Analysis of the Effect of Fraud Triangle Dimensions, Self- Effi cacy, and Religiosity on Academic Fraud in Accounting Students

* Corresponding author, Email: 1ireniamelati20@gmail.com, 2wilopo@perbanas.ac.id, indah.hapsari@perbanas.ac.id ABSTRACT Academic fraud is a bad habit done by students, particularly the college students. This bad habit includes cheating, opening the answer key during the test, or involving teamwork on individual homework. This academic fraud occurs because of several factors. This research aims to analyze the factors that infl uence academic fraud consisting of fraud triangle dimensions (pressure, opportunity, and rationalization), self-effi cacy, and religiosity. This study used the respondents of the accounting students at STIE Perbanas, Surabaya. They were taken by using a convenience sampling method. This research is a quantitative research conducted using a multiple regression analysis. The results of this research show that the variables of pressure and rationalization have an effect on academic fraud, while the variables of opportunity, self-effi cacy, and religiosity have no effect on academic fraud.


INTRODUCTION
Fraud is a problem that often occurs around us, both on small and large scale, including corruption. Almost every day, the mass media expose news related to fraud such as corruption, money laundering, gratuities, bribery, and so on. In Indonesia, day to day fraud is increasingly entrenched and diffi cult to overcome. This is because the fraud perpetrators are educated and experienced people. Research proves that more than 70% of corruptors have the Bachelor's degree (Wilopo 2016: 37).
Education plays an important role for developing the national character and preventing the corruption. Therefore, good education is expected to be capable of providing good values for students. Unfortunately, the education system in Indonesia has long been indicated by dishonesty by the students, including in colleges, commonly referred to as academic fraud. This dishonesty trend poses a variety of threats in the business world. So academics are challenged to avoid this dishonesty and are expected to be able to respect educational ethics and the moral development of undergraduate education (Deliana, at al, 2017).
Academic fraud that occurs in Higher Education, especially on accounting students at STIE Perbanas, vary, ranging from cheating during the test and getting the answer during the test via cell-phone to serious violations such as asking other student to sign their attendance list, falsifying sick permission document, falsifying signatures of their parents and even their lecturers. This is evidenced by the notifi cation of face publications, the identity of the perpetrators, and written statements of perpetrators of fraud on the campus wall magazine board. The consequences they have to receive are very diverse, such as aborting courses that they have cheated and even suspension. However, in fact these sanctions did not provide a deterrent effect on students instead they still dared to cheat to get what they wanted.
Academic fraud always occurs at STIE Perbanas Surabaya, especially during Midterm Test (UTS) and Final Test (UAS). This is evidenced by the data of students who did dishonesty in the form of cheating, opening notes, and the like, starting from the odd period of 2013/2014 to even period of 2017/2018 as represented in Figure 1 and Figure 2.  Figure 1 shows that academic fraud by the Accounting students often occurs in the even period. Total fraud during the odd period was 38 students over a period of 5 years. The highest fraud occurred in the odd period of 2015/2016, or 15 students. In addition, Figure 2 presents a description of academic fraud viewed from the midterm test or fi nal test sessions. Figure 2 presents data that academic fraud tends to be higher at midterm test than at fi nal test. This is the evidence of the greater fraud rate during midterm test than the fi nal test in each semester. The highest fraud occurred in the fi nal test in odd semester of 2015/2016, or 13 students, and 10 students in even semester of 2017/2018, 8 students in odd semester of 2013/2014, while the lowest fraud occurred in midterm test in even semester of 2013/2014 and 2017/2018, and 2016/2017. Dishonesty in the world of education, hereinafter referred to as academic fraud or academic dishonesty, can be interpreted as fraudulent acts committed by students, such as cheating using small paper or cell-phone, copy pasting from the internet, cooperating with friends during exams, and others (Santoso and Yanti, 2015). Academic fraud can be defi ned as a method or action carried out with an intentional element to achieve a goal (good results) that comes from dishonest behavior. Academic fraud conducted by students, according to Fitriana dan Baridwan (2012), is an attempt to get something dishonestly.

Research Objective
Based on the results of the above phenomena and problem formulation, the purpose of this research is to develop the previous research and to test as well as to analyze the effect of the Fraud Triangle Dimensions, self-effi cacy, and religiosity on the occurrence of academic fraud by the Accounting students.

