Does the Occasion Justify the Denunciation ? : a Multilevel Approach for Brazilian Accountants

Frauds represent large losses to the global economy, and one of the main means for their containment is by means of denunciations within organizations: whistle blowing. This research aims to analyze whistle blowing within the Brazilian context, considering the influence of costs and intrinsic benefits as well as aspects of the individual's interaction with his/her organization, profession and society at large. By means of a questionnaire answered by 124 accountants, a multilevel model was applied to analyze these aspects. The results demonstrate the importance of situational aspects as a positive influence in favor of denunciations. These results are useful for organizations and regulatory institutions in developing institutional mechanisms to encourage denunciation. Moreover, the results are also useful for teachers of professional ethics and members of the Federal and Regional Accounting Councils, which are dedicated to the assessment of alleged deviations from the professional code of ethics.


INTRODUCTION
he act by which members or former members of a certain organization denounce illegal, immoral or illegitimate practices conducted by another person to people or competent entities is known as whistle blowing (NEAR; MICELI, 1985).Thus, this mechanism is not only an alternative route of "voice" for employees in their workplace (BRIAN et al., 2012), but also an efficient way for an organization to keep exert internal control in order to prevent and detect unethical conduct (TAYLOR; CURTIS, 2010).
Interesting examples of the importance of whistle blowing in the corporate world date back to the cases of Enron and WorldCom, which occurred in 2002 and led to the publication of the Sarbanes-Oxley Law (SOX), which in turn required the implementation of a formal denouncing channel for all business companies covered by the scope of the law, aimed at strengthening corporate governance mechanisms.
It is noteworthy that despite research studies on whistle blowing not being recent (major studies date back to the 1980s), the international literature in the area focuses almost exclusively on the analysis of personal aspects such as the individual's commitment to the profession (KAPLAN; WHITECOTTON, 2001;TAYLOR;CURTIS, 2010) and the company (BECKER, 1992;BECKER;BILINGS, 1993;TAYLOR;CURTIS, 2010), as well as his/her position on the aspects regarding society and the public interest (LUKÁCS et al., 2012).
These aspects have as their main characteristic the fact that they are kept constant by the individuals.
However, specific characteristics of the offense may also exert an important influence on the denunciation (MESMER-MAGNUS; VISWESVARAN, 2005).Robinson, Robertson & Curtis (2012) noted, however, that to date, little is known about the effects of contextual attributes on the decision to denounce.
Another common practice in research on whistle blowing is focusing on the Anglo-Saxon context.Bashir et al. (2011) noted that certain elements of national culture can also interfere.Thus, it is clear that much remains to be explored in other national contexts.In Brazil, for example, when the peculiarities of the national culture are taken into account (e.g. "jeitinho", cordiality and personalism), it has been suggested that denunciation may in fact not be the natural course in organizations.From this perspective, this study aims to analyze the impact of situational aspects in the practice of whistle blowing in the Brazilian context.Therefore, a multilevel approach that seeks to analyze situational intrinsic incentives that would drive the individual to decide in favor of denunciation is employed, as well as to identify how the relationship between the individual and the conjunctural scenario affects his/her decision to denounce.This methodological approach has been increasingly used in the field of Administration (HITT et al., 2007;GOLDZMIDT;BRITO;VASCONCELOS, 2007;FÁVERO, 2008).

THEORETICAL BACKGROUND
Whistle blowing may be seen as a procedural mechanism, in which the first process begins when the individual is faced with a suspicious behavior and evaluates the possibility of such behavior being fraudulent.When this evaluation is positive (in this case, when the individual actually classifies the behavior as reprehensible), the individual is then directed to the second process: the decision to denounce (or not) the observed behavior (NEAR; MICELI, 1985MICELI, , 1996)).
In this section, the theoretical framework used to develop the research will be presented.
In Section 2.1, important results presented in the literature are analyzed, emphasizing how personal aspects are widely recognized as an important factor in the decision making process to denounce.In section 2.2, the main argument of the article is presented by explaining the process through which the situational aspects influence the interest in denouncing by the alleged whistle-blower and, more specifically, the role that the intrinsic costs and benefits can have.

