Construct Validity Test of Integrity and Suryomentaram-style Introspection in Creating Anti-Fraud

The high level of corruption and fraud in Indonesia is one of the crucial problems that must be resolved immediately by the Government of Indonesia. Various efforts to eradicate corruption and the creation of anti-fraud will not be optimal if the habit of applying the meaning of integrity in actions and behavior, which begins with the determination of an employee or public official, is still mixed with the desire to enrich oneself and one’s class. The identification of the problem in this study includes 3 (two) things. First, does the construct of measuring instrument of integrity and Suryomentaram-style introspection supports the efforts to prevent fraud and corruption and fulfill unidimensionality by using confirmatory factor analysis? Second, does the measuring instrument of integrity and Suryomentaram-style introspection in this study consists of items that fit in measuring integrity and Suryomentaram-style introspection using the application of the polytomous IRT model? Third, do the items in the measuring instrument of integrity and Suryomentaram-style introspection in this study contain a response bias based on gender that can be detected through differential item functioning (DIF)?. This research was conducted with the aim of obtaining measuring instrument of integrity and Suryomentaram-style introspection that supports steps to prevent fraud and corruption at the level of implementing employees to managers in Indonesian companies and/or state institutions.


INTRODUCTION
The high level of corruption and fraud in Indonesia is one of the crucial problems that must be resolved immediately by the Government of Indonesia. During the Covid-19 pandemic, corruption and fraud have increased without the support of incessant arrests of perpetrators by law enforcement agencies. Since the enactment of Law Number 23 of 2014 Jo Law Number 22 of 1999, crimes of corruption and fraud occur not only in the Central Government sector, but also in Regional Government. Even in terms of modus operandi, corruption and fraud often involve collaboration between policy makers in the Central Government and Regional Governments as well as third parties in the process of procuring goods and services.
Various models of approach strategies have been developed to prevent corruption and fraud, one of which is the approach to improving employee integrity. Integrity is a tool to prevent corruption and fraud (Hariyani et al., 2015). Anti-corruption is formed based on the significant influence on moral integrity, lifestyle and organizational culture (Putri & Nihayah, 2019). Integrity is formed from honesty, discipline, obedience and responsibility which are the basis for the personal moral strength of workers in improving their performance in carrying out their functions and duties (Herman, 2019). However, the effectiveness of the integrity pact in the context of eradicating corruption and minimizing fraud is still far from being expected (Basuki, 2010).
Another strategy that is still rarely found in application in the scope of work life is the introspective approach strategy. This strategy has often been used in the world of health, especially in introspection survey. This introspection survey is generally carried out to find out how well the general public understands the dangers of a disease which is supported by other supporting factors. Introspection survey is a data collection activity by local cadres, leaders and youth related to the development of Healthy Villages (Kemenkes RI, 2019). The purpose of conducting an introspection survey is to make the surrounding community understand health problems, recognize the potential of villages that can be used in solving health problems, and create public awareness to minimize the emergence of health problems by optimizing the potential around them (Menteri Kesehatan, 2019).
The introspection that will be studied is the introspective model of Ki Ageng Suryomentaram. Ki Ageng Suryomentaram is a nobleman from the Javanese kingdom who is simple, unpretentious, populist, egalitarian, and has no desire for the throne and wealth within the kingdom. K.A. Suryomentaramstye introspection comes from the writing of kawruh jiwo (knowledge of the soul) in understanding oneself which includes: Kaweruh Jiwa Ngundari Reribet, Rasa Hidup, Rasa Bungah Susah lan Mulur-Mungkret, Jiwa Kramadangsa, dan manungso tanpo ciri (The knowledge of the soul removes trouble, the sense of life, the sense of joy, sorrow and grief, the soul of manhood, and flawless human beings).