THEORETICAL FRAMEWORK AND HYPOTHESIS Academic Fraud
Academic fraud is dishonest behavior carried out by students to get the results they want (Artani and Wetra, 2017). This is also explained by Albrecht et al., (2009) that fraud can occur because of three factors: pressure, opportunity and rationalization, or referred to as Fraud Triangle. This certainly can be applied not only in the accounting fraud scheme, but also in the academic fraud scheme. Furthermore, Purnamasari (2013) defi nes fraudulent behavior as an act carried out by students to cheat, obscure or deceive the instructor until the instructor thinks that the work or task collected is the task of the student's work.
According to McCabe & Trevino (1997) in Bolin (2004), academic fraud includes cheating on friends' work in any way, opening a notebook without the knowledge of the examiner, assignments done by other colleagues, copying the answers of other colleagues during the exam, and so on.

Pressure
Pressure is a condition where the perpetrators are in a state of urgency so that they inevitably are forced to do so to cover their needs (Tuanakotta, 2010: 208). Pressure can be said as an encouragement or motivation from within a person (internal factors) or from the surrounding environment (external factors) that causes a person to do an action (Apriani, at al 2017). Someone who is under pressure will tend to act fraudulently, and vice versa. This is intentionally done to be able to realize something he wants.
The pressure, in relation to academic fraud, felt by students varies, such as pressure from parents, peers, college where he studies, or demands from companies for high GPA standards to become employees (Murdiansyah, at al, 2017) . In addition, by doing academic dishonesty, students hope to be seen as a successful person, worthy of trust and can infl uence other colleagues (Artani dan Wetra, 2017). If a student feels a lot of pressure in him, there will be a possibility that he will commit fraud (Becker et al., 2006).
Pressure, according to Becker et al., (2006), includes assignments and examinations that are diffi cult to do by students, the graduation standards set, and inability to manage time.

Opportunity
Opportunity is defi ned as a situation that underlies a person to cheat. Opportunity generally exists before fraud occurs. Opportunity is obtained by the perpetrator from observing the situation around him (Tuanakotta, 2010: 211).
Opportunity is the second element in fraud triangle. A person can keep all pressures from anyone for what they feel, but if the perpetrators of fraud have a perception that they have the opportunity to commit fraud, they will immediately commit the fraud without being known by others (Tuanakotta, 2010: 211). A simple example of opportunity is e.g., the students can determine the right time to open the answers they have in their prepared sheet, by paying attention to the surrounding environment, especially the lecturer or supervisor of the exam. When the test supervisor is off guard or doesn't pay attention to the classroom, they think it is time for them to cheat.

Rationalization
Rationalization is usually done before committing a crime, not afterwards. Finding rationalization is a part that must exist in the crime itself, not part of the motivation to commit a crime (Tuanakotta, 2010: 212). (Nursani dan Irianto, 2016) Rationalization is generally in the form of reasons, such as: "no one is harmed for this action", "I did this for good purposes", or "some have suffered because of this, my integrity and reputation" (Wilopo, 2016: 283 -284).
After the crime is committed, this rationalization is abandoned and no longer needed. At fi rst the perpetrators feel guilty because they have resisted the existing rules, but when repeating the act a second time or so, they will feel easy and familiar (Tuanakotta, 2010: 212). Rationalization, according to Apriani, et al., (2017), includes: (1) cheating is something in common, (2) cheating to get a high score, (3) cheating to raise self-esteem, (4) cheating is an instant way to get value desired.

Self-Effi acy
Self-effi cacy is an expectation in humans that determines how much effort and how long someone will try to survive in facing problems and unpleasant experiences (Bandura, 1997) in Pudjiastuti (2012). According to Bandura (1997) in Ghufron and Risnawita (2011: 75), self-effi cacy is the result of a cognitive process in the form of decisions, beliefs, or expectations about the extent to which individuals estimate their ability to carry out tasks, or a certain action needed to achieve desired results. In everyday life, self-effi cacy leads us to determine challenging ideals and survive in the face of diffi culties. Bandura (1997) in Ghufron and Risnawita (2011: 78) explains that one's self effi cacy comes from four things, including: experience of success, other people's experiences, verbal persuasion, and physiological conditions.