THE INFLUENCE OF INDIVIDUAL ASPECTS IN DENUNCIATION
The work developed in international literature demonstrates that the rules that dictate and regulate certain social relations can significantly affect the choices of individuals (CIALDINI et al., 199;ROBINSON;O'LEARY-KELLY, 1998;TAYLER;BLOOMFIELD, 2011).For whistle blowing, studies show that variables such as the relationship of individuals with their organization, with norms of their profession and of society as a whole might influence the alleged complainant towards the protection of the offender (i.e., not denouncing) or pursuing his/her punishment (HOLLINGER; CLARK, 1983;NEAR;MICELI, 1996;KEENAN, 1990;MESMER-MAGNUS;VISWEVARAN, 2005).different organizational contexts, help to explain the (anti)ethical behavior.In the organizational environment, the individual may be committed to his superior, to his closest co-worker and the very organization itself (BECKER, 1992).Bilings & Becker (1993) identify that the individual commitment can occur in local (co-workers) or global (organization) terms.Thus, the trade-off formation is highlighted, in which co-workers and the organization are arranged as two extremes of a continuum, and the individual will choose between protecting one of the two sides (TAYLOR; CURTIS, 2010).Thus, the denunciation would tend to occur in organizations where the whistle-blower perceives a high congruence between his/her personal values and the values of the organization (NEAR; MICELI, 1985MICELI, , 1996)).
Still regarding the importance of the commitment, Meyer & Allen (1997) (2001) have identified that the professional identity is positively related to the perception of the responsibility to denounce identified unethical acts.Borges & Medeiros (2007), while analyzing the case of Brazilian accountants, have highlighted the relationship between professional commitment and ethical aspects.Cardoso, Mendonça Neto & Oyadomari (2010) show evidence of different professional posture by the Brazilian accountant and all others.
The ethical standards of professional conduct are usually implicitly or tacitly determined (FLYNN; WILTERMUTH, 2010), making it difficult for members of organizations to accurately identify what types of behaviors are well regarded or welcomed in the profession (TREVIÑO, 1986).
The denunciation involves a conflict between the employee's loyalty and the protection of the public interest (LUKACS et al., 2012).The denunciation also plays a social role, since, if the denounced fraud is properly investigated and stopped, the members of the organization, its stakeholders and the entire society tend to benefit from this interruption (MICELI; NEAR; SCHWENK, 1991).Rainey & Steinbauer (1999) and Perry & Hondeghem (2008) consider this altruistic motivation as a public motivation that would influence an individual's decisions toward ethical behavior.This factor would lead individuals to compare their personal values and those of the organization, analyzing whether the exercise of their role provides opportunities to satisfy their willingness to cooperate with society (STEIJN, 2008;WRIGHT;PANDEY, 2008).Thus, more specifically, the decision to denounce could be altruistically Based strictly on the standpoint of making purely rational decisions, the denunciation will take place when the personal benefits exceed the private costs.The individual identifies the need to denounce the identified activity by weighing the costs and benefits of the possible denunciation for the organization, the fraud agent and the complainant him/herself (RAPP, 2007).In this perspective, the incentive structure (costs and benefits) experienced by the alleged complainant may be briefly divided into two categories: extrinsic and intrinsic.
Extrinsic benefits are represented by rewards offered by companies to the complainant.
In certain contexts, for example, the denunciation is seen as a reputable practice, able to bring glory and fame to the complainant (FAUNCE;JEFFERYS, 2007).Not infrequently the complainants also face significant costs when choosing to denounce misconduct.Extrinsic costs are associated with any type of punishment or loss (material or reputational) imposed on the complainant by the company or by other agents (such as co-workers).Although whistleblowers are sometimes protected by "anti-retaliation" laws and/or regulations, in practice however, cases in which complainants bear a number of negative consequences as a result of their actions are constant (LUKACS et al., 2012).The complainants are frowned upon within organizations, possibly facing retaliation from all parties involved, either directly from their co-workers or the company's high management (PARMERLEE;NEAR;JENSEN, 1982;JOS;TOMPKINS;HAYS, 1989).It should be noted that research on whistle blowing has had its focus primarily on the structure of extrinsic incentives, showing how financial incentives and the possibility of retaliation would affect the attitude of the individual.However, such studies have shown unclear and/or inconclusive results (GREENBERG, 2011).
Intrinsic costs are related to self-imposed moral negative effects by the complainant for having made the denunciation, such as the discomfort and constraint for having denounced an offense committed by a close co-worker.On the other hand, the intrinsic benefits are associated with the well being of having interrupted an immoral or illegal activity that the complainant deems serious or feels responsible to denounce, considering the effectiveness of the denunciation to stop the fraudulent conduct (NEAR; MICELI, 1985MICELI, , 1996)).At this point, Benabou & Tirole (2006) identified that, when analyzing their role of utility, individuals take into account the image that their decision will convey about themselves (to themselves and others).Moreover, Sims &Keenan (1998) andO'Fallon &Butterfield (2005) found few studies in literature that analyze the flaws that individuals commit upon making a moral judgment when evaluating certain situations.
Considering these intrinsic benefits and costs, two more hypotheses are developed to be tested in this study: H2. Occasions remitting individuals to a higher moral intensity lead to greater determination in denouncing fraudulent acts.

