RELATION BETWEEN SELF-DECEPTION AND TAX EVASION IN BRAZIL

Objective: This paper intends to identify the relationship between self-deception and tax evasion in Brazil, considering that these variables may be related, given the taxpayers’ behavioral diversity . Theoretical framework: Contents related to self-deception and human behavior. Method: Data collection by applying questionnaires of the Self-Deception Enhancement and the Impression Management subscales in February and March 2020. Logistic regression was used under univariate and multivariate analysis to identify the relationship between the variables. Results and conclusion: The indicators of each construct obtained in factor analysis acknowledged that the increase in propensity to self-deception decreases the chance of tax evasion. Intrinsic and extrinsic motivations of taxpayers suggest the need for public/private management to establish greater tax transparency of socioeconomic and tax factors, aiming to reduce and inhibit tax evasion practices. Implications of the research: The findings are helpful to assist Public Administration in the application of effective strategies to promote inhibition of tax evasion levels and effects, considering the diversity of taxpayers’ behavior, without ignoring their intrinsic motivations, since these are essential to understand the attitude towards illegal practices. Considering the consequences brought to the economy and society, tax evasion’s inhibition is necessary for public funding’s monitoring and constitutional principles’ guarantee of respect. Originality/Relevance: The study approaches the topic from a perspective that investigates human behavior through social strategies that can influence the practice of tax evasion and contribute to tax gap reduction.


INTRODUCTION
Self-deception involves a false belief of the individual, who may exhibit a truthconscious behavior.Trivers (2000) characterizes such a reinterpretation as an unconscious falsification of reality present in consciousness.In this sense, Mele (2001) states that theories about self-deception emphasize cognitive and perspective biases, such as misinterpretation and selective attention; however, he also considers that there is an influential social component in this behavior, in which other people can be the means to justify the self-deception of the individual, i.e., he is able to deceive other individuals, hide information or deceive himself.Mele (2001) suggests that self-deception raises two paradoxes: one concerns the state of mind of the self-deceiver, termed "static".The other, to the process of self-deception, called "dynamic" or "strategic".For the author, the static self-deception simultaneously presents contradictory beliefs, while the dynamic intends to deceive itself without making its intentions ineffective.
These paradoxes led philosophers to state that self-deception is conceptually possible or even coherent (Paluch, 1967;Haight, 1980).Borge (2003) holds that accounts of selfdeception relinquish core elements of psychological notions of "self" or "deception" to avoid paradoxes, allowing the individual to think that a particular attitude is justified.In this sense, Sahdra and Thagard (2003) argue that the individual chooses a distinct type of motivated irrationality.
From this perspective, considerations such as those of Rorty (1988) and Pears (1981) allow the phenomenon of self-deception to be organized into two groups of individuals: those who maintain that the cases of self-deception are intentional, called "intentionalists", and those who deny this statement, called "revisionists".Intentionalists find the model of intentional interpersonal deception appropriate, as it explains the selectivity of deception and the apparent responsibility of the deceiver, as well as providing a clear way to distinguish deception from other types of motivated beliefs (e.g.illusions).In turn, revisionists reformulate the requirement of intention, the requirement of belief, or both.For Bermúdez (2000), reviewing these aspects does not involve beliefs, but an attitude; thus, self-deception also assumes intentional behavior.
The problem that intentional models of self-deception face is that of the dynamic paradox, which postulates that it seems impossible to form an intention of the individual to make himself believe in what he currently does not believe or believes to be false (Bermúdez, 3 2000).In this sense, the author states that intentionalists can consciously deceive themselves and believe this, as well as losing their belief over time, completely forgetting their original misleading intention.Thus, the individual involved in illegal behavior can destroy records of such behavior and create evidence that obscures it, forgetting to have carried out illicit acts for a period.
Discussion of the topic sheds light on how motivation affects belief retention, questioning the notion of belief and the limits of these psychological concepts.To the extent that self-deception represents, individually and collectively, an obstacle to self-knowledge, there arises a problem of existential concern, suggesting a distinct possibility for individuals to live with distorted views of themselves, others and the world, which can make them alien to themselves and blind to the nature of their significant moral commitments to themselves and society.Dings (2017) argues that while self-deception tends to be conceptualized as an individual process, it can also be a process influenced by the social context.For Mazer and Ariely ( 2006), self-deception presents evidence to explain dishonesty.
Given the estimates of tax evasion in Brazil, where, in 2018, the National Union of Prosecutors of the National Treasury estimated that 345 billion reais in taxes were evaded, in addition to the research revealed that the behavior of the individual influences the illicit practices (Mittone, 2006;Torgler, Schaffner, & Macintyre, 2007;Thomas, 2015), the present study is justified for exploring the strategies of self-deception, aiming to know and understand how they interfere in the behavior of individuals in the practice of tax evasion.Such understanding may provide for effective measures to minimize its effects.
Therefore, it is worth asking whether the strategies of self-deception can explain the tax evasion of Brazilian taxpayers.Thus, the overall objective of the research is to identify the relationship between self-deception and tax evasion in Brazil.To test this hypothesis, the Balanced Inventory of Desirable Responding (Paulhus, 1991) scale consists of 40 questions, with 20 questions used to evaluate self-deception, called the Self-Deception Enhancementsubscale, and the others used for print management, through the sub-scaleImpression Management.
As for originality and theoretical purpose, this study approaches the topic from a perspective that investigates human behavior through social strategies that can influence tax evasion.As to practice, according to Guthrie (2014), the accountant plays an important role in ensuring the effective and sound functioning of tax systems, as well as assisting in the fight against tax evasion.
The research intends to contribute with the office of the accountant in order to guide taxpayers, from the behavioral aspects, as to the consequences of the actions taken in the face of tax evasion, and also to collaborate with the Public Administration in the possibility of determining actions that may discourage tax evasion and dishonesty.

