<<

Effects of Accountability, Knowledge and Ethics On the Quality of Auditor's Work in KAP Jakarta Selatan.

Khairul Saleh L Tobing [email protected] Universitas Nasional, Jalan Sawo Manila, Pejaten, Pasar Minggu, Jakarta Selatan

Abstract

This study aims to determine: 1) Effect of Auditor Accountability on Quality of KAP Audit Work Results in South Jakarta; 2) The Effect of Auditor Knowledge on the Quality of South Jakarta KAP Audit Audit Results; 3) Effect of Auditor Ethics on the Quality of South Jakarta KAP Audit Work Results.

This study intends to explain the causal relationship between the variables through testing the formulated hypothesis. The population in this study is the auditor who works at the Public Accountant Office in South Jakarta. The population used in this study was 122 auditors who worked at 11 Public Accounting Firms (KAP) in South Jakarta. The technique uses nonprobbility sampling with purposive sampling. To measure the number of samples used the slovin formula with a significant level of 5%. The technique was carried out using the method, the researcher distributed to 127 respondents and only 125 returned questionnaires and used 122 questionnaires according to the measurement results of the Slovin formula. The questionnaire was tested for validity and reliability before the research data was collected, the test results showed that all instruments were valid and reliable. Prior to the analysis, the analysis prerequisite tests are conducted including the normality test, linearity test, multicollinearity test, and test. Analysis of the data used to test the hypothesis is by using multiple techniques, t test and F test. The results of the study show that: 1) There is a positive and significant influence of Accountability, Knowledge and Ethics on the Quality of Auditor's Work Results in KAP South Jakarta.

Keywords: Accountability, Knowledge, Auditor Ethics, and Audit Quality.

I. INTRODUCTION

Auditor services for financial statements are the best known services and these services are services that are often used by external parties who want to assess the company and make decisions on the company. The auditor is able to provide guarantees that the financial statements are relevant and reliable. In general auditing is a systematic process of evaluating evidence objectively about statements about a company's economic activities and events, with the aim of determining the level of appropriateness between statements and established criteria. From the audit results the auditor draws and conveys the conclusions of the financial statements to interested parties. The company's external parties lay the basis of its decision on the work of the auditor, and the Auditor draws conclusions based on the audit work that has been done. This can be interpreted that whether or not the quality of the work of auditors will influence the decisions to be taken by external parties. Logically the quality of the work of the Auditor can be influenced by accountability, that is a social psychological impulse owned by an Auditor in completing his obligations which will be accountable to his environment. Many social psychology studies have proven that there is a relationship and the effect of one's accountability on the quality of his work, because the accountability of an auditor can improve the auditor's cognitive process in making decisions. The of accountability with knowledge also determines the quality of the work of the auditors, and the Audit must also be carried out by someone who has sufficient technical expertise and knowledge in the audit field. This knowledge directly affects the ability of auditors, and this knowledge is obtained from formal education or courses, seminars and workshops related to Accounting. From some of the discussion above, the author would like to examine matters relating to the quality of the work of auditors who are influenced by the auditor's accountability, knowledge and ethics, using auditor respondents who work at KAP in the South Jakarta area who are willing to be respondents in this study. Researchers want to prove whether the results of this study will be the same or different from previous researchers, if the research is carried out in different locations and work environments (different KAP). Different mindsets and perspectives or the way auditors perform their duties will bring different understanding in the quality of their audit work. Based on this the authors are motivated to conduct this study with the title "The Effect of Auditor Accountability, Knowledge and Ethics on the Quality of Auditor's Empirical Study Results in KAP South Jakarta"

B. Problem Formulation Does the Auditor's accountability, Knowledge and Ethics have an influence on the quality of the auditor's work C. Purpose and Use of Research This study aims to determine the effect of accountability on the quality of the work of auditors, determine the effect of knowledge on the quality of the work of auditors, and to determine the effect of ethics has on the quality of the work of auditors. Research is expected to increase knowledge for researchers in the field of accounting, especially in the field of auditing and about the effect of accountability, knowledge and ethics on the quality of the work of auditors. In addition, this research is expected to also be able to trigger better research on auditor quality in the future