Religiosity
Religion is a belief in God, belief in the creator of the universe. But, religiosity is defi ned as the level of religion that humans have or simply is the level of human belief in God.
Religiosity is manifested not only when a person performs ritual behavior (worship only), but also when doing good deeds which are driven by inner and outer powers. Religiosity leads a person to be able to avoid wrong actions. Almost all religions teach goodness in behavior. Religiosity has a negative effect on academic fraud (Purnamasari, 2013), it means that someone who has high religiosity will feel fear of God where he will believe in the existence of Karma or consequences for every action done in this world. Indicators of measuring one's religiosity, according to Glock and Stark in Ghufron and Risnawita (2011: 170), can be seen from the following dimensions: dimension of faith, dimension of worship, dimension of devotion and appreciation, dimension of religious knowledge, and dimension of practice and consequences.

The Effect of Pressure on Academic Fraud in Accounting Students
The biggest pressures experienced by the students include compulsion or coercion to graduate, competition to get high grades, too much workload and insuffi cient study time. Too much pressure can make students feel burdened in doing anything. This problem often occurs in students who feel that they have not found compatibility between something that they like and what they are interested in. Too much pressure will make students stress or depression because they cannot achieve the targets they have set. An achieving the targets, a student sometimes legalizes all ways so that the pressure gradually disappears. So, the higher the pressure perceived by students, the higher the possibility to cheat (Apriani, dkk 2017). Hypothesis 1: Pressure has an effect on academic fraud in accounting students.

The Effect of Opportunity on Academic Fraud in Accounting Students
Opportunities can occur because there is a gap or weak control that is utilized improperly. Opportunities for academic fraud among students include cheating, opening answer key during exams, opening mobile phones in the exam room, looking for answers from colleagues outside the class, and so on.
The results of research conducted by Deliana, et al (2017), Nursani and Irianto (2016), Fitriana and Baridwan (2012) and Becker et al., (2006) show that the variable of opportunity has an infl uence on academic fraud. For example, research conducted by Nursani and Irianto (2016) fi nds that the opportunities seen by students come from other sources, such as internet technology, class conditions, and connection with their senior. This explains that opportunities can occur when the lecturer is leaving the examination room, supervision during the exam is weak, students are in an environment where cheating often occurs, or students have strategic positions to commit fraud. So, these will further encourage students to cheat during the exam. Hypothesis 2: Opportunity has an effect on academic fraud in accounting students.

The Effect of Rationalization on Academic Fraud in Accounting Students
Rationalization, according to Padmayanti et al. (2017), is a process or a way to make something irrational become rational or can be accepted by common sense. Rationalization is generally in the form of reasons, such as: "no one else is harmed for this action", "I do this for good purposes" (Wilopo, 2016: 283) Research on the effect of the variable of rationalization on academic fraud was done by Padmayanti et al. (2017) where out of ten statements, there are three indicators with the highest scores: 1) if the test questions given by lecturers are easy, I can get good grades without having to cheat, 2) I often see my colleagues cheating, and this motivates me to take part in cheating, 3) I cheat only when I am pressed. The three indicators show that students rationalize any reason to be able to save themselves. They sometimes think that their actions are right without thinking about the long term behavior. So, if students have various rationalizations for their behavior, they will think that cheating is something in common. Hypothesis 3: Rationalization has an effect on academic fraud in accounting students.

The effect of Self-Effi cacy on Academic Fraud in Accounting Students
Self-effi cacy is often defi ned as one's selfconfi dence. Self-effi cacy is a belief in one's ability to regulate and carry out a series of actions needed to achieve his/her desires (Ghufron and Risnawita 2011: 73). Self-effi cacy, in relation to the academic fi eld, can be understood as the student's confi dence in his ability to do something. People who have high self-effi cacy are the people who have high selfconfi dence that they will be able to complete their tasks well. Conversely, if students have low self-effi cacy, they will perceive that the ability they have is not necessarily able to make him succeed through an exam or a problem. So it can be concluded that the higher the selfeffi cacy students have, the more likely they are to believe in their abilities and avoid cheating. Hypothesis 4: Self-effi cacy has an effect on academic fraud in accounting students.