Various efforts to eradicate corruption and the creation of anti-fraud will not be optimal if the habit of applying the meaning of integrity in actions and behavior, which begins with the determination of an employee or public official, is still mixed with the desire to enrich oneself and one's class. On the other hand, employees and public officials are still lacking in self-awareness in order to avoid acts of corruption and fraud. One of the actions that show the decreasing level of integrity and self-awareness of an employee or public official is to engage consciously, and supported from the heart, in the practice of bribery. The practice of giving facilitation payments or bribes is still common. The latest case is the arrest of 22 suspects within the Probolinggo Regency Government in 2021 regarding the selection of positions. The provision of facilitation payments becomes an inherent characteristic and is considered as a wrongful act that is justified so that massive corruption is formed, carried out with a small nominal and little frequency. (Transparency International Indonesia, 2014).
According to Schwab, (2019), the Global Competitiveness Report 2019 places Indonesia's competitiveness at number 50, down 5 places from 141 countries in 2018. This ranking is seen from the aspect of a stable financial system supported by a high increase in technology adoption. However, the quality of getting access to information is still relatively low. The richness of dynamic business culture is not supported by professionalism. The quality and quantity of work intention of workers in Indonesia is included in the low category (Schwab, 2019). From the results of Indonesia's ranking, it can be seen that the integrity of workers in Indonesia needs to be studied more deeply. According to Schlenker et al., (2017), Integrity is a reflection of life's vision to be more positive, increase a sense of spirituality, and minimize irrational actions. The irrational action in question is that the worker takes an action without regard to the applicable written rules and regulations. He emphasizes counter productive actions such as committing fraud and/or corruption, that is realized from the heart.
Counterproductive actions in fraud and/or corruption are reflected in Indonesia's ranking in 2019 which was at number 85 with a value of 40 out of 180 countries (Transparency International, 2019). Meanwhile, Indonesia's ranking in 2020 has decreased by occupying number 102 with a value of 37 out of 180 countries (Transparency International, 2020). Indonesia's ranking in the Corruption Perceptions Index for 2019 & 2020 further creates a large gap in understanding the integrity that is reflected in action. Because if employees uphold the value of integrity, it will result in high quality work and create good governance for the sake of creating an environment that is free from corruption (Rosmi & Syamsir, 2020). Therefore, it is necessary to have a measuring instrument for the level of employee integrity in supporting the achievement of the goals of the company or state institution. With the advancement of technology in various sectors of life, the level of integrity of workers and prospective workers is increasingly needed to be able to produce valid and reliable results (Berry et al., 2007;Hurt & Tomoyasu, 1995;Killinger, 2010;Ones et al., 2003;Schlenker et al., 2017).
To bridge the gap in understanding the meaning of integrity, employees need to focus and strengthen their feelings and desires and do not commit acts of fraud and corruption (Nawawi et al., 2019) Strengthening feelings and desires is part of K. A. Suryomentaram's introspection as an independent human being who has peace in his heart. Introspection will arise a thought that if a person wants pleasure and gives pleasure to his neighbor through dirty means, it is the same as forming a rope knot to tie his own neck and then dragging his neighbor to be entangled together (Nilam, 2008;Sumedi, 2012).
There are very few integrity measuring instruments in Indonesia. The integrity measurement tool studied by Permatasari (2012) uses the Moral Identity Theory (MIT) approach popularized by Blasi (2004). Meanwhile, this study uses Item Response Theory (IRT) which was popularized by Hambleton et al., (1991); Rogers (1961). The Item Response Theory (IRT) approach is believed to be able to improve the measurement of validity, reliability, accuracy and free from sample bound (Berry et al., 2007;Ones et al., 2003;Schlenker et al., 2017). Data that has a high level of validity and reliability can be used as consideration for developing steps to improve integrity (Falani et al., 2020) and introspection in Suryomentaram style (Kholik & Himam, 2015).