Effect of Accountability on the Quality of Auditor's Work Audit quality is influenced by a sense of accountability (accountability) owned by the auditor in completing audit work. Accountability is a social psychological impulse that a person has to fulfill obligations and convey responsibility for answering and explaining the performance and actions of an organization to those who have the right or authority to request information or responsibility. Accountability can improve the quality of work results if it is supported by knowledge and ability in solving problems encountered. There are three indicators that can be used to measure individual accountability. First, how much is their motivation to complete the work. Second, how confident are they that their work will be examined by superiors. Third, how much effort (power of thought) is given to complete a job.

2. Effect of Knowledge on the Quality of Auditor's Work The amount of effort expended by the auditor in completing a job varies according to the level of knowledge possessed. And the level of knowledge of a person can improve the quality of his work. Knowledge can affect the relationship of accountability with the quality of the work of the auditor if the complexity of the work to be faced is of medium / medium level. 2017 Public Accountant Professional Standards (SPAP) on general standards, explains that in conducting an audit, auditors must have sufficient expertise and knowledge structure . Based on this explanation it can be concluded that audit knowledge in carrying out a work will influence the auditor in selecting errors and detecting risks that arise during the auditing process. The results obtained will be able to influence the decisions taken.

3. Effect of Ethics on the Quality of Auditor's Work Ethics is a science of good and bad judgments about the rights and obligations of morality. Professionals in professional ethics imply a pride, commitment to quality, dedication to the interests of clients and a sincere desire to help problems so that the profession can become a trust for the community. In research conducted by Futri and Juliarsa (2014) shows that professional ethics have a positive effect on audit quality. By upholding professional ethics, it is expected that fraud will not occur among the auditors, so that they can provide audit opinions in accordance with the financial statements presented. Furthermore the research of Kurnia et al. (2014) supported by Rahayu and Suryono (2016) research shows that auditor ethics has a positive effect on audit quality, which that the higher the ethics owned by the auditor, the more audit quality will be produced

Analysis Framework The analytical framework explains the influence and relationship between several variables studied.

Figure 2.1 Analysis Framework Chart

Accountability (X1)

H2 Kualitas Hasil Kerja (Y)

Knowledge H1 (X2)

H3

Ethics (X3) Hypothesis :

Based on the analysis framework, the hypotheses in the study are: H1: Accountability has an influence on the quality of the auditor's work. H2: Knowledge has an influence on the quality of the auditor's work. H3: Ethics has an influence on the quality of the auditor's work.

II. RESEARCH METHODS The object of this research is the accountability, knowledge, ethics of auditors and their influence on the quality of the work of auditors implemented at the Public Accounting Firm (KAP) located in the South Jakarta area. Sources of data used are data taken directly from the informants namely the auditors of the Public Accounting Firm (KAP) in South Jakarta. The type of data used is primary data. The population consists of 176 auditors in 11 public accounting firms located in the region: The sampling technique uses a nonprobability sampling technique that is a sampling technique that does not provide the same opportunity or opportunity for each element (member) of the population to be selected as a sample member. Nonprobbility sampling used is Purposive Sampling, a data source sampling technique with certain considerations. In determining the sample size, this study uses the Slovin formula in order to know how many samples will be taken. The formula used is as follows:

1 + () Information: n = Sample Size N = Population Size 1 = Constant e = Expected precision level does not deviate.

The Slovin formula is used with a significant level of 5% of the number of samples used and can be calculated as follows:

176 = 122 1 + 176 (0.05 )

From these calculations, it can be seen that the sample that will be used in this study is 122 auditors.

1. Data Collection Techniques and Tools Data collection techniques used are through field research. Data obtained directly from the first party (primary data) by providing questionnaire to the research subjects, namely the auditors, then the data will be processed using SPSS version 23.00. The measurement scale uses a Likert scale, which is used to measure the attitudes, opinions and perceptions of a person or group of people about social phenomena. The questionnaire was prepared with a 5-level Likert scale consisting of strongly agree, agree, neutral, disagree and strongly disagree.