H3. Occasions remitting individuals to a greater constraint lead to lesser determination
in denouncing fraudulent acts.

DATA COLLECTION AND QUESTIONNAIRE
To obtain the variables used in the study, a questionnaire consisting of three parts (as shown in the Appendix) was applied.The first part is related to the characteristics of the respondents, aiming at understanding the profile of the analyzed individuals.The second part consists of statements to be judged by the respondents according to their degree of agreement.
In the third part, four scenarios were developed to be evaluated by answering five questions.
The scenarios were developed emphasizing possible interactions in the workplace that could drive individuals to commit fraud, changing both the offender and the motivations that would drive him/her to commit the offense in each scenario.This procedure was adopted to enable greater variability in the overall assessment of the scenarios performed by the respondents.Scenario A shows a fraud committed by a co-worker in response to a request from a high position in the organization.These facts would highlight the relationship between the individual and his/her closer co-workers as well as the constraints associated to the relationship with the superiors.Scenario B shows a fraud committed by a co-worker who decides to practice the wrongful act at his/her own will, to personally benefit at the expense of the organization, other co-workers and society.Scenario D resembles the characteristics of scenario B, but it highlights a hierarchical superior as the offender.Scenario C describes a non fraudulent and easy to analyze situation; this was included in order to ensure that respondents were reading and correctly interpreting the proposed scenarios; therefore, respondents who considered scenario C as fraudulent had all their answers discarded.

PARTICIPANTS
A pilot test with 20 business administration post-graduate students was conducted in order to identify possible flaws in understanding and to establish the average time required to complete the entire questionnaire.After adjustments, the questionnaire was applied to 162 accountants in micro and small Brazilian private enterprises, in hard copy, during professional development seminars organized by the Federal Accounting Council (CFC).After excluding questionnaires in which the respondents considered Scenario C as fraudulent, 124 questionnaires were in fact analyzed.
The choice of such a group was motivated by the fact that accountants have access to varied and valuable information, which is essential to the organization and enables contact with fraudulent situations in the organizational context.Moreover, since the profession is regulated, in case the accounting professional does not denounce a fraudulent act that he/she is aware of, he/she shall be treated as an accomplice of the offender, liable to punishment (ALVES et al., 2007).

VARIABLES USED
The dependent variables used in this research study are the "will" and "perseverance" to denounce, as discussed by Taylor & Curtis (2010).The will (Vont) attempts to identify the will that the individual has to denounce certain actions as fraudulent.Perseverance (Pers) represents the level to which a person would be willing to denounce the fraudulent act and see the offender punished.The respondent should assess the variables for each scenario on a fivepoint scale.The questions and the alternatives are presented in the third part of the questionnaire.For such, the variable 'determination to denounce' (Obst) was created, which corresponds to the product of the variables Vont and Pres.After all, if the respondent has no desire to denounce (Vont = 0), it does not matter whether they would persevere in the denunciation, so the determination would be null.Likewise, if the respondent says that he/she would not persevere with the denunciation (Pers = 0), it does not matter if he/she wants to denounce, since the determination would also be null.On the other hand, if the respondent has an exacerbated desire to denounce (Vont = 4), and would persevere to the extreme (Pers = 4), he/she is indeed determined to denounce (Obst = 16).
The explanatory variables of the model are divided into two levels.At the level of the scenarios (level 1), two variables representing the costs and benefits involved in the situational analysis of whether to denounce or not are used.For the benefits, the variable used is Moral Intensity (IM), obtained by the individual's judgment as to the "seriousness of the fraud" and the "responsibility to denounce it".As for the costs, the variable used is Constraint (Const), which highlights the moral difficulties in denouncing a fraud within the organization.
The variables were measured for each of the proposed scenarios and were then coded on a scale between one to five points.The questions and the alternatives are presented in the third part of the questionnaire.
At the level of individuals (level 2), the variables used enable the analysis of the relationship of individuals with their organization, their profession and society as a whole.
These variables purport to explain the influence of concerning aspects (inherent) to the individual in his/her determination to denounce, regardless of the situation in which he/she finds him/herself.The assertions to be judged by the respondents are presented in the second part of the questionnaire.
The individual's relationship with the organization is measured by the variable Local Commitment (LC), a reflective latent variable obtained through three questions arising from the works of Aranya & Ferris (1984) and Jeffrey & Weatherholt (1996), aiming at contrasting the trade-off between the individual's loyalty to the organization and to his/her co-workers.
For Graham (1986), the denunciation will only occur if the individual is more committed (loyal) to the organization at the expense of his co-workers.The relationship of the individual with society is measured by the variable Public Motivation (MP), a latent variable obtained through five questions proposed by Kim (2005) and defined as the individuals' altruistic motivation to serve the interests of the community in which they are inserted (RAINEY; STEINBAUER, 1999; PERRY; HONDEGHEM, 2008).