According to Hambleton et al., (1991), the Item Response Theory (IRT) approach will produce high validity, reliability, and accuracy if it fulfills the assumptions of unidimensionality and local independence. To fulfill these assumptions, confirmatory factor analysis is used to prove whether the Suryomentaram-style theory of integrity and self-awareness has a valid construct. The items in the Item Response Theory (IRT) use the Rasch Model in order to have invariant sampling items that can be used repeatedly on other respondents (Hambleton et al., 1991). Therefore, the Suryomentaram-style integrity and selfawareness measurement model that will be developed is the Polytomous Item Response Theory (PIRT) (Baker, 2001). Another advantage of Item Response Theory with the Polytomous IRT measurement model is that it is able to detect Different Item Functioning (DIF) (Baker, 2001;Hambleton et al., 1991).
Based on the background of the problem above, the identification of the problem in this study includes 3 (two) things. First, does the construct of measuring instrument of integrity and Suryomentaram-style introspection supports the efforts to prevent fraud and corruption and fulfill unidimensionality by using confirmatory factor analysis? Second, does the measuring instrument of integrity and Suryomentaram-style introspection in this study consists of items that fit in measuring integrity and Suryomentaram-style introspection using the application of the polytomous IRT model? Third, do the items in the measuring instrument of integrity and Suryomentaram-style introspection in this study contain a response bias based on gender that can be detected through differential item functioning (DIF)?.
This research is conducted with the aim of obtaining measuring instrument of integrity and Suryomentaram-style introspection that supports steps to prevent acts of fraud and corruption at the employee level from executives to managers in Indonesian companies and/ or state institutions. The limitation of the research is that in obtaining respondent data, the researchers only use online questionnaires without conducting faceto-face meetings and/or focus group discussions (FGDs).
This research is expected to provide benefits to the development of measuring instrument of integrity and introspection with Suryomentaram style in preventing acts of fraud and corruption in Indonesian companies and/or state institutions. In addition, this research is also expected to provide benefits to Indonesian companies and/or state institutions in finding a minimum standard of the level of integrity and Suryomentaram-style introspecion and in recruiting employees or promoting employees to manager level.

LITERATURE REVIEW AND HYPO-THESIS Integrity
Integrity is a word adapted from Latin which in the Big Indonesian Dictionary (KBBI) has similarities with the state of being in a complete unit, with a high degree of quality. Integrity is formed from honesty, wisdom (Peterson & Seligman, 2004), discipline, obedience and responsibility in action (Herman, 2019). According to Blasi (2004), Integrity is formed from uniformity in thoughts, emotional control, actions and speech that are calm and peaceful in a stable manner over time and conditions. This study uses Item Response Theory (IRT) (Rogers, 1961) with a focus on 4 (four) factors such as upholding honesty, upholding discipline in thinking and behaving, strong self-control, and upholding self-esteem. Peterson & Seligman (2004) provide an explanation of 4 (four) basic factors for the formation of integrity. First, upholding honesty with words and acting according to a conscience that holds fast and is committed to the truth. Second, upholding discipline in thinking and behaving based on determination by carrying out obligations and refusing to bribe and being bribed, rejecting fraudulent actions to meet material needs. Third, strong self-control as evidenced by the individual's ability to speak politely and act without being easily provoked by emotions and not imitating the actions of others in taking advantage of loopholes to commit fraud. Fourth, upholding selfesteem is the belief of every individual in trust in the moral values that are believed and always trying to think positively that his actions are in accordance with moral principles that are far from fraudulent acts (Blasi, 2004).

Suryomentaram-style Introspection
Ki Ageng Suryomentaram-style introspection was originally raised from the moral and social values of Kawruh Jiwa (knowledge of the soul) to the community. The Kawruh Jiwa was recorded in the Kawruh Jiwa Wejanganipun Ki Ageng Suryomentaram I (published in 1989) and the Kawruh Jiwa Wejanganipun Ki Ageng Suryomentaram II (published in 1990). According to Yoshimichi (2006) in (Kholik & Himam, 2015), understanding humans are not always in physical form but also in soul or feeling in thinking and acting as described in the kawruh jiwa of Ki Ageng Suryomentaram.