A. Operational Definition The operational definition of this study contains indicators that enable researchers to collect data relevant to these variables. The independent variables in this study are accountability (X1), knowledge (X2), and ethics (X3), and the dependent variable is the quality of the auditor's work (Y).

Variable Operations Variabel Definisi Variabel Indikator Pengukuran Accountability Accountability or circumstances to - Effort Likert (X 1) be held accountable or - confidence circumstances to be held accountable to parties who have rights

Knowledge Mental components that result - Level of work (X 2) from any process, whether born experience from innate or achieved through - education Likert experience . Ethics A set of rules or norms or - Education (X 3) guidelines governing human - Environment behavior, both those that must be done and those that must be Likert abandoned that are adhered to by a group or group of people or society or profession. Quality of Auditor's The number of correct responses - Audit quality Work given by someone in completing a - Problem solving job compared to the standard work knowledge Likert results or predetermined criteria.

B. Method of Analysis and Testing Hypotheses Descriptive are used to provide a description or description of data seen from the (standard), , , maximum, minimum, sum, , and or skewed distribution.

1. Data Quality Test Validity test is used to measure the validity or validity of a questionnaire. A questionnaire is said to be valid if the questions on the questionnaire are able to reveal something that will be measured by the questionnaire. The technique used is the Pearson Correlation Technique by correlating the score of each item with its total score. The criteria in testing validity are as follows:

1) If r count is positive and r count> r table, then the question item is valid. 2) If the r count is negative and r count

Reliability measurement is done in two ways, namely:

1) Repeated Measure: here a person will be asked the same question at different times, and then see if he remains consistent with the answer. 2) One Shot or measurement only: measurement is only done once and then the results are compared with other questions or measure the correlation between answers to questions. SPSS provides facilities to measure reliability with Cronbach Alpha (a) statistical tests. A construct or variable is said to be reliable if it gives a Cronbach Alpha value> 0.70 (Ghozali, 2016: 47-48).

2. Classical Assumption Test To test the classical assumptions the researcher used the normality test, the multicollinearity test, the test for heteroscedititisity, the test and the multiple analysis.

a. Normality Test, aims to test whether in the regression model, or residual variables have a . The tool used to test normality is to use the (P-P plot). Normality can be detected by looking at the spread of data (points) on the diagonal axis of the graph. b. Multicollinearity Test Multicollinearity test aims to test whether the regression model found a correlation between independent variables (independent). To detect the presence of multicollinearity can be seen from the value of tolerance or the value of Variance Inflaction Factor (VIF). If the tolerance value> 0.10, it means that there will be no multicollinearity. Conversely, if the tolerance value <0.10, it means that multicolonierity occurs. Whereas decision making when looking at the VIF value is if the VIF value <10 then it means that there is no multicollinearity. Conversely, if the VIF value> 10 then it means that there is multicollinearity.

c. Heteroscedasticity Test Heteroscedasticity test aims to test whether in the regression model there is an inequality of variance from the residuals of one observation to another. If the residual variance from one observation to another is fixed, then it is called and if different is called heterokedasticity. Detection of the presence or absence of heteroscedasticity can be seen from the presence or absence of certain patterns on scatterplot charts. If there is a certain pattern, it indicates that heteroscedasticity has occurred. But if there is no clear pattern and the points spread above and below the number 0 on the Y axis, then there is no heteroscedasticity (Ghozali, 2016: 134) d. Autocorrelation Test The autocorrelation test aims to test whether in the linear regression model there is a correlation between the disturbance error (disturbance term-ed.) In the t period and the disturbance error in the previous period (t-1). The autocorrelation test can be done with Durbin-Watson. Decisions about the presence or absence of autocorrelation are:

1. If dU (4-dL), then a negative autocorrelation occurs. 4. If (4-dU)

e. Multiple Linear Regression Analysis Multiple linear regression analysis is a linear relationship between two or more independent variables (X1, X2, X3 ... Xn) with the independent variable (Y). This analysis is to determine the direction of the relationship between the independent variable and the dependent variable, whether each independent variable is positively or negatively related and to predict the value of the dependent variable if the value of the independent variable has increased or decreased. As per the hypothesis tested, the multiple linear regression equation is

Y = α + β1X1 + β2X2 + β3X3 + β4X4 + e

Information:

Y = Quality of Auditor's Work α = Constant 1 = Regression coefficient  X1 = Auditor Accountability 2 = Regression coefficient  X2 = Auditor Knowledge 3 = Regression coefficient  X3 = Auditor Ethics 4 = Regression coefficient  X4 = Accountability, Knowledge, Ethics e = error / term interference

3. Hypothesis test The results of the coefficient of determination are analyzed to measure how far the model's ability to explain the variation of the dependent variable. The coefficient of determination between zero and one, the value of the coefficient of determination is small (near zero) means the ability of independent variables in explaining the variation of the dependent variable is very limited. While the low coefficient of determination (close to one) means that the independent variables provide almost all the information needed to predict variations in the dependent variable. In knowing the value of the coefficient of determination in this study is to see the value of R square in the model summary generated by the SPSS program.

a. Model Feasibility Test (F Test) F test is performed to prove whether the independent variables simultaneously have an influence on the dependent variable. F test is done with the aim to test the entire independent variable on one dependent variable freely with a significance of 0.05 can be concluded. If a significant value <0.05 then Ha is accepted and H0 is rejected, this means that all dependent or free variables have an influence together on the dependent or dependent variable. Whereas if a significant value> 0.05 then Ha is rejected and H0 is accepted, this means that all independent or independent variables do not have a joint influence on the dependent or dependent variable.

b. T test T-test is used to test whether each independent variable significantly influences the dependent variable. This hypothesis test is done by comparing the value of t arithmetic with t table obtained based on a significant level of 5% or 0.05. The basis for decision making is as follows:

1) H1 is accepted if probability <0.05 2) H2 is accepted if probability <0.05 3) H3 is accepted if the probability is <0.05 4) H4 is accepted if the probability is <0.05

VI. RESEARCH RESULTS AND DISCUSSION

A. Research Results Questionnaire Return Rate No Informastion Jumlah Kuesioner Presentase 1 Questionnaire distributed 127 104% 2 Returned questionnaire 125 102% 3 Questionnaire processed 122 100% Source: primary data, processed by the author

1. Analysis of Characteristics of Respondents Characteristics of respondents that have been observed in the form of gender, age, level of last education, position, length of work, length of time occupying the position of respondent. The distribution results in frequency are presented in the following: a. Gender Based on table 4.3 it can be seen that of 122 respondents there were 96 respondents who were male or 79% while 26 respondents were female or 21%. So it can be concluded that the majority of respondents are male. b. Age Based on table 4.4 it can be seen that of 122 respondents 82 aged <30 years or 67%, 30 respondents aged 31-40 years or 25%, and 10 respondents aged 41-50 years or 8%, based on the questionnaire distributed were not there are respondents aged> 51 years. c. Last education Based on Table 4.5 it can be seen that of 122 respondents there were 102 respondents who had the last education S1 or 84%, while 20 respondents had the last S2 education or as much as 16%. So the majority of respondents have the last education S1. d. Long time working Based on Table 4.6 it can be seen from 122 respondents there were 78 respondents who had worked less than 5 years or 64%, 25 respondents who worked for 5-10 years or 21%, 13 respondents who worked for 10-20 years or 11% , while 6 respondents worked less than 20 years or 4% e. Position Based on Table 4.7, it can be seen from 122 respondents that there were 82 respondents who served as junior auditors or 67%, 28 respondents served as senior auditors or 23%, and 6 respondents who served as managers and supervaisors or 5%. So the majority of respondents served as junior auditors.