The concept of self-deception
Self-deception (self-deception) is a psychological construct defined as the way of organizing knowledge, in which a person can simultaneously store both true and false knowledge as if they were both true, either by socialization or by the very nature of the act, so as to make it possible to deceive himself and others (Trivers, 2000).Giannetti (2005) states that self-deception permeates much of the choices and judgments we make, and that this phenomenon is based on the capacity we have to sincerely believe that we are what we are not.Dings (2017) defines self-deception as a phenomenon that involves a process and originates in a motivation or intention that leads to self-deception, which can be formulated in a new belief or used in the maintenance of an existing belief or attitude.For this author, such a process results in a final state of self-deception that, when it occurs without interference from others, is termed "solitary self-deception"; however, with the involvement of others, it is termed "social self-deception".Social self-deception, unlike solitary self-deception, is a strategy employed by self-deception to deceive other people.The author states that the social selfdeception includes in its repertoire verbal statements, facial expressions, body language, among other behaviors and attitudes.Dings (2017) still conceptualizes social self-deception in positive versus negative and in intentional versus unintentional.For the author, an individual may encourage another in a positive sense, making their beliefs and/or attitudes similar to each other, while intervention is negative when the self-deceptive actively prevents other individuals from forming a belief.
The requirement that deceivers hold contradictory beliefs increases the "static" paradox, as it seems to represent an impossible state of mind, namely consciously believing that p and ~p at the same time.As a deceiver, she must believe that ~p and, as deceived, she must believe that p.
Intentionalists agree with the proposition that self-deception is intentional.When the deceivers themselves are not aware of their contradictory beliefs or their misleading intentions, no paradox seems to be involved in the act of deceiving oneself.Many approaches use some combination of psychological and temporal division (Bermúdez, 2000); however, Gendler (2007) evaluates self-deception as a state of imaginative pretension, rather than as a belief.
An example of social self-deception in tax evasion can be characterized when an individual induces another individual to adopt illicit practices, claiming to have carried out an illegal practice and not having been identified, i.e. by adopting a persuasive strategy towards social self-deception.Belief induction can be made through false or misleading information: when the individual adopts the illicit practices without influencing others, he practices solitary self-deception; however, when he encourages one or more individuals to adopt practices similar to those of self-deception, it is considered as social self-deception.Robert (1982) argues that self-deception does not generate belief in the usual sense of the term, as the individual feels willing to admit as true or false a relevant proposition.It states that self-deception creates a conflict between different types of attitudes, namely, complete beliefs and mere statements.The author explains this phenomenon through an individual selfdeceived by an unfaithful spouse, maintaining a belief in the doxastic alternative.
Among the main approaches used to explain self-deception is the deflationary account, which is marked by the rejection of literal interpretation, since the individual abandons the attribution of unconscious and inaccessible belief in the face of the doxastic alternative (Mele, 2001).This view of self-deception is characterized by a state of mind, its product being the motivated false beliefs; that is, the individual believes in the false or unjustified proposition that is the content of his desire.
However, Velleman (2000) and Gendler (2007) criticize the accounts of self-deception because such authors reject belief explanations that distinguish them from other cognitive activities solely on the basis of the role of attitude motivation.Velleman (2000) presents in his research several examples with the purpose of showing that various other cognitive attitudes, such as imagination, can motivate action, as is used by other researchers, to exclude belief and fall short of truth.
In this direction, Gendler (2007) argues that if belief is the attitude of accepting a proposition with the aim of accepting something as true, then in situations of self-deception, thoughts and actions are guided by theories that are not really true.For Pinheiro, Penãloza, Monteiro and Nascimento (2014) individuals who have the belief in nature as an important personal interest can provide the human being with a greater tendency to pro-environmental behaviors, corroborating that the belief can influence the actions of individuals in various environments.
Another variable that has been shown to influence the behavior of individuals is the feeling of justice.Issifou and Beuren's (2020) study revealed that the theoretical implications on ethnic, that sense of justice contributes to explaining distributive, procedural, informational and interpersonal perceptions.
That said, one realizes that there is a consensus in the literature that self-deception exists as a psychological construct and is influenced by a mental state that can undergo alterations in function of the object of its beliefs.Accordingly, in the context of this study, the following basic hypothesis for research is forecast, to be tested empirically: the practice of tax evasion of Brazilian individuals is influenced by self-deception.