The first phase of kawruh Jiwa (knowledge of the soul) is ngudhari reribet which focuses on the positive consequences of individual speech and behavior by having a sense of independence in wealth and throne (Pratisti & Prihartanti, 2012). The second phase is rasa hidup, rasa bungah susah dan mulur-mungkret that focuses on a sense of desire to have and always feel lacking and choose wealth and excess physical needs such as luxury food (Pratisti & Prihartanti, 2012;Wahyuningrum, 2017). The third phase is jiwa kradamangsa which is a soul that is filled with greed and is not aware of the work being done but wants excessive material results (Nilam, 2008;Sumedi, 2012). The fourth phase is manungso tanpo ciri which is the last phase after going through the first to third phases by having a conscience and a calm, peaceful mind, always thinking positively, and being aware of his status.

Item Response Theory (IRT)
Item Response Theory (IRT) is a modern approach that has a high level of validity, reliability, and accuracy in assessing the Item Characteristic Function (ICF) (Hambleton et al., 1991;Peterson & Seligman, 2004;Rogers, 1961). Item Characteristic Function (ICF) is expected to further increase respondents' scores higher in producing abilities/traits in the questionnaire questions. Item Response Theory (IRT) must meet the assumption of unidimensionality, namely the factor item that has a dominant score on the respondent. In addition, Item Response Theory (IRT) must meet the assumption of local independence, namely item factors that affect respondents' responses to a series of items that are not related to the formation of the core factor. With the fulfillment of these 2 (two) assumptions, it is hoped that in a series of differential item functioning (DIF) tests, it is possible to know the bias caused by differences in characteristics and functions in certain groups (Baker, 2001;Hambleton et al., 1991).

METHODS Research Methods and Design
This study uses a quantitative method with a Typical Performance Test (TPT) as a measuring tool. The selection of a typical performance test (TPT) as a measuring tool is intended to measure a person's trait/ personality which is interpreted in the form of questions or questionnaires that do not contain wrong or correct answers (Freidenberg, 2011). Answers to questions are given in the form of a Likert scale with 4 (four) types of responses starting from "very appropriate; appropriate; inappropriate and very inappropriate". The choice of the Likert scale is intended so that respondents provide varied responses and avoid the natural tendency of respondents to answer neutrally.

Type and Source of Data
The type of data used is primary data derived from respondents' answers to questionnaires. The assessment on the scoring of the answers to the questionnaire using the Suryomentaram-style integrity and introspection measuring instrument in Table 1.
Sources of data derived from respondents' answers to questionnaires must meet the following criteria: first, the age range is 21 to 56 years; second, having a minimum of 1 year of work experience; third, having a minimum of undergraduate education (Bachelor's Degree / S1): fourth, having faith in religious beliefs; fifth, working on Sumatra Island, Java Island, Bali Island, Kalimantan Island.

Technical Analysis Tool
The measuring tool used to analyze items of integrity and introspection in the style of Suryomentaram in preventing acts of fraud and corruption in Indonesian companies and/or state institutions is confirmatory factor analysis (CFA) assisted by LISREL version 8.7 for data processing. The flow in the analysis process can be seen in Figure  1.

RESULTS AND DISCUSSION Demographics and Research Results
Demographic analysis is intended to provide an overview of the distribution of respondents in the study. Of the 760 questionnaires distributed, 537 questionnaires can be used according to the criteria. The following is the demographic distribution of respondents based on gender, employment status, and position, as shown in Figure 2.
Respondents who have employment status as State Civil Apparatus/ASN with female gender are 106 respondents or 19.74%, while respondents who  Integrity theory carried out by Rogers (1961) and Suryomentaram-style introspection theory carried out by Kholik & Himam (2015); Munif, 2017;Pratisti & Prihartanti (2012) become the basis of researchers in designing measuring instruments in this study. This study conducts several tests such as using the first order and second order CFA methods, person Fit and item Fit as a test tool for Item Response Theory (IRT) and Differential Item Functioning (DIF).