MULTILEVEL MODEL
Most data that are studied in social sciences are from phenomena naturally arranged in a hierarchical or nested fashion, in which the subjects belonging to the same group share a set of stimuli that favors homogeneity (RAUDENBUSH; BRYK, 2002).In this perspective, the use of simple linear regression models becomes problematic, inflating the test statistics to be evaluated.Alternatively, multilevel models are used, also called hierarchical linear model (HLM).
Multilevel models are a generalization of regression methods with the advantage of taking into account the analysis of hierarchically structured data (OCONNELL, 2002).Thus, they not only identify the structural relationships but also the residual variability occurring at each level (GELMAN, 2006, p. 432-435), enabling the establishment of a specific model to express the relationship between variables within a given level and specifying how the variables within this level influence the relationships established at other levels (DRAPER, 1995, p. 115-147).For Hofmann (1997) the main advantages of using hierarchical models are: (i) recognition of the partial interdependence between individuals of the same group, while separating the individual residues and the group; (ii) verification of the variation of the different levels in the outcome.
The hypotheses were tested using HLM, estimated by maximum likelihood, on 2 levels: scenarios (level 1) and individuals (level 2).This approach allows identifying the influence of the level of the scenarios and the level of the individuals by estimating the variance within and between the groups by means of a static model.Thus, the evaluations are subject to the dependence of responses (since it consists in different evaluations of the same individual), but without dynamic correlation (since the evaluated scenarios have no relation with each other).
The models test the hypotheses as recommended by Raudenbush & Bryk (2002).So, initially, the Null Model was estimated -which uses no explanatory variable at any level -for the dependent variable: Determination to Denounce (Obst).The analysis of this model is essential in the multi-level approach, since from its estimation, information is obtained on the proportion of the variability at each level of analysis; i.e., the results of this model indicate whether the use of the multilevel approach is appropriate or not.The null model for the Obst variable was estimated according to the equations below for levels 1 and 2. Finally, the entire model is complete: For the model, Obstij is the determination to denounce of respondent j in scenario i; γ00 is the average determination to denounce of all respondents; rij is the deviation of the scenario i in relation to the average of the respondent's determination to denounce j; u0j: deviation of respondent j in relation to the average determination to denounce of the respondents.
The second estimated model considered the individual control variables (level 2): age, gender and marital status.The age variable is continuous, while the others were treated as binary: gender (1.female, 0. male); marital status (1.married or cohabiting; 0. others.).
The third model input variables of individual control (level 2) as evidenced in the literature: Commitment Lo9cation(LC); Professional Identification (IP); Public motivation (MP).These variables were treated as continuous and were centered around the mean of the average responses of the individuals.
The fourth model considered the intrinsic situational variables (level 1).The variables of level 1 (Const and IM) were centered around the mean level reported by the respondents themselves among the three proposed scenarios, thus identifying, how the variation in the responses to each scenario varied according to the respondent.The fourth model for the determination to denounce is given by the following equations: For the model: Const is the constraint of respondent j in scenario i; IM is the moral intensity of respondent j in scenario i; LC is the degree of commitment to the organization of respondent j; IP is the professional identification of respondent j; MP is the level of public motivation of the respondent j.The rest of the variables presented in the model follow the definitions presented earlier.
Models 2 and 3 were not specified according to their equations with the purpose of giving greater parsimony to the section; however, it is noteworthy that its specification follows the same principles of those developed in models 1 and 4.