The Effect of Religiosity on Academic Fraud in Accounting Students
Religiosity is the level of one's belief in the existence of God. This belief will later help in determining whether the deed is good or not. Someone with high religiosity will tend to avoid the deeds as perceived to be detrimental to his life in the future. He will also consider the existence of karma or consequences for bad deeds he /she has ever done. Thus, he/ she will always be careful in doing things and prefer a good way.
The higher the student's religiosity, the lower the academic fraud is. Conversely, the lower the student's religiosity, the higher the student's motivation to commit academic fraud. This study supports the research conducted by Herlyana, et al (2017), Zamzam, et al (2014), and Purnamasari (2013). Hypothesis 5: Religiosity has an effect on academic fraud in accounting students.

Framework
Based on the results of the explanation of the relationship between variables that have been stated, it can be made a framework of the effect of fraud triangle dimensions, selfeffi cacy and religiosity on academic fraud in accounting students. The framework is shown in Figure 3.

RESEARCH METHOD Population, Sample, and Sampling Technique
The population in this study is all students of STIE Perbanas Surabaya. The sample used in this study is students majoring in Bachelor's degree in Accounting. The aim of selecting the accounting students as the research sample in this study is that they are expected to e become accountants who uphold the professional code of ethics and integrity.
The sample size was measured done using Slovin formula with the aim to fi nd out the minimum number of samples that must be taken from the total population. Samples are taken from active students majoring in Bachelor's degree in Accounting at STIE Perbanas Surabaya during even period 2017/2018 consisting of 1,169 students, where the students are in the second, fourth, sixth, and eight or more semesters. By using the Slovin formula which is set at an error tolerance level of 10%, there will be 97.6 or 98 students who are sampled. The value of 98 is a minimum value for sampling, so that more than 98 students should be sampled in order to cover questionnaires that are not worthy of testing. The sample taken is 130 Accounting students at STIE Perbanas Surabaya, where the students in the second, fourth, sixth, and eighth semesters are represented by 30 students, while the students who are over eighth semester are represented by 10 students.
The sample was taken using a nonprobability sampling, with the Convenience Sampling technique. By using this technique, samples can be obtained from the members of the population easily to provide information to researchers.

Data and Data Collection Method
This research is a quantitative research, in which the data were obtained from questionnaires distributed to the accounting students of STIE Perbanas Surabaya. There were 130 questionnaires distributed to the students of the 2nd, 4th, 6th, and 8th semesters. It was represented by 30 samples, while students who were over the eighth semester were represented by 10 samples. Data from the questionnaire were processed when the questionnaire had bee fi lled in completely.

Research Variable
The variables include the dependent variable (academic fraud) and independent variables (fraud triangle dimensions, selfeffi cacy, and religiosity).

Defi nition and Variable Measurement Academic Fraud
Academic fraud is fraudulent action carried out by students intentionally, such as violating applicable rules, conducting dishonesty and performing unethical things. The fraudulent actions that are usually carried out by students include cheating on exams,

Figure 3
Research Framework opening answers that have been taken from home, opening answers written on mobile phones, copying friend's answers, helping friends cheat, falsifying sources of friend assignments, collaborating to complete tasks, asking other people to replace themselves during the exam, and so on.
The academic fraud variable was measured based on previous research conducted by Nursani and Irianto (2016), Purnamasari (2013), Fitriana & Baridwan (2012) with reference to the measurement indicators proposed by MCCabe & Trevino (1997) as stated in Bolin (2004). They are as follows: (1) trying to copy the work of a friend somehow, (2) using a notebook without the knowledge of the supervisor, (3) carrying out the task by another pressure, (4) copying the answers of other colleagues, (5) cooperating in doing homework, (6) helping others cheat, (7) quoting without mentioning the source, (8) studying the question and answer models of the next class, and (9) recognizing the work of other colleagues as his own work. The nine indicators are measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree".

Pressure
Pressure can be defi ned as a condition where someone needs to do something. The pressure perceived by students, either from internal or from external factor, will encourage their intention to cheat. Pressure that is often perceived by students includes pressure due to academic demands on campus, peer demands, and demands from parents. For this variable, the researchers refer to the measurement of variables carried out by Becker et al., (2006), where the indicator of the pressure variable is measured by: (1) excessive tasks for students, (2) the test given is too diffi cult to answer, (3) students fi nd it diffi cult to meet graduation standards for certain courses, (4) students cannot manage their time properly.
The four indicators was later developed into 10 item statements. The ten items will be measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree". The higher the assessment number given, the higher the tendency to cheat.