Sample and research instrument
107 individuals taxpayers of income tax, domiciled in Brazil and who submitted income tax declaration in one of the last five years participated in the study.The sampling used was classified as non-probabilistic and for convenience, as the data were not obtained randomly.
The questionnaires included self-deception, print management measures, as well as two questions about tax evasion.To measure self-deception, the Self-Deception Enhancement subscale was used, from the desirable response inventory (Paulhus, 1991).The Self-Deception Enhancement of 20 items measures sincere measurements of beliefs in desirable selfdescriptions, with sample items including "I am totally in control of my own destiny" and "I never regret it".A Likert Scale of 7 points was used (1 for untrue and 7 for true), and higher scores indicate the self-deceptive trend.The Impression Management subscale was measured by using 20 items from the desirable response inventory management subscale (Paulhus, 1991).Such a subscale includes items such as "I always declare everything at customs" and "I would never throw garbage on the street".The Self-Deception Enhancement measures positive views of their beliefs, while Impression Management measures the social friendliness of individuals' public image.Therefore, they were used to improve the evaluation of self-deception as a function of the self-correlation between subscales.
To identify individuals who have evaded taxes and to see if there was a difference between individuals who have declared and those who have not evaded, two questions were included from the Taxpayer Opinion Survey (2015), a public survey that collected opinions and evaluations from U.S. taxpayers: "In the last five years have you gone a little too far in presenting expenses to pay less tax?"; and "In the last five years have you failed to declare any income?"Thus, greater self-correlation was expected between Self-Deception Enhancement and Impression Management among participants who claimed to have evaded tributes in the past five years.

Methodological procedures
The database was made up of a sample of 107 respondents, and 61 categorical variables were analyzed, 40 of which were related to a Likert Scale that constituted two different constructs: one related to self-deception; the other to print management.
The answers related to the constructs were based on a Likert scale ranging from 1 to 7, with 1 assigned to "Not to be true" and 7 assigned to "Truth" with respect to the questions presented.Confidence intervals strictly less than 4 (mid-point of the scale) show disagreement with the item; intervals strictly greater than 4 indicate agreement; and intervals containing point 4 of the scale do not show agreement or disagreement (i.e., express impartiality).The bank's first 20 questions related to the "self-deception" construct, and questions 21-40 related to the "print management" construct.
To verify the linearity of the data and the relationship between the indicators of the constructs created, initially the correlations of the variables par Spearman (Hollander & Wolfe, 1999) were analyzed, since a significant correlation coefficient at the level of 5% is indicative of the existence of linearity.In addition, Bartlett's test (Mingoti, 2005) was performed to verify linearity in each construct.
In describing the questions following a Likert Scale, the mean, standard deviation and central trend measures were used, in addition to the bootstrap method (Efron & Tibishirani, 1993), which was used to calculate the confidence intervals of the means.The Mann Whitney test (Hollander & Wolfe, 1999) was used to verify whether there was a significant difference in the responses of individuals in the questions that considered whether tax evasion existed on the part of the respondents.
In order to create indicators representing each of the two constructs worked ("selfdeception" and "print management"), a factorial analysis was used.
To verify the convergent validity, the criterion of the Mean Extracted Variance was used (Fornell & Larcker, 1981).To measure reliability, Cronbach's Alpha and Composite Reliability were used (Chin, 1998).Discriminating validity is guaranteed when the variance extracted from one construct is greater than the variance shared from it with the others (Fornell & Larcker, 1981).To verify the dimensionality of the constructs, the criterion Kaiser (1958) was used.
To verify whether the constructs "self-deception" and "print management" exercised influence on the tax evasion of individuals, a logistic regression was adjusted (Agresti, 2002), and the backward method was used for variable/issue selection.For the backward method a level of 5% significance was adopted.In addition, the statistics of the Pseudo R² (Nagelkerke, 1991) and the Hosmer-Lemeshow test (Agresti, 2002) were used to verify the quality of fit of the model.The software used in the analyzes was R (version 3.6.1).

