent of counterproductive work behavior is moderated by employees' demographic features (education, age, sex, length of service and type of work).
The goals will be achieved using a survey conducted in March 2018 on a sample of 1488 professionally active people in Poland. To achieve the objectives, two research hypotheses were adopted (see the research model in Figure 1  I expect this study to represent a significant contribution to the relevant literature in two key areas. This study begins with a thorough description of the impact that the quality of interpersonal relationships at work has on counterproductive work behaviors. This study also describes how this impact is moderated by employees' basic demographic features (education, age, sex, length of service Hypothesis 1 (H1): Quality of interpersonal relationships at work has a negative influence on the degree of counterproductive work behavior.

Hypothesis 2 (H2):
The influence that interpersonal workplace relationship quality has on the degree of counterproductive work behavior is moderated by the demographic features of employees, including: (H2a) education, (H2b) age, (H2c) sex, (H2d) length of service and (H2e) type of work.
I expect this study to represent a significant contribution to the relevant literature in two key areas. This study begins with a thorough description of the impact that the quality of interpersonal relationships at work has on counterproductive work behaviors. This study also describes how this impact is moderated by employees' basic demographic features (education, age, sex, length of service and type of work). In the subsequent sections, I first propose a theoretical framework for this study. Next, I propose a suitable methodology to test my research model. I also set out the empirical results of this study and discuss the contributions and implications of those findings. Lastly, I discuss the limitations and future study directions.

Quality of Interpersonal Relationships at Work
Every dimension of human activity is grounded in interpersonal relationships [31]. An organization is the largest incubator of such relationships because its members are "doomed" to frequent interactions and mutual closeness [32]. The relationship includes two complementary components: the task-related and the interpersonal (see also [33]). The first of these predominates and aims at the proper performance of tasks and involves the exchange of task-related resources [33]. Research into relationships at work thus focuses more on the task-related component and its effects on the organization and the organization's results [32]. The interpersonal component applies to personal relationships between employees, and for these to occur mutual acquaintance is required-the deeper that is, the greater the share of the interpersonal component, and thus the parties become closer and begin to see each other as partners or even friends [32].
The concept of interpersonal relationships at work is not clearly understood. Gabarro [34] (p. 81) defined them as an "interpersonal relationship that is task-based, nontrivial, and of continuing duration". It is "a series of interactions between two people, involving interchanges over an extended period of time" [35] (p. 37). They can also be described as a "sequence of interactions between two people that involves some degree of mutuality, in that the behavior of one member takes some account of the behavior of the other" [36] (p. 9). The interpersonal dimension of this relationship means that it is a mixture of the following forms of exchange: verbal (e.g., conversation), para-verbal (e.g., exclamation of surprise), non-verbal (e.g., a smile, proximity during a conversation) and physical (e.g., touching, kissing) [37].
Work relationships have a dual character, i.e., they can be positive (identified with high-quality relationships) or negative (low-quality relationships) [38][39][40][41]. In the first case, their benefits (in terms of vitality, emotions, etc.) are felt on both sides [33,41,42]. By contrast, in negative relationships, at least one party experiences adverse effects (e.g., stress, discomfort, worsened mood) [7]. High-quality relationships are, among other things, more personal, intimate, vital, frequently interactive, and abundant in a variety of (mainly positive) emotions; they are more based on open communication, reciprocity, trust, respect and cooperation, and bring positive energy, and the parties help each other [3,11,16,19,21,[43][44][45]. Meanwhile, negative relationships are more short-lived, impersonal, limited to task-related matters, based on suspicion, formalized, devoid of emotion (and possibly abundant in negative emotions) [3, 11,16].
The quality of employee relationships reflects a team's well-being [46], but many factors complicate the examination of the quality of these relationships. First, it is a continuum of sorts, i.e., quality is graded, and rarely extreme in its form, while relationships can also be seen as neutral or indifferent [47]. Moreover, in practice, it seems easier to qualify relationships as high quality than to consider them to be low quality [48]. It should also be borne in mind that the presence of negative aspects (or the lack of positive aspects) does not necessarily mean that the relationship is wholly negative (and vice versa).
It is thus a dynamic construct [15], which means that [37,49]: • the same relationships can be positive at some times and negative at others, • they may include both positive and negative aspects (interactions), • their intensity may change.
Furthermore, this category is highly subjective, i.e., it depends on a person's individual perception as to which aspects of the relationship prevail and on their assessment of whether their expectations of the relationship have been met [50]. The same subjective perceptions apply to the costs incurred by the parties to the relationship, and of the benefits they obtain from their involvement in the relationship [7,50]. This is because the goal of the relationship is not quality itself, but the creation of values that will meet the needs of the partners and of the organization itself [51]. An additional difficulty in researching the quality of relationships at work is caused by the multi-dimensional nature of the variable, which contains various aspects of exchange within the interaction, and which is further influenced by many conditions of individual and contextual significance.
All this also means that the concept of interpersonal relationships at work is not clearly understood. Table 1 presents selected perspectives on this question. Table 1. Selected approaches to quality of interpersonal relationships at work.