DESCRIPTIVE STATISTICS AND FORMATION OF VARIABLES
Most respondents are men (64.5%).The number of married or cohabiting respondents constitutes the majority of the respondents (60.5%) and most have children (56.5%).The average age of the respondents was 37.6 years, and ten subjects did not report their ages.
Regarding the type of work, 54% of respondents perform management accounting functions and 41.9% perform activities related to financial accounting.Other highlighted activities were managerial functions and audit activities, totaling 44.4%.This information is in Table 1.
For analysis of the latent variables' composition, a Confirmatory Factor Analysis with the aid of the SmartPLS software was performed, estimated by the method of Partial Least Squares (PLS).Table 3 shows the correlation matrix of the variables used in the study, in addition to the average and standard deviation of the variables.The relatively high correlations between the binary variables of married respondents with children depict this trend, which was confirmed by a chi-square test (χ2 = 100.9,df = 1).Additionally, there is a predominance of negative correlations between the variable on Constraint and other variables, providing evidence of a negative influence of such aspects on the denunciation.The high correlations between the situational variables and the variables of the dependent group also highlight its eventual importance in explaining whistle blowing.Finally, it is important to analyze the specific characteristics of the scenarios, according to their descriptive statistics presented in Table 4.As a result, it is noteworthy that respondents tend to consider situations in which superiors demand fraudulent practices from their co-workers as more embarrassing to be denounced than those situations in which the individuals themselves (regardless if they are in a similar or higher position in the organization hierarchy) decide to incur in fraudulent practices.This result can be explained by the fact that in the first case (Scenario A), there is a consideration that the fraud received the consent of the organization, which would inhibit the denunciation (Treviño, 1986;Simons, 2002;Flynn & Wiltermuth, 2010).

MODEL RESULTS
According to Short, Ketchen, Bennett & Toit (2006, p. 259-284), the significance of individual changes in performance is tested in two ways.The first is a χ2 test that compares the statistics of deviations between the situational effect model and the null effect model.The second features a t-test for fixed effects and χ2 test for variance components.From this conceptual framework, the four proposed models will be analyzed.
Initially, the null model, which shows the variance decomposition between levels, by calculating the intra-class correlation coefficients (ICC), was applied.The results for Determination are in Table 5, which shows that 75.34% of the variability occurs between individuals (χ2 = 23.09) and a significant percentage of variance of the dependent (24.66%) was due to the situational aspects of each scenario.Jointly, the results show that both levels are important in the explanation of the variability of the Determination to Denounce.Though seemingly small, the ability to explain the situational aspects is important, especially when considering the complexity of relations in the field of Social Sciences.Thus, as stated by Goldszmidt, Brito & Vasconcelos (2007), an ICC of 24.66% is outstanding in its explanatory power.This fact thus supports the first hypothesis proposed.
To attempt to explain the variables that justify such variability, as well as expand the explanatory power of the analyzed model, explanatory variables for both levels were incorporated.Initially, the variables related to level 2 were inserted.Neither explanatory variable was significant in the model.Further, the model as a whole also was also not significant in explaining the variability of the dependent variable (χ2 = 6.05, gl = 6), which can be demonstrated by the small decrease in the value of the degree of model misalignment (∆Deviance = 0.25). 1 Table 6 shows the results of the model.
Level 2 Finally, the model is estimated by inserting the variables pertaining to the first level, to test hypotheses 2 and 3.Both situational variables were statistically significant and with signs aligned with the expectations.Thus, situations that highlight the moral intensity of the individuals (with regard to the seriousness of the fraud and the responsibility of the individual to denounce it) increase the determination of the individual to denounce (γ20 = 3,04 ** ).In parallel, situations that lead to a greater constraint on behalf of the complainant reduce the Determination to Denounce (γ10 = -1.09** ).Furthermore, analyzing the model as a whole, it is noted that, in this case, it was statistically significant in explaining the variability of the dependent variable (χ 2 =218.03;gl=8), which is evidenced by the significant decrease in the value of the degree of model misalignment (∆Deviance=79.46).Table 7 shows the results of the model.
1 A similar result is obtained by considering only some of the level-2 variables, i.e., personal control variables.The model as a whole was also not significant in explaining the variability of the dependent variable (χ 2 =3.18; gl=4), which is evidenced by the significant decrease in the value of the degree of model disadjustment (∆Deviance=0.28).
Obst General Average (γ00) 6.88 Considering the results, the importance of considering a second level of analysis regarding the denunciation in organizations is clear.More specifically, the results of the multilevel model showed that the situational variables that date back to the intrinsic costs and benefits present themselves as being statistically significant and with coefficients in agreement with that expected, i.e., a positive effect for Moral Intensity (IM) and negative effect for Constraint (Const).Thus, analyzing the scenarios in more detail, it is noted that the average constraint observed was higher for the case where the co-worker was forced to commit fraud (Scenario A) and had a lower mean in the situation where the superior commits fraud (Scenario D).This fact makes it clear that, with respect to moral aspects (intrinsic), any misconduct by superiors tends to be seen as subject to more severe punishment.
In contrast, the explanatory variables at the individual level are not significant in any of the presented models.This result may stem from several factors.Initially, as emphasized in the literature, features of Brazilian culture may influence whistle blowing in a more general perspective, i.e., the alleged existence of a third level of analysis.Thus, these results may only be valid for the Brazilian environment.
The results of the non-significant effect of the Location of Commitment (LC), although going against much of the literature -as exposed by Taylor & Curtis (2010) may have led to a problem of social desirability, i.e., it was denoted to the respondent an intention to present opinions and views that are well accepted by society (and by the regulatory entity of the profession) instead of expressing real opinions (FISHER, 1993).
Finally, although the gender variable was included only as a control variable, it should be noted that its significance in the final model (indicating that women tend to have greater Determination to Denounce fraudulent acts than men) deserves to be noted, according to the perspective adopted and developed by Nguyen et al. (2008).