RESULTS AND DISCUSSIONS
To verify the linearity of the data, the correlations of the variables pair by pair were initially analyzed, since a significant correlation coefficient at the level of 5% is indicative of linearity.Through the Spearman correlation matrix (Hollander & Wolfe, 1999), 175 (22.44%) relationships were significant at the 5% level.
In addition, Bartlett's test (Mingoti, 2005) was performed to verify linearity in each construct.Thus, in all other constructs p-values lower than 0.05 were observed, indicating that there is significant evidence of linearity.

Factorial analysis
In order to create indicators that represented the constructs "self-deception" and "print management", factorial analysis was used in each one of them.The validation of each of the indicators created by the factorial analysis was analyzed by checking the convergent validity, the discriminant validity and the reliability of the indicators referring to each of the constructs.
To verify the unidimensionality of the constructs "self-deception" and "print management", the criterion of Kaiser (1958) was used.In both constructs, Cronbach's Alpha and/or Composite Reliability were greater than 0.60, thus evidencing their reliability.According to Kaiser's criterion (1958), all constructs were one-dimensional.There was converged validation for the "print management" indicator, since its extracted variance was greater than 0.40.Although the variance extracted from the indicator related to the "selfdeception" construct was 0.39, it was considered that the convergent validation assumption was achieved for this case by the fact that all factor loads of this indicator were greater than 0.50 and its extracted variance was very close to 0.40.There was discriminant validation on all constructs, since the maximum shared variance of each was less than the respective variance extracted.

Description
Table 1 presents the descriptive analysis of the categorical variables of the sample studied.One can see that all individuals think they can help others in many ways.The descriptive analysis for the only numerical variable was age, and the respondents had an average age of approximately 31 years, with a standard deviation of 9.84 years, the minimum age being 17 years and the maximum age being 65 years.Thus, it can be seen that the sample consisted of individuals less sensitive to penalties (Richardson, 2006;Gërxhani, 2007).
Table 2 presents the descriptive analysis and comparison of the items according to the construct to which they belonged, according to the Likert Scale.The data presented in Table 2 allow to infer that regarding the construct "selfdeception", individuals tended to agree with the Q1, Q5, Q6, Q8, Q10, Q13, Q15, Q17 and Q19 questions, since all their confidence intervals presented values higher than the midpoint of the scale.Thus, the data allow to infer that the sample studied presents a predisposition to selfdeception for these questions.
Respondents tended to disagree with item Q11, as their confidence interval showed values lower than the midpoint of the scale.In addition, Q11 was statistically lower than all other questions in the self-deception construct, since its confidence interval has values lower than all other confidence intervals of the questions that constituted this construct.
As for the "print management" construct, individuals tended to agree with Q26, Q28, Q33, Q35, Q36, Q38, and Q40 questions, since all their confidence intervals had values higher than the midpoint of the scale.Respondents tended to disagree with items Q23, Q25, Q27, Q29, Q31, Q32, Q37 and Q39, as their confidence intervals showed values lower than the midpoint of the scale.Individuals tended to be impartial to Q21, Q22, Q24, Q30, and Q34, as their confidence intervals covered values with the midpoint of the scale.
Given the above, the sample presented a propensity to self-deception and print management; therefore, individuals are expected to be willing to practice tax evasion.