Source
Concept [52] Relationship quality is identified with its strength and means the existence of links between the parties that lead to satisfaction and commitment. In turn, the strength of a relationship is "a (probably linear) combination of the amount of time, the emotional intensity, the intimacy (mutual confiding), and the reciprocal services which characterize the tie" [53] (p. 1361).
[ 54] Relationship quality is the level of mutual respect, trust and sense of duty between employees.
[55] (p. 265) "Quality of relationship entails a pervasive, intentional, and constructive focus on mutual support and on members as individuals." [50] Relationship quality is an evaluation of how well a relationship meets the parties' expectations, needs, predictions, goals and aspirations.
[56] Relationship quality is an evaluation of how far a relationship is based on the principles of reciprocity.
[36] Relationship quality is an evaluation of what actions the parties take towards each other, their feelings and attitudes, and the results that the relationship brings them. [6] High-quality relationships are "relationships built on the interpersonal closeness of employees, as expressed in mutual interest, kindness and willingness to cooperate, contributing to creating a positive organizational climate conducive to effective communication, trust, loyalty and commitment to work".
Based on the above definitions, it can be stated that the quality of interpersonal relations at work is "each party's subjective evaluation (feelings) regarding the degree to which these relationships meet expectations in terms of their results. This quality is the result of many personal and contextual conditions" [57] (p. 141). The determinants of this quality constitute an extremely important issue, though the literature does not agree on which are relevant, or to what extent [45,50]. Such determinants certainly exist, because the process of relationship development differs in pace and degree in each case [34].
The literature usually indicates the following determinants of quality of relationships at work: satisfaction, the mutual dependence of employees, commitment, trust, the traits and similarity of parties, duration of relationship, frequency of interaction, emotions, investment in relationships, communication, organizational culture and atmosphere, and relationships outside of work [57][58][59][60][61][62]. Importantly, none of them independently explains the quality of employee relationships [63].
Based on complex qualitative and quantitative research, Szostek [57] operationalized the determinants of interpersonal relationships at work, dividing them into four categories and using them to create a validated instrument for measuring this quality, which has been used in the research conducted for the needs of this article. These categories are: organizational climate (e.g., atmosphere at work, honesty, trust, how parties treat one another), 2.
interpersonal ties (e.g., sharing personal information, contact after work, helping each other, celebrating important occasions together), 3.
interpersonal relationship building methods (e.g., caring for how the workplace is equipped, meetings with employees, surveying their opinions, the holding of company events), 4. distance resulting from management style (e.g., fair treatment by the supervisor, the "human approach" of the boss, private contact after work).
Furthermore, the author has divided the manifestations of the quality of these relations into two dimensions, i.e., the causes of the quality versus its effects (some of the manifestations play a dual role, e.g., trust between employees), and organizational manifestations (initiated by the organization) versus individual (initiated by the employee).