Opportunity
Academic fraud committed by students is based on the existence of opportunity, where students feel they are in the right condition that allows them to act fraudulently. Opportunities that are often seen by students are related to observations that they do to the surrounding environment. The measurement of opportunity variable is based on the previous research conducted by Deliana, et al (2017), Apriani, et al (2017), Nursani and Irianto (2016) which refer to the measurement indicators of the variable proposed by Becker et al. (2006) as follows: : (1) the teacher does not check the plagiarism of students, (2) the teacher does not change the pattern of questions and tests, (3) the students often observe the surrounding environment, (4) the teacher does not take precautions when students cheat.
The four indicators will later be developed into 10 items of statement. The ten items will be measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree". The higher the assessment point, the higher the tendency to cheat.

Rationalization
Rationalization is defi ned as the action of attempting to explain or rationalize behaviour or an attitude with logical reasons, even if these are not appropriate. Rationalization is often the reason for students to cheat because they have strong and "right" reasons according to their point of view, but are wrong in in general. This usually comes from the internal confl ict of students as a basis for legalizing fraud that they did (Nursani and Irianto, 2016). The measurement of rationalization variable is based on previous research conducted by Apriani, et al (2017) with indicators referring to research conducted by Becker, et al., (2006) as follows: (1) the instructor does not provide an explanation for dishonest behavior, (2) there are no strict sanctions for students who cheat, (3) the faculty does not detect fraud, (4) the sanctions given are not commensurate.
The four indicators will later be developed into 10 items of statement. The ten items will be measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree". The higher the assessment point given, the higher the tendency to cheat.

Self-Effi cacy
Self-effi cacy is self-confi dence or the ability of oneself to organize and carry out an action to achieve the results of an effort. Selfeffi cacy is a belief about what someone can do. For students, self-effi cacy can help them not to conduct fraudulent action. Self-effi cacy, in the academic fi eld, has some measurements. One of them is based on the cognitive processes of each individual. The statement is based on research conducted by Albert Bandura (1997) in Ghufron and Risnawita (2011: 80-81) which is measured based on several factors, including: (1) experience of success, (2) experience of others, (3) verbal persuasion, (4) physiological conditions.
The four indicators was later developed into 8 statement items. The eight items were measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree". The higher the assessment point given, the higher the tendency not to cheat.

Religiosity
Religiosity can be defi ned as the level of one's commitment to his religion. Religiosity can also be interpreted as the level of depth of an individual to believe in a religion that is balanced with the knowledge and experience felt by the individual by practicing religious values in the form of obeying God's predetermined rules, and performing all religious obligations with sincerity in daily life. One aspect of the religiosity variable that will be examined is the way a person behaves. The moral measurement uses the indicators proposed by Glock and Stark in Ghufron and Risnawita (2011: 170) as follows: (1) dimensions of faith, (2) dimensions of worship, (3) dimensions of devotion and appreciation, (4) dimensions of religious knowledge, (5) dimensions of practice and consequence.
The fi ve indicators will later be developed into 7 items of statement. The seven items will be measured using a 1-5 Likert scale. Point 1 shows "Strongly Disagree" and point 5 shows "Strongly Agree". The higher the assessment point given, the higher the tendency not to cheat.