Univariate analysis
In order to make a comparison between individuals who evade and do not evade tax, a univariate analysis was performed for each of the constructs ("self-deception" and "print management").It was considered that the individual evaded taxation if he had answered "Agree" to the question "In the last five years have you ceased to declare some income?" or to the question "In the last five years have you increased a little in the presentation of expenses to pay less taxes?".
Table 3 presents the comparison for the construct 'self-deception', where it can be observed that only Question 19 significantly influenced the tax evasion of individuals, i.e. it was the only question of the construct 'self-deception' which showed significant difference in the responses of individuals to tax evasion As shown in Table 3, 74.9% of the respondents stated that they had not evaded taxation in the last five years, however, only Q19 presented statistical significance at a level of 5%.These data show that, although individuals have a predisposition to self-deception, in view of the fact that 70% of categorical variables showed averages above 4 for respondents who claimed to have evaded tax and 75% of categorical variables showed averages above 4 for respondents who did not evade tax in the last five years, self-deception did not influence respondents to the practice of tax evasion.
Thus, it can be conjectured that the sample studied did not show sensitivity to uncertainty, in view of having been made up of young individuals, consequently prone to selfdeception (Richardson, 2006;Gërxhani, 2007).
It is worth pointing out that, although only Q19 presented statistical significance and there were no statistical differences between the individuals who declared to have evaded and those who declared not to have evaded tax, the data revealed inferences that may assist in the construction of a theoretical framework about self-deception, allowing to establish relationships with individuals of another age group.
Table 4 presents a comparison for the "print management" construct, where it can be observed that only Questions 26 and 30 significantly influenced the tax evasion of individuals; that is, these questions of the "print management" construct presented significant differences in their responses regarding the presence or not of tax evasion.As noted, although individuals had a predisposition to print management, only two categorical variables presented statistical significance as to the influence on tax evasion.However, it can be seen that there was no significant statistical difference between tax evaders and non-evaders.
In the light of the above, the data can contribute to future research, starting from analyzes that involve specific groups of taxpayers, in view of the incipient researches that study the relationship between print management and tax evasion.

Multivariate analysis
The logistic regression response variable was defined as follows: if the individual had answered "Agree" to the question "In the last five years have you ceased to declare some income?" or to the question "In the last five years have you increased a little in the presentation of expenses to pay less taxes?", it was considered that the respondent evaded tax; otherwise, that the individual had not evaded tax.Items with a p-value of less than 0,250 in the univariate analysis were selected for the multivariate analysis.Thus, two logistic regressions were adjusted: one considering the significant items of the univariate analysis of the "self-deception" construct; another, the significant items of the univariate analysis of the "print management" construct.For this logistic regression model, the backward method was used for the final selection of variables, considering a significance level of 5%.
Table 5 presents the initial and final multivariate models for tax evasion, according to the selected items of the "self-deception" construct.According to the final model, it can be concluded that only Question 19 was significant for the tax evasion of the individual, with each increase of one unit on its Likert Scale leading to a decrease of 23% [5%; 37%] in the chance of tax evasion.Thus, the greater the propensity to self-deception, the lower the chance of tax evasion.In addition, by the Hosmer-Lemeshow test, the model presented a suitable adjustment (p-value = 0.996), not rejecting the null hypothesis of the logistic model adjustment to the worked data.The Pseudo R² Nagelkerke showed that Question 19 was able to explain 4.6% of the variability of the occurrence of tax evasion in individuals.
Table 6 presents the initial and final multivariate models for tax evasion according to the selected items of the "print management" construct.According to the final model, it can be concluded that only Question 30 was significant for the tax evasion of the individual, with each increase of one unit on its Likert Scale leading to a decrease of 20% [4%; 33%] in the chance of tax evasion; that is, the greater the propensity to print management, the lower the propensity to tax evasion.In addition, by the Hosmer-Lemeshow test, the model presented a suitable adjustment (p-value = 0.795), not rejecting the null hypothesis of the logistic model adjustment to the worked data.The Pseudo R² Nagelkerke showed that Question 30 was able to explain 4.5% of the variability of the occurrence of tax evasion in individuals.
Considering that only Q19 of the "self-deception" construct and Q20 of the "print management" construct presented statistical significance to explain tax evasion, in which the propensity to tax evasion reduces as the propensity to self-deception and print management increases, the data does not corroborate the statement of the scholars when they claim that selfdeception does not aim at the truth; that is, for the sample studied, the greater self-deception and print management, the lower the chance of tax evasion.
This statement can be verified through indicators developed for the purpose of factor analysis, as shown in Table 7.The data in Table 7 shows the comparison for the constructs "self-deception" and "print management", where it can be observed that none of the indicators significantly influenced the tax evasion of individuals; that is, none of them presented significant difference in the responses of individuals regarding tax evasion.For the sample studied, however, it is noticed that an increase in the Likert Scale generates an increase in the indicators 1ºQ, 2ºQ, and 3ºQ between the individuals who declared not to have evaded tax and those who declared to have evaded, evidencing that self-deception and print management decrease the propensity to tax evasion, as this, in turn, increases self-deception and print management.
It was expected that, the higher the average of the indicators for each of the constructs, the greater the propensity to self-deception and consequently to tax evasion; however, the relationship was inverse for the sample studied, which can be differentiated for a larger sample, since the average of the indicator "self-deception" for the individuals who evaded was 3.91, close to the average of the Likert Scale.
In the light of the above, it can be inferred from the data that, although the sample is not statistically significant, it can be observed that, when the self-deception propensity indicator increases, the propensity to tax evasion is lower.This is because the taxpayer may understand that evading is not an illegal practice.