Counterproductive Work Behavior
Counterproductive behavior is also described as negative, erroneous, pathological, deviant, dysfunctional or unethical, although these concepts are not synonymous and do not express the essence of such behavior. The fundamental difficulty in defining these behaviors derives from the fact that they manifest in multiple different ways, some of which are very serious and others trivial [64][65][66]. As a result, different authors approach CWB differently, depending on which types of behavior they believe are predominant (see Table 2). "Set of distinct acts that share the characteristics that they are volitional (as opposed to accidental or mandated) and harm or intend to harm organizations and/or organization stakeholders, such as clients, coworkers, customers, and supervisors." [70] (p. 418-419) "Set of negative behaviors that are destructive to the organization by disturbing its operational activities or assets, or by hurting workers in such a way that will overcome their efficiency." [71] (p. 14) "Problem that violates significant organizational norms and threatens the wellbeing of an organization, its members, or both." Source: author's own work, based on [57].
However, leaving conceptual differences aside, behavior can be considered counterproductive if the following three conditions are met [72]: it results in a violation of the standards in force in the organization, 2.
the conduct was engaged in voluntarily, and 3.
it harms (including potentially) the organization and/or its stakeholders.
It is impossible to list all the potential examples of CWB, but some authors have decided to organize them, and have proposed various classifications. The best known are listed in Table 3.
Counterproductive behaviors can also be directed against sustainable development (e.g., environmental pollution, wasting resources, using environmentally unfriendly products), and there is thus increasing discussion of "counterproductive sustainability behaviors" (CSB) [5]. Furthermore, taking into account the negative correlation between CWB and OCB [12], it can be assumed that an increased propensity for counterproductive work behavior reduces employees' sustainable behaviors, which are often a form of organizational citizenship behaviors [22]. Meanwhile, unless employees engage authentically in such behavior, there can be no question of building a culture of sustainability [77,78].
Currently, the most frequently cited classification is that proposed by Spector et al. [69]. The authors distinguish two dimensions of these behaviors, i.e., CWB-I (individual-oriented) and CWB-O (organization-oriented). They also propose five categories of behaviors, namely: 1.
abuse against others-behavior harmful to other stakeholders of the organization (e.g., lying, gossiping, harassment), 2.
production deviance-performance of duties by the employee such that the work cannot be properly completed (in terms of quantity and/or quality of results), 3.
sabotage-deliberate destruction of the organization's property (not only tangible but also intangible, e.g., its image), 4.
theft-intentional appropriation of property belonging to the organization or other people, 5.
withdrawal-limiting working time to below the minimum required to properly achieve the organization's goals.
It should be noted that the causes of counterproductive behavior have not been fully identified [72], and studies instead relate primarily to how such behavior manifests. The situation is complicated by the fact that many of the analyzed causes of CWB are also identified as results of counterproductive actions.

Sampling Procedures and Participant Characteristics
The survey was conducted from March to May 2018 using a triangulation of measurement methods, i.e., an online survey (approximately 90% of collected data), a face-to-face survey and an auditory survey (approximately 10% of data collected from 5 local enterprises). The measurement involved professionally active people in Poland, who were selected non-randomly (by a combination of deliberate selection and selection by the "snowball" method). An invitation for their employees to complete the questionnaire was sent to the institutional email addresses of: It was also sent to employed persons that the author knows personally, including via Facebook profile (about 300 people in total). Additionally, in the case of the online survey, the cover letter asked respondents to send the questionnaire to professionally active people they knew.
Respondents' anonymity was guaranteed despite the fact that a large part of the questionnaires was sent to the potential respondents' employers (municipal offices and enterprises). Respondents received the author's original message with a link to an external questionnaire and were informed that only the author had access to the results of the study. In addition, participation in the survey was voluntary, and completion of demographic questions was also optional. Respondents had a high sense of anonymity. This is indicated by the fact that most of them provided their demographic data. Besides, responses to sensitive questions (mostly CWB-C) had a relatively high variance.
The characteristics of respondents in terms of main demographic variables are presented in Table 4.

Measurement Scales
The extent of counterproductive work behavior was measured using the CWB-C [85] scale (Counterproductive Work Behavior Checklist) by Spector et al. [69]. The scale is used to measure multiple manifestations of such behaviors simultaneously (the most extensive version contains 45), dividing them into the aforementioned five categories (abuse against others, production deviance, sabotage, theft, withdrawal) and two dimensions (individual-or organization-oriented). The respondents assessed how often they had engaged in manifestations of CWB, indicating one of the options: never, one or two times, one or two times per month, one or two times per week, every day-see see Appendix A (Table A1).
Meanwhile, the quality of interpersonal relationships at work was measured using QIRT-S (Quality of Interpersonal Relationships in the Team Scale) [57]-see Appendix A (Table A2). This scale contains 58 wordings, and the respondent is asked to respond to them by indicating 1 of the options: I strongly disagree, I somewhat disagree, difficult to say, I somewhat agree, I strongly agree. The statements can be divided into the aforementioned four categories (organizational climate, interpersonal ties, interpersonal relationship building methods, distance resulting from management