To find out the suitability between the measuring instrument model and the respondent's answer data, statistical testing is needed using the Chi-Square test and the Root Mean Square Error of Approximation (RMSEA). The use of the Chi-Square test is intended to determine the poor "fit" of a measuring instrument model. The Chi-Square probability is expected to be insignificant (p > 0.05) in order to make the model "fit" with the data. RMSEA produces a value of < 0.10 indicating the model "fit" with the data. RMSEA has the following criteria: first, RMSEA produces a value of 0.08 to 0.10 indicating that the model "fit" is sufficient; second, the RMSEA produces a value < 0.08 indicating that there is a reasonable error originating from the model.
The first step to test the "fit" of an integrity model with 30 items and Suryomentaram-style introspection with 30 items is carried out through first order CFA testing to 100 respondents as shown in Tables 3 and 4. The first order CFA test produces indicators that will be used in the model as a measuring instrument for latent variables. The researchers make a measurement model in this test with 2 (two) latent variables (integrity and Suryomentaram-style introspection) and the measurement indicators are 30 items that are measured directly.
The test results of the first-order CFA measurement model with integrity as a latent variable show that the Chi-Square p-value = 0.02531, or less than 0.05 (P<0.05), which means that the model is not fit. The measurement results with RMSEA value = 0.048, or less than 0.10 (RMSEA < 0.10) means that the model is fit. Meanwhile, the test results of the first order CFA measurement model with Suryomentaram-style introspection as a latent variable show that Chi-Square p-value = 0.02794, or less than 0.05 (P <0.05), which means that the model is not fit. The measurement results with RMSEA value = 0.054, or less than 0.10 (RMSEA < 0.10) means that the model is fit.
Based on the results of the first order CFA measurement, the researchers eliminate several items that have small values, as an indication that these items have a small contribution to the measurement model. Re-testing for measuring the "fit" of an integrity model with 20 items and Suryomentaram-style introspection with 22 items is caaried out through first order CFA testing to 100 other respondents as shown in Tables 5 and 6.
The test results of the first order CFA re-mesurement model with integrity as a latent variable show that the Chi-Square p-value = 0.42, or more than 0.05 (P>0.05), which means that the model is fit. The measurement results with RMSEA value = 0.026, or less than 0.10 (RMSEA < 0.10) means that the model is fit. Meanwhile the test results of the first order CFA remeasurement model with Suryomentaramstyle introspection as the latent variable show that Chi-Square p-value = 0.48, or more than 0.05 (P>0.05), which means that the model is fit. The measurement results with RMSEA value = 0.034, or less than 0.10 (RMSEA < 0.10) means that the model is fit.
The assumption of unidimensionality as one of the requirements for the Item Response Theory (IRT) approach to be implemented has been fulfilled by a measuring instrument with the latent variables of integrity and Suryomentaramstyle introspection. Table 7 in Appendix 1 shows the 20 items for the latent variable of integrity. Meanwhile, 22 items for the latent variable of Suryomentaram-style  Table 8 in Appendix 1. The CFA measurement also produces information for each item with a fit or unfit index value. The results of the index per first order CFA measurement item for the integrity measuring instrument can be seen in Table 9 in Appendix 2, while the Suryomentaram-style introspection measurement instrument can be seen in Table 10 in Appendix 3. Tables 9 & 10 show the entire index per item t-Value which is greater than 1.96 (t-value > 1.96).