CONCLUSION
After examining three different scenarios of fraudulent situations, this study provides evidence on the relative importance of fraud shaping the whistle blowing process.Overall, the results highlight the importance of individual and situational aspects also towards the denunciation, noting the importance of considering this phenomenon from a multilevel perspective.More specifically, it demonstrates the importance of situational intrinsic incentives arising from fraudulent situations and, in contrast to the results obtained in the international literature, the variables used to measure the individual level were not statistically significant for the Brazilian context.Finally, we highlight some of the major empirical contributions that this work brings.
Due to non significant effect of Professional Identification, perhaps it would be pertinent that the accounting students be more often exposed to teaching cases involving situations of fraud that were perpetrated by means of accounting information.Additionally, the Federal Accounting Council (CFC) could also consider the possibility of undertaking efforts to enhance the relevance of the work of the accountant in the prevention of fraud, considering, for example, the adoption of the code of ethics of the International Federation of Accountants (IFAC).
With the consequent internationalization of enterprises, it should be noted that the importance of companies complying with the requirements of laws such as the Sarbanes-Oxley Law, which determines the implementation of a formal Denouncement Channel so that This section (back page), 4 scenarios (A, B, C, D) characterizing possible situations in your professional activity are presented.For each scenario, as well as their respective action, the best alternative (only one) in each of the assertions.

3
Disagree Agree Thus, it is expected that this study contributed to the growing literature on the relationship between internal controls, organizational norms and individual behavior (TAYLER; BLOOMFIELD, 2011).As a main suggestion for future research, the role performed by the respondent within the organization may be considered, taking the significant found by O'Higgins & Kelleher (2005) into consideration.
am very happy to have chosen to be an accountant.

TABLE 1 -DESCRIPTIVE STATISTICS OF RESPONDENTS
Of the indicators, only the one referring to personal values (Vals) presented a low factor and statistically non-significant loading, being excluded from the analysis.The rest showed high and statistically significant factor loadings.The significance of the correlations was obtained by means of the bootstrapping method, with a thousand iterations.The convergent validity was confirmed by cross loadings tables, which show high factor loadings in only one of the constructs, in addition to the AVE (Average of Variance Explained) Root of each factor being greater than 0.5.For the discriminating validity analysis, it was noted that the correlations between the factors presented low values and, in addition, considering the criterion ofFornell & Lacker (1981), it was noted in all cases that the AVE Root was higher than the values of the correlations.The reliability was perceived by the high values of CR (Composite Reliability), all greater than 0.7, and high Cronbach Alphas.The results are shown in Table2.