Reliability Values
A total of 1488 correctly completed questionnaires were received, and these were analyzed using IBM SPSS Statistics and IBM SPSS Amos v. 16 (research data file with all observations is available in Supplementary Materials). Then, variables with very low variance were eliminated from the analysis, i.e., those for which the share of "never" replies (CWB-C) was at least 95% (see Table 5). Confirmatory factor analysis (without rotation) was carried out in the SPSS program using unweighted least squares method for extraction. The results are presented in Table 6. In the next stage, confirmatory factor analysis was carried out, which made it possible to select those variables that had the highest factor loadings and most significantly influenced the categories of relationship quality and dimensions of counterproductive behavior. Table 7 lists the variables used in the further analysis in structural equation models (variables are marked "Q" for quality of relationships and "C" for counterproductive behaviors, alongside a number on a given scale). The table also contains Cronbach's Alpha coefficients as a measure of reliability. Further investigation began with structural modeling of the impact that the category of quality of interpersonal relationships at work had on the dimensions of counterproductive work behavior. The model was evaluated using the maximum likelihood method, adopting a significance factor of 0.05.
It was decided not to model using the quality dimensions of interpersonal relations because they are strongly correlated with one another (partly because they are created to some extent using the same variables). Presumably, the proposed dimensions could largely be measuring the same thing, which undermines the sense in distinguishing them in the analysis.