RESEARCH RESULTS AND DISCUSSION Descriptive Test
Descriptive analysis is used to provide an overview of the variables in this study, such as the variables of pressure, opportunity, rationalization, self-effi cacy, and religiosity. The analysis of these variables is outlined on Table 1. Table 1 is a recapitulation of the answers of 130 respondents. The results of the analysis according to the fraud triangle dimensions (pressure, opportunity, justifi cation) show that the highest value is in the variable of pressure (X1). In this case, the respondents feel that they get academic pressure. This pressure can be in the form of the desire to get a high GPA, obtaining satisfactory grades and the obligation to pass in certain subjects. This is evidenced by the high average value of the variable of pressure (X1) of 3.41. In addition, the value of the variable of opportunity (X2) is at the interval "doubtful", and the variable of rationalization (X3) is at the interval "disagree. Nonetheless, the value of the variable of rationalization approaches is in the interval "doubtful". Viewed from the data presented, the variable of academic fraud (Y) is at the interval disagree", meaning that students do not agree with academic fraud. However, there is possibility that students will continue to cheat. This is because students still have several other factors that can strengthen their intentions to act dishonestly, such as feeling depressed, seeing opportunities, and having reasons to justify actions that violate the rules. Source: processed Data SPSS Table 2, shows the recapitulation of respondents' answers grouped by the year. Viewed from the highest mean value of the variable of academic fraud (Y), the class of 2014 is ranked fi rst, the second is for the class of 2013 and 2012, and the third is for the class of 2015. The three of them are indeed at the interval "disagree", but the class of 2014 is the highest.
Academic fraud that occurs in students in the class of 2014 can be caused by a number of factors, such as the feeling of being pressured to graduate soon, or they understand very well about their campus environment so that they can fi nd loopholes, and even they have their own logical reasons to justify actions that actually violate the rules.  Table 3 is a recapitulation of answers grouped by GPA. The highest mean value of academic fraud comes from students who have a GPA <2.00 -2.75, and the lowest value comes from students with a GPA > 3.50. Students with a lower GPA tend to cheat. This can be caused by a lack of sharpening of their own potential and low self-effi cacy, or there are other factors that are more dominant, such as feeling depressed during in college, seeing benefi cial opportunities, or having reasons to justify wrong actions.
Students with a high GPA tend to not commit dishonesty during exams or when doing assignments. This is because they have tried to improve their abilities and potential and had ways to deal with diffi cult situations. This is sometimes reversed with the condition of students who have a lower GPA. Even so, it does not mean that students with a low GPA have a shortage of everything. It could be because there are other factors that make their academic performance become not optimal.
The results of the recapitulation of respondents' answers in Table 4 show that there is a difference in the mean value of 0.03 on the variable of academic fraud (Y) that comes from students who do not know or have not known the term academic fraud. Even though it is a small value and both are still at the interval "disagree", it is still possible for students to cheat. If it is reviewed, this is due to the high mean value of the variables of (X1) (pressure), X2 (opportunity) and X3 (rationalization) for students who do not know the term. So, these three factors can also affect students to do unfair work while doing assignments or examinations.

Validity and Reliability Testing
Validity test is used to test the validity of a questionnaire. A questionnaire is said to be valid if the question in the questionnaire is able to express something that will be measured by the questionnaire (Ghozali, 2016b: 52). While reliability test is used as a tool to measure a questionnaire, which is an indicator of a variable or construct, whether the questionnaire is reliable or not (Ghozali, 2016b: 47). The following Table 5 shows a summary of the validity and reliability tests in this study.
The requirement of research using good primary data is that the statement indicators must be valid and reliable. In this study, the fi ve variables and each measurement indicator are valid. In addition, this validity test also indicates that the variable tested is able to be measured properly through the indicator or statement item displayed.
This study uses 55 items of statement as indicators of the measurement of the six  Table 5 above, there are three variables with an asterisk '*', this is because at the beginning of reliability testing, the statement indicators of the three variables are still not reliable. The method that can be taken is by removing a number of statement items until the Cronbach's Alpha value is more than 0.6, so that all reliable statement items are 49 statement items only. However, the remaining statement items can still refl ect the indicators of each variable properly.

Classical Assumption Test
Classical assumption test is used to test whether, in a regression model, the independent variables and the dependent variable or both have a normal distribution or not. In addition, the classical assumption test is also used to test whether there is a correlation between independent variables in the regression model. The results of classical assumption test in this study are as follows:

Normality Test
Normality Test is used to test whether in the regression model, the confounding variable or residual has a normal distribution. The statistical tool used to test the normality of the data is Kolmogorov-smirnov with a signifi cance level of 5%. Regression equations can be declared normal if the signifi cant value is ≥ 0.05, Ghozali (2016b: 157). The results of the normality test can be seen in Table 6.

Multicollinearity Test
Multicollinearity test aims to test whether in the regression model there is a relationship between independent variables. The tool used to conduct multicollinearity test is Variance Infl ation Factor (VIF). If the VIF value is < 10 or 0.1, this means that multicollinearity does not occur. However, if the VIF value is > 10, the data variable experiences multicollinearity (Ghozali, 2016b: 103). Multicollinearity test results can be seen on Table 7. Overall, the VIF value of each variable is below 10, so the fi ve independent variables have no symptoms of multicollinearity

Heteroscedasticity Test
The heteroscedasticity test is done to test whether in the regression model there are variance inequalities from one observation to another observation. If the signifi cance value is ≥ 0.05, this means that there is no heteroscedasticity, and if the signifi cance value is <0.05, this means that there is heteroscedasticity. The results of heteroscedasticity test can be seen in Table 8.