CONCLUSIONS AND RECOMMENDATIONS
The present study aimed to identify the relationship between self-deception and tax evasion in Brazil.The analysis and confirmation of the research hypothesis were carried out by statistical tests obtained through logistic regression.Self-deception was analyzed and considered, using the Self-Deception Enhancement subscale and Impression Management, from the desirable response inventory to measure self-deception.In order to deepen the analysis, the relationship between self-deception and tax evasion was studied between the groups of individuals who declared and those who declared that they had not evaded tax.
Based on the statistical tests obtained through logistic regression under the univariate and multivariate perspective, from the analysis of each construct variable, there was no statistical significance.However, using the indicators of each construct obtained in the factorial analysis, it was confirmed that an increase in the propensity to self-deception decreases the chance of tax evasion.
Faced with the theory of self-deception, it is understood that the individual stores true and false knowledge as if they were both true, through socialization or by the nature of the act, deceiving himself.In the tax environment, self-deception is exemplified by the practices of excessive expenditure or under-declaration of income, where the individual acts rationally by analyzing the cost-benefit of the act and does not assess the harm caused to society by internalizing a sense of justice that he already contributes enough fiscally or that he does not get the expected return from public services.
From this perspective, it is suggested to increase the levels of tax transparency of socioeconomic and fiscal factors, aiming to reduce the information asymmetry between public management and citizens, as well as to increase their level of trust in institutions.The results obtained in this research reinforce that the motivational factors of self-deception are dependent on the intrinsic and extrinsic stimuli of individuals, being conditioned by behavioral factors.
Therefore, the findings are useful to assist the Public Administration in implementing effective strategies in promoting the inhibition of levels and effects of tax evasion.These results highlight that the bodies responsible for curbing the adoption of illegal practices, in the tax sphere, must consider the diversity of taxpayers' behavior, without ignoring their intrinsic motivations, since this is determinant for understanding their attitude to illegal practices.Such an understanding also helps the public administration to define a more effective way of communication with citizens, generating greater transparency of the tax system.Thus, tax authorities will be able to draw up tax enforcement strategies targeting taxpayers who fail to comply properly with tax obligations, with a view to reducing the tax gap.Because of the consequences for the economy and society, a ban on tax evasion is necessary for the monitoring of public funding and for ensuring respect for constitutional principles.
The limitations of this research are explained so that due care is taken in the use of the results found here: although the questionnaire has been validated and its reliability assessed, the respondents' interpretations of the questions can be distinct because they reflect the understanding/judgment that each one has of the questions.Therefore, the results obtained cannot be generalized to any self-deception and tax evasion behavior.
Future work may be directed to investigate a defined target audience, as well as done with other data collection instruments.It is recommended to investigate other older taxpayers and other biases related to self-deception, such as confirmation, in order to improve empirical evidence about the individual subjective factors that determine an individual's fiscal conduct.

Table 1
Descriptive analysis of categorical variables

Table 2
Descriptive analysis of variables by "self-deception" and "print management" constructs Source: Search data.

Table 4
Univariate analysis of "print management" construct issues regarding tax evasion

Table 5
Multivariate analysis of "self-deception" construct items

Table 6
Multivariate analysis of "print management" construct items Source: search data.

Table 7
Univariate analysis of indicators created by factor analysis