(H1): Quality of Interpersonal Relationships at Work Has a Negative Influence on the Degree of Counterproductive Work Behavior
The hypothetical structural equation modeling (SEM) structural model is presented in Figure 2 and illustrates the assumed structural relationships between the categories of relationship quality and the dimensions of counterproductive behavior. Importantly, scientific reflection on the results of the author's research and that of other authors led to the further assumption that a relationship existed between categories of this quality. The illustrated model does not take into account the measurable variables making up individual factors, but it is consistent with those listed in Table 7.   The external SEM model was evaluated using the maximum likelihood method [86,87]. The results of the estimation of the external SEM model (factor analysis) are presented in Table 8, while the internal model (regression analysis) is shown in Table 9. Measures of the degree to which the model fits the data are shown in Table 10, while Table 11 lists standardized total effects of the impact that individual quality categories relations have on CWB-I and CWB-O. The external SEM model was evaluated using the maximum likelihood method [86,87]. The results of the estimation of the external SEM model (factor analysis) are presented in Table 8, while the internal model (regression analysis) is shown in Table 9. Measures of the degree to which the model fits the data are shown in Table 10, while Table 11   The external SEM model was evaluated using the maximum likelihood method [86,87]. The results of the estimation of the external SEM model (factor analysis) are presented in Table 8, while the internal model (regression analysis) is shown in Table 9. Measures of the degree to which the model fits the data are shown in Table 10, while Table 11 lists standardized total effects of the impact that individual quality categories relations have on CWB-I and CWB-O.    In factor analysis, all variable loadings are statistically significant. For some variables, the P value could not be calculated, because they needed to be assigned constant variance in order to ensure model identifiability [88].
For most categories of relationship quality, the impact on both CWB dimensions is negative and this relationship is statistically significant (parameters β 7 , β 10 , β 11 , β 12 , β 13 ; see Table 9). This means that improvement (or deterioration) in these aspects of relationship quality leads to a decrease (or increase) in the degree of such behavior. As already mentioned, and as other authors have already confirmed (e.g., [18][19][20]) this relationship between these constructs is logical.
For the categories "interpersonal relationship building methods" and "interpersonal ties" only, despite the fact that they negatively affect CWB-I (β 10 , β 13 ), they also noticeably positively influence CWB-O (β 6 , β 9 ). This is one of the few observed paradoxes for which some logical justification can be found.
The category "interpersonal relationship building methods" includes various organizational activities (e.g., caring for how the workplace is equipped, organizing meetings with employees) and its positive influence on CWB-O (β 6 ) may be explained by employee cynicism. This is a situation in which employees begin to exploit the activities that the organization conducts to build work relationships for their own interests (e.g., the organization promotes teamwork, limiting the ability to control the activities of a single employee, which may increase social loafing). This relationship stands as a warning to the organization-its activities for creating relationships between employees must be prudent and must not lead to personnel being reoriented from organizational interests to individual interests.
The category "interpersonal ties" is in a way the essence of workplace relationship quality (it includes, for example, private conversations, displays of emotion, helping out). When ties are strong, relationships between employees often move on to a non-professional footing. This may explain this category's positive relationship with CWB-O (β 9 )-e.g., according to the team, the organization's activities are directed against an employee (e.g., unfair treatment, dismissal) and then counterproductive behavior may be a form of opposition by the rest of the team or a display of revenge against the organization for such an act. This means that, paradoxically, an organization may be disadvantaged when interpersonal relations at work are too good (employees become reoriented away from the organization and towards the team).
The negative impact of the "interpersonal relationship building methods" category on CWB-I (β 10 ) can be explained by the fact that the organization has certain tools that it can use to reduce such behaviors (e.g., employing the right workers, promoting dialogue between employees). If the organization's activities in this respect are inappropriate or lacking, the tendency of personnel towards CWB-I may increase. The "interpersonal ties" category is extremely interpersonal, and so its negative influence on CWB-I (β 13 ) is more understandable, and requires no further comment-quite simply, the stronger the bonds between employees, the less they tend to behave counter-productively towards each other.
The negative impact of the "distance resulting from management style" category on CWB-I (β 12 ) also seems logical. The variables comprising this category primarily express how management treats employees (e.g., clear division of responsibilities, ease of communication). Therefore, the higher the quality of the relationship in this aspect, the lower personnel's tendency to direct CWB at a supervisor. In the opposite case (e.g., unfair or disrespectful treatment of subordinates), counterproductive behavior would be a means of retaliating against a supervisor.
The "organizational climate" category influences both dimensions of CWB negatively (β 7 , β 11 ) and in both cases this relationship is statistically significant. Despite the fact that this category consists of both organizational and individual variables (e.g., honesty, trust, atmosphere at work), responsibility for the climate lies mainly with the organization. Therefore, the higher the quality of relationships in this domain, the lower personnel's tendency towards CWB-O. In the opposite case, such behavior may be a manifestation of revenge against the organization for a bad workplace climate. Particular categories of relationship quality also interact with one other. The categories "interpersonal relationship building methods" and "distance resulting from management style" significantly positively influence the category "organizational climate" (β 7 , β 11 ). This is understandable because, as previously mentioned, workplace climate is most heavily influenced by the organization, whereas both categories of relationship quality emanate from the organization's activities in this regard.
The category "interpersonal ties" is positively influenced by the categories "distance resulting from management style" and "organizational climate" (β 3 , β 5 ). In the first case (β 3 ) a supervisor's approach to subordinates affects not only vertical interpersonal relationships, but also horizontal ones (e.g., unfair treatment of subordinates leads to mutual jealousy between them, while a clear division of labor reduces task-related conflicts within a team). The influence of organizational climate on interpersonal ties at work (β 5 ) is equally understandable-when it is based on trust, cooperation, solidarity and discretion, ties between employees are strengthened. In the opposite case they atrophy.
The relationship between the categories "interpersonal relationship building methods" and "interpersonal ties" is not statistically significant (β 4 ).
The interaction that the model shows between categories of relationship quality made it possible to determine the indirect and total influence of these categories on the analyzed dimensions of CWB. For example, the category "distance resulting from management style" affects CWB-O less directly and more indirectly-through two other categories: "organizational climate" and "interpersonal ties".
Considering the standardized values of total effects (see Table 11), CWB-O is most strongly influenced by the category "interpersonal ties" (0.575), and least by "distance resulting from management style" (0.095). This last category also has the strongest influence on CWB-I (−0.576).
Regarding the estimation of the model's degree of fit to the empirical data (see Table 10), it should be emphasized that the Incremental Fit Index (IFI) value is 0.827, while Root Mean Square Error of Approximation (RMSEA) is 0.054, which leads to the conclusion that the model's fit is correct and satisfactory. The CMIN/DFstatistic deviates slightly from the norm (it is above 2), but in the case of SEM models, each measure of their quality has some limitations, and the choice between them is usually subjective [88].
IFI (Incremental Fit Index) is one of many measures of the relative fit of a model and is calculated based on the comparison of chi-square statistics and the degrees of freedom of the estimated and base model. In this case, the base model is understood to be an independent model in which the analyzed variables are not intercorrelated at all. The IFI returns values in the range < 0; 1 > and the higher it is, the better the model's fit to the data [89].
RMSEA (Root Mean Square Error of Approximation) is the discrepancy between the theoretical and population matrices of variance-covariance, corrected by the number of degrees of freedom. This is one of the few measures for which fairly widely accepted thresholds obtain, i.e.,
CMIN (chi-square statistic) is a model fit test statistic. It can be used to calculate the "p" significance of a model's misfit to the data. When examining large samples, it is difficult not to reject the null hypothesis, i.e., that the model fits the data (p > 0.05). In such situations, the CMIN/DF ratio (normed chi-square statistic) is considered, where DF is the number of different elements of the variance-covariance matrix minus the number of estimated parameters. Researchers recommend that models in which CMIN/DF exceeds 2 be rejected, although other authors adopt less strict limits (of 5 or even 10) [89].