Autocorrelation Test
Autocorrelation test serves to test whether there is a relationship between confounding errors in the period of this study and those in previous studies. The tool used in this test is Durbin Watson. The results are presented in Table 9. The value of Durbin Watson in this study is 2.165, so the value is greater than Durbin Upper (dU), but smaller than the 4-dU value. Therefore, if it is translated in numbers, it will be 1.7941 < 2.165 < 2.2059. This means that in this study there is no autocorrelation in it.

Results of Analysis and Discussion
As shown on Table 10, it shows that of the fi ve variables, only the variable of pressure and the variable of rationalization that have an infl uence (because the signifi cance values are below 0.05), while the variables of opportunity, self-effi cacy, and religiosity have no infl uence. In addition, adjusted R 2 value is 0.447 or 44.7 percent, so it can be concluded that the independent variable is able to explain the dependent variable of academic fraud by 44.7 percent, while the remaining 55.3% (100% -44.7% = 55.3%) is explained by other variables outside of this research model. The result of the F-test indicated by table of F-calculated is 21.848 with a signifi cance probability of 0.000 < 0.05, therefore, it can be concluded that H1 is accepted which means the regression model is fi t. Thus, it can be used to determine that the variables of pressure, opportunity, rationalization, self-effi cacy and religiosity simultaneously have an infl uence on the variable of academic fraud.

Analysis of the Effect of Pressure on Academic Fraud
Based on the data in Table 10, the regression coeffi cient for the variable of pressure is positive 0.329, meaning that each increase one unit in the variable of pressure, there will be an increase in the variable of academic by 0.329 units, assuming other variables are considered constant. In addition, the signifi cance value is 0.002 (or below 0.05) so that it can be said that the variable of pressure has an effect on academic fraud. This result explains that Hypothesis 1 (H1) is accepted.
Pressure has an effect on academic fraud. This is caused by several factors, such as the presence of pressure due to the demands of parents, wanting to get a high GPA, demands to graduate in certain subjects with minimum score B, unable to manage time, and so on. In this study, the indicator that has the highest score is the statement "I want to get a high GPA" and "I have to maintain my GPA so that my performance looks good". From these two statements prove that students feel that results are the most important. They want good results without going through the complicated process. As a result, if the pressure perceived is high, the level of academic fraud will also be high. This supports the Fraud Triangle Theory which explains that pressure is one of the causes of fraud. This result also supports the theory of reasoned action which explains that a person's behavior can be seen from his intention. In this case students intend to improve their academic performance in an instant manner, so that dishonest behavior is caused by the intention.
Pressure or stress can be reduced by cultivating habits to be able to manage stress for students and creating a synergic and pleasant learning environment. If they are implemented well, it will be able to minimize the level of stress and academic fraud.

Analysis of the Effect of Opportunity on Academic Fraud
The regression coeffi cient of the variable of opportunity is positive 0.139, meaning that every increase one unit in the variable of opportunity, the variable of academic fraud will increase by 0.137 units, assuming other variables are considered constant. The signifi cance value of this variable is 0.139 which is more than 0.05. This shows that the variable of opportunity has no effect on the variable of academic fraud. This result explains that Hypothesis 2 (H2) is rejected.
Opportunity has no effect on academic fraud. It can be caused by various factors. The statement that has the highest score is "I have been ever caught cheating during the exam". From this item statement shows that the regulations in STIE Perbanas Surabaya have been said to be good because they have successfully detected and prevented students from cheating. The positive impact is that the students will be trained to believe in their own abilities and be more independent, and not dependent on others.
This study does not support the Fraud Triangle Theory where opportunity is one of the causes of fraud. This study does not support the theory because the opportunity in this study does not affect the occurrence of academic fraud. However, this study supports the Theory of Reasoned Action. This is because students' behavior can be refl ected in their intentions. They are able to assess profi table opportunities because they have demands to be fulfi lled.
Research related to this variable can be used as a benchmark that the existing rules have been implemented properly. In addition, students are expected not to do things that violate the rules because it will harm themselves and other partners who are also involved. It would be better if students prepare the material tested carefully so that so there is no need to depend on other colleagues.