B. Discussion 1. Test Results Descriptive statistics are used to provide a description or description of data seen from the mean (standard), standard deviation, variance, maximum, minimum, sum, range, kurtosis and skewness (skewed distribution). The test results can be seen in table 4.8. Descriptive statistics Descriptive Statistics Std. Minimum Maximum Sum Mean Deviation ACCOUNTABILITY 122 17 25 2741 22.47 1.846 KNOWLEDGE 122 15 23 2411 19.76 1.859 ETHICS 122 16 25 2601 21.32 1.904 QUALITY OF WORK 122 18 25 2688 22.03 1.686 RESULTS Valid N (listwise) 122 Source: ouput processed by the author, using SPSS v.23 Based on the results of table 4.8 above, it can be seen that: 1) Accountability Variable (X1) has a total data (n) of 122 respondents, showing that the lowest value (minimum) is 17 and the highest value (maximum) is 25, the value (sum) is 2741 and the mean value is 22.47, while for the standard deviation of accountability is 1.846. 2) Knowledge variable (X2) has data (n) of 122 respondents, showing that the lowest (minimum) value is 15 and the highest value (maximum) is 23, the value (sum) is 2411, the average value (mean) is 19 , 76, while for the Knowledge deviation standard of 1.859. 3) Ethical Variable (X3) has a total data (n) of 122 respondents, showing that the lowest (minimum) value is 16 and the highest value (maximum) is 25, the sum (sum) is 2601 and the mean value is 21.32, while for the Ethics standard deviation of 1.904. 4) Variable Quality of Work (Y) has a total data (n) of 122 respondents, showing that the lowest (minimum) value is 18 and the highest value (maximum) is 25, the sum (sum) is 2688 and the average value (mean ) amounting to 22.03, while for the standard deviation of Accountability is 1.686.

2. Data Quality Test Results

a. Validity Test Results Validity test is used to measure the validity or validity of a questionnaire. A questionnaire is said to be valid if the questions on the questionnaire are able to reveal something that will be measured by the questionnaire. One way to test the validity developed by comparing the value of r arithmetic with r tables. Based on the table r (corrected item-total correlation) and in this study using a significant level of 5% or 0.05. If N is 122, then rtable = 0.177. Thus, if r count> r table, then the statement can be said to be valid. The validity test results can be seen in table 4.8 as follows

Accountability Validity Test Results Correlations AK2 AK3 AK4 AK TOTAL_AK AK1 Pearson 1 .213 * .329 ** .221 * .208 * .592 ** Correlation Sig. (2 -tailed) .018 .000 .014 .022 .000 122 122 122 122 122 122 AK2 Pearson .213 * 1 .142 .275 ** .250 ** .604 ** Correlation Sig. (2- .018 .119 .002 .005 .000 tailed) 122 122 122 122 122 122 AK3 Pearson .329 ** .142 1 .413 ** .267 ** .645 ** Correlation Sig. (2- .000 .119 .000 .003 .000 tailed) 122 122 122 122 122 122 AK4 Pearson .221 * .275 ** .413 ** 1 .240 ** .740 ** Correlation Sig. (2- .014 .002 .000 .008 .000 tailed) N 122 122 122 122 122 122 AK Pearson .208 * .250 ** .267 ** .240 ** 1 .582 ** Correlation Sig. (2- .022 .005 .003 .008 .000 tailed) N 122 122 122 122 122 122 TOTA Pearson .592 ** .604 ** .645 ** .740 ** .582 ** 1 L_AK Correlation Sig. (2- .000 .000 .000 .000 .000 tailed) 122 122 122 122 122 122 Source: primary data, processed by the author Hasil Uji Validitas Pengetahuan Correlations Correlations TOTAL_ PENG1 PENG2 PENG3 PENG4 PENG5 PENG AK1 Pearson 1 .320 ** .127 .386 ** .306 ** .648 ** Correlation Sig. (2- .000 .164 .000 .001 .000 tailed) N 122 122 122 122 122 122 PENG2 Pearson .320 ** 1 .142 .472 ** .405 ** .741 ** Correlation Sig. (2-tailed) .000 .118 .000 .000 .000 N 122 122 122 122 122 122 PENG3 Pearson .127 .142 1 .032 .090 .435 ** Correlation Sig. (2-tailed) .164 .118 .730 .322 .000 N 122 122 122 122 122 122 PENG4 Pearson .386 ** .472 ** .032 1 .314 ** .722 ** Correlation Sig. (2-tailed) .000 .000 .730 .000 .000 N 122 122 122 122 122 122 PENG5 Pearson .306 ** .405 ** .090 .314 ** 1 .642 ** Correlation Sig. (2-tailed) .001 .000 .322 .000 .000 N 122 122 122 122 122 122 TOTAL_ Pearson .648 ** .741 ** .435 ** .722 ** .642 ** 1 PENG Correlation Sig. (2-tailed) .000 .000 .000 .000 .000 122 122 122 122 122 122 Sumber : Source: primary data, processed by the author