(H2): The Impact of Interpersonal Work Relationship Quality on the Extent of Counterproductive Work Behavior is Moderated by Employees' Demographic Features (Education, Age, Sex, Length of Service and Type of Work)
In order to verify the hypothesis about the impact of the quality of interpersonal relationships at work on CWB dimensions being moderated by the demographic features of employees (education, age, sex, length of service and type of work), estimated models were analyzed in subgroups that were distinguished in terms of these features.
For testing the differences between corresponding coefficients in the models, a T-test was used. Parameters for which the statistic value is greater than |1.96| are statistically significantly different between analyzed groups [90].
Firstly, the examined persons were divided into two subgroups by level of education. More than half had a higher education, so to make groups of the same size, only two subgroups were distinguished: I-people with a higher education, II-people with a middle school, vocational or secondary education. The results of the internal model estimation are shown in Table 12.
For people without a higher education (group II), the influence that the categories "interpersonal relationship building methods", "organizational climate" and "interpersonal ties" have on CWB-I proved to be statistically insignificant (β 10 , β 11 , β 13 ). In addition, for this group, almost all categories of workplace relationship quality had a less influence than they did for people with a higher education, i.e., those from group I (a higher P value). These differences for parameters β 9 and β 11 were statistically significant (T-test value greater than |1.196|). Of course, this does not mean that people from group II are less inclined towards CWB, but it can be stated that relationship quality is not as important a factor in shaping these behaviors as it is for group I people (e.g., this may derive from the type of work performed, which for people with a higher education may be a profession with more frequent interactions with others).
The respondents were also divided into two groups based on age, taking 35 as the cut-off boundary (which made the groups more or less equal in size, but is also often used in practice as a boundary to divide people into the young and the mature). The results of the internal model estimation are shown in Table 13. No significant differences were found in the influence of relationship quality on CWB in either group, with two exceptions: in the group of people aged over 35 the category "organizational climate" was found to influence CWB-I (β 11 ), while among those aged 35 or less the category "interpersonal ties" was found to influence CWB-I. Furthermore, the influence of the category "organizational climate" and "interpersonal ties" on CWB-O (β 7 , β 9 ) and "distance resulting from management style" on CWB-I (β 12 ) was stronger in group II.  When the population was divided into sex subgroups, for men the categories "interpersonal relationship building methods", "organizational climate" and "interpersonal ties" had no statistically significant influence on CWB-I (β 10 , β 11 , β 13 ). A similar situation was noted for the impact of the category "distance resulting from management style" on CWB-O (β 8 ). However, it can be seen that in almost all the considered combinations the impact of relationship quality on CWB-I is of higher significance for women (see P value and T-test). The results of the model estimation are shown in Table 14.
Respondents were then divided into groups by length of service in their given position, with the dividing cut-off set as 8 years (the median). The results of the model estimation are shown in Table 15. For people of shorter service periods, the category "interpersonal relationship building methods" does not significantly influence CWB-I (β 10 ); the influence of the category "organizational climate" on CWB-I was also found to be insignificant (β 11 ). Meanwhile, for employees of longer service, the category "distance resulting from management style" does not significantly influence CWB-O (β 8 ).
The final part of the investigation involved dividing respondents into subgroups by type of work. The results of the model estimation are shown in Table 16, and of T-test in Table 17. For blue-collar employees, two categories of relationship quality ("organizational climate" and "distance resulting from management style") had a statistically non-significant influence on CWB-I. The second of these categories also negatively affected CWB-O for this group of employees. Regarding employees in managerial positions, the categories "interpersonal relationship building methods", "organizational climate" and "distance resulting from management style" turned out to have an insignificant influence on CWB-I, as did the category "interpersonal ties" on CWB-O.