The second step to test the "fit" of the integrity model with 20 items and Suryomentaram-style introspection with 22 items is carried out through a second order CFA test on 537 respondents. The results can be seen in Tables 11 and 12. The assumption of unidimensionality as one of the requirements for the Item Response Theory (IRT) approach to be implemented has been fulfilled by a measuring instrument with latent variables of integrity and Suryomentaramstyle introspection based on the results of the second order CFA test. The test results of the second order CFA measurement model with integrity as a latent variable show that the Chi-Square p-value = 0.41 or more than 0.05 (P>0.05), which means that the model is fit. Measurement with RMSEA value = 0.024 or less than 0.10 (RMSEA < 0.10) means that the model is fit. Meanwhile, the test results of the second order CFA re-measurement model with Suryomentaram-style introspection as the latent variable show that the Chi-Square p-value = 0.46 or more than 0.05 (P>0.05), which means that the model is fit.
Measurement with RMSEA value = 0.031 or less than 0.10 (RMSEA < 0.10) means that the model is fit. Based on the results of p-value and RMSEA value, the latent variables of integrity and Suryomentaramstyle introspection have been successfully fulfilled, so it can be concluded that these models with the data presented have resulted in a truly fit value.
The results of index per second order CFA measurement item for the integrity measuring instrument can be seen in Table 13 in Appendix 4, while the Suryomentaram-style introspection measuring instrument can be seen in Table  14 in Appendix 5. To find out how well the criteria for the measurement model produce test results that can be explained, it is necessary to re-test the infit mean square value. Infit mean square is a measurement instrument used in the polytomous IRT model that measures Pearson's fit and item fit. According to Adams, R.J., & Khoo (1994), infit mean square test serves to find out the compatibility between the model and the data, which is different from the t-value test which functions to find out how sensitive the item is to the number of samples. The infit mean square testing tool model that produces values in the range of 0.77 and 1.30 is said to be "Fit" (Adams, R.J., & Khoo, 1994). The fit mean square test is carried out using the QUEST application. Table 15 in Appendix 6 shows the results of the integrity measuring instrument, while Table 16 in Appendix 7 shows that The last test is to find out whether there is a tendency for similar answers between gender groups. Response bias towards male and female gender can occur in answering items that measure integrity (Ones, 1993) and introspection (Ones et al., 2003;Pratisti & Prihartanti, 2012). The test is carried out using a Differential Item Functioning (DIF) measuring instrument assisted by the QUEST application. According to Adams, R.J., & Khoo (1994), testing using DIF is carried out to detect the possibility of a response bias contained in items that support the measurement instrument of integrity and Suryomentaram-style introspection. Bias can be detected if the p-value is less than 0.05 (p-value <0.05) (Adams, R.J., & Khoo, 1994). Table 17 in Appendix 8 shows the results of the integrity measuring instrument and Table 18 in Appendix  9 shows the Suryomentaram-style introspection measurement instrument in analyzing the possibility of the emergence of DIF response bias in the answers to items on the measuring instrument. The results of the DIF test show that the items on the measuring instruments of integrity and Suryomentaram-style introspection, from the beginning, do not contain DIF from the gender group. Therefore, this measuring instrument can be used to measure the level of integrity and Suryomentaramstuly introspection in the male and female genders.

CONCLUSION
Based on the results of the first order CFA test on 30 items of integrity measuring instrument and 30 items of Suryomentaram-style introspection measuring instrument, the Chi-Square value indicates not fit, while the RMSEA value indicates fit. This is because the Chi-Square has more sensitivity to the number of sample respondents and the normality of the data. Therefore, the researchers retest the first order CFA by eliminating items that have a small value as an indication that these items have a small contribution to the measurement model. So, there are only 20 items of integrity measuring instrument and 22 items of Suryomentaram-style introspection measuring instrument for re-testing the first order CFA, where the results indicate fit. The results of second order CFA test on 20 items of integrity measuring instrument and 22 items of Suryomentaram-style introspection measuring instrument indicate fit. The results of the first order test and 4 (four) second order factors CFA integrity and 4 (four) second order factors CFA Suryomentaram-style introspection fit with respondent data and data normality. The results of the first order CFA test and second order CFA test show that the items in integrity and Suryomentaram-style introspection measurement instruments fit and meet the assumption of unidimensionality.