Discussion and Conclusions
The article discusses the impact that quality of interpersonal relations at work (the categories of this quality are: organizational climate, interpersonal ties, interpersonal relationship building methods, distance resulting from management style) has on counterproductive work behavior (the dimensions of these behaviors are: those directed against other people and those directed against the organization itself). It was also investigated how this impact is moderated by employees' basic demographic features, namely: education, age, sex, length of service and type of work. Based on the analyses carried out using SEM structural modeling, it must be concluded that there are no grounds for rejecting either research hypothesis, i.e.: Hypothesis 1 (H1): Quality of interpersonal relationships at work has a negative influence on the degree of counterproductive work behavior.

Hypothesis 2 (H2):
The impact of interpersonal work relationship quality on the extent of counterproductive work behavior is moderated by employees' demographic features (education, age, sex, length of service and type of work).
The relationship between these constructs is, however, far more complex, as evidenced by the paradoxes that are discussed, and primarily due to cases in which certain categories of relationship quality have a positive effect on CWB-O or CWB-I.
The subject of the article is closely related to the area of sustainability. Counterproductive behaviors often violate the principles of sustainability in each of three areas: environmental (e.g., increased waste material), economic (e.g., theft of raw materials) and social (e.g., deterioration of working conditions, unequal treatment). These behaviors are often a consequence of inappropriate interpersonal relationships at work (e.g., they are a form of revenge on an organization or other people in the organization). Furthermore, due to the negative relationship between CWB and OCB, an increased propensity for counterproductive behavior limits employees' sustainable behaviors, which are a manifestation of organizational citizenship behaviors. So, the matters addressed in the article are particularly important for organizations that want to operate in a sustainable manner. Every organization that wants to contribute to sustainable development should monitor the quality of interpersonal relationships at work to minimize employees' propensity to engage in CWB and CSB.

Theoretical Contributions and Managerial Implications
The considerations made in the article contribute greatly to developing management science theories, in particular in terms of the impact that quality of interpersonal relationships at work has on counterproductive work behavior. Until now, this issue had principally been understood intuitively or based on unverified paradigms. The author has empirically confirmed that the said influence exists-and furthermore, that it is moderated by such personnel demographic features as education, age, sex, length of service and type of work. Almost all categories of workplace relationship quality had a less statistically significant influence for employees without higher education than they did for people with a higher education. Only a few significant differences in the influence of relationship quality on CWB were found between age groups. Almost all considered categories of relationship quality had a more significant impact on CWB for women than for men. Furthermore, for employees of longer service (compared to those of shorter service) and those with managerial position (compared to "blue collar" workers), more categories of relationship quality had a significant influence on CWB.
The article also confirms the complexity of the relationship between relationship quality and CWB. Looking at this issue in brief it can be stated that the organization should take care that relationships between employees are as good as they should be (the results of the study prove that this does not mean that the quality of relationships should be as high as possible; it should be of such a level that does not reorient the employee from the interests of the organization to the interests of the team).
This will reduce their engagement in various types of counterproductive behavior (mainly those aimed at colleagues, but also in counterproductive sustainability behaviors). Furthermore, employees will be more likely engage in various types of OCB, including sustainability behaviors. Not only does the whole organization benefit from this, but so too does its environment, including the natural environment and society. Thus, each organization should use various available tools that can influence the quality of relationships at work (e.g., company events, employee meetings, promoting teamwork). Conversely, if these relationships are of low quality, employees will more readily engage in various forms of CWB.
However, a more thorough analysis of what impact the quality of interpersonal relationships at work has on counterproductive work behavior reveals many paradoxes that should sensitize managers to the issue of managing these relationships. The key guidelines that need to be considered when managing the quality of interpersonal relationships at work include: 1.
Building high-quality interpersonal relationships at work should be moderated so as not to reorient the employee away from the organization and towards him/herself or the team. Such a reorientation may result in a greater degree of CWB-O (e.g., in the name of solidarity with the team) or of CWB-I (e.g., abuse of colleagues' trust).

2.
The organization must not forget that retaliation is one of the main reasons for employees to engage in counterproductive behavior against the employer. Hence, imprudent actions by managers (e.g., non-payment for overtime, undervaluing subordinates) may intensify such staff behavior. 3.
The organization's prevailing climate (which should be based on honesty, solidarity, altruism, etc.) is very important for the quality of relationships at work, and thus for employees' propensity for CWB.

4.
Of equal importance for the relationship between quality of relationships and CWB are employees' demographic features. (The relationship is strongest for employees with higher education, those in senior positions, those with longer service, and women. Apart from these, only a few significant differences in the influence of relationship quality on CWB were found between age groups.) 5.
When recruiting new people, attention should be paid not only to candidates' knowledge, experience and qualifications, but also to their propensity for CWB, as well as to the impact of a given person on the quality of relationships in the team (in this case, integrity tests or information from former employers or from Facebook can be used). 6.
The process of socializing employees within the organization should be thought out and balanced in terms of orientating the person towards his/her own interest, and that of the team and the workplace. Open communication and consistency in action play an important role here. 7.
Employees should be trained in the competences that play a key role in the quality of interpersonal relationships at work and in counterproductive work behavior. These might include not only traditional training, but also atypical activities (e.g., strategy games, going out to play paint ball together). 8.
The quality of relationships between employees and the degree of CWB should constantly be monitored so as to respond sufficiently early to any disturbing situations. The available measuring scales can be used for this (e.g., those discussed in the article-the CWB-C and the QIRT-S).
It is necessary to ensure respondent anonymity, which will increase data reliability. Naturally, after completing a survey, employees should be informed of their results, and the necessary actions should be taken to mold the quality of relationships between employees and to prevent CWB. Table 18 presents the main shortcomings of the study, as well as actions taken to mitigate the effects of identified restrictions. This limitation is mitigated by the fact that the sample in the study was relatively large in number and demographically diverse, including in terms of education, age, sex, length of service and type of work, but also employment sector (private vs. public).

Limitations and Future Study Directions
2 Partial application of face-to-face survey methods in CWB measurement Face-to-face surveys reduce employees' sense of anonymity. Therefore, only 10% of the total data was collected in this way, and the rest using an online survey (guaranteeing anonymity).
Furthermore, about 20% of data collected by face-to-face survey had zero variance (practically the only answer was "never"). It was therefore right to use a triangulation of measurement methods, including indirect methods. The analysis excluded those questionnaires for which variance of CWB was 0. 3 The quantitative nature of scales for measuring quality of interpersonal relationships at work and counterproductive behaviors Quantitative research has certain limitations, particularly for such complex and dynamic issues as the quality of relationships and CWB. Nevertheless, validated scales were used in the measurement, making the collected data highly reliable, as measured by Cronbach's alpha coefficient.
Source: author's own work.
In considering the above limitations of the author's research, it should be stated that the presented results are exploratory in nature. It is therefore important for subsequent research in this area to measure the impact that quality of relationships at work has on CWB using a representative sample of employees. It would seem important to differentiate the organizations surveyed by industry, size, ownership type, degree of internationalization, etc. It would also appear of interest to analyze the non-governmental sector separately, given that its operations have a certain particularity.
In future studies on the impact that quality of interpersonal relations at work has on CWB, an augmented research model should be proposed that would take into account the impact not only of employees' demographic features, but also of the determinants of relationship quality mentioned in the first part of the article (and these should first be operationalized). These are: satisfaction, the mutual dependence of employees, commitment, trust, the traits and similarity of parties, duration of relationship, frequency of interaction, emotions, investment in relationships, communication, organizational culture and atmosphere, and relationships outside of work.
It should also be remembered that the use of a measuring scale created in specific cultural conditions (e.g., CWB-C, QIRT-S) requires adaptation to the conditions of any study, and separate validation. This is indicated by, for example, the fact that, in the structural model of the impact that category of quality of interpersonal relationships at work has on particular CWB categories as presented by Szostek [57], as many as three of the categories had to be eliminated (sabotage, work avoidance and disruption of work). This was due to the low values of Cronbach's alpha for these categories (below 0.5). This may mean that the variables that comprise these categories are not measuring the same phenomena.   Relationship quality categories and their related statements: (1) organizational climate: 8,12,19,24-27,29-38,44,45,47,50-55,57,58; (2): interpersonal ties: 1-7,9-11,13-16,48,49,56; (3) interpersonal relationship building methods: 39-43; (4) distance resulting from management style: 17,18,20-23,28,46. Source: [57] (p. 244-247).