Hasil Uji Validitas Etika Correlations ETIK1 ETIK2 ETIK3 ETIK4 ETIK5 TOTAL_ETIK ETIK1Pearson Correlation 1 .232 * .286 ** .299 ** .295 ** .646 **

Sig. (2-tailed) .010 .001 .001 .001 .000 N 122 122 122 122 122 122 ETIK2 Pearson .232 * 1 .318 ** .295 ** .369 ** .677 ** Crelation Sig. (2-tailed) .010 .000 .001 .000 .000 122 122 122 122 122 122 ETIK3 Pearson .286 ** .318 ** 1 .352 ** .181 * .675 ** Correlation Sig. (2-tailed) .001 .000 .000 .046 .000 122 122 122 122 122 122 ETIK4 Pearson .299 ** .295 ** .352 ** 1 .308 ** .663 ** Correlation Sig. (2-tailed) .001 .001 .000 .001 .000 122 122 122 122 122 122 ETIK5 Pearson .295 ** .369 ** .181 * .308 ** 1 .632 ** Correlation Sig. (2-tailed) .001 .000 .046 .001 .000 122 122 122 122 122 122 TOTAL_ Pearson .646 ** .677 ** .675 ** .663 ** .632 ** 1 ETI Correlation Sig. (2-tailed) .000 .000 .000 .000 .000 122 122 122 122 122 122 Sumber : Source: primary data, processed by the author

Hasil Uji Validitas Kualitas Hasil Kerja Aditor Correlations KHK1 KHKK2 KHK3 KHK4 KHK5 TOTAL_KHK KHK1 Pearson 1 .354 ** .265 ** .176 .174 .605 ** Correlation Sig. (2-tailed) .000 .003 .053 .055 .000

122 122 122 122 122 122

KHKK2 Pearson .354 ** 1 .200 * .206 * .247 ** .615 ** Correlation Sig. (2-tailed) .000 .028 .023 .006 .000 122 122 122 122 122 122 KHK3 Pearson .265 ** .200 * 1 .271 ** .254 ** .655 ** Correlation Sig. (2- .003 .028 .003 .005 .000 tailed) 122 122 122 122 122 122 KHK4 Pearson .176 .206 * .271 ** 1 .308 ** .634 ** Correlation Sig. (2- .053 .023 .003 .001 .000 tailed) 122 122 122 122 122 122 KHK5 Pearson .174 .247 ** .254 ** .308 ** 1 .636 ** Correlation Sig. (2- .055 .006 .005 .001 .000 tailed) 122 122 122 122 122 122 TOTAL_ Pearson .605 ** .615 ** .655 ** .634 ** .636 ** 1 KHK Correlation Sig. (2- .000 .000 .000 .000 .000 tailed) 122 122 122 122 122 122 Sumber : Source: primary data, processed by the author

The results of the validity test in table 4.9 above state that each item of each variable from each dependent variable is the Quality of Work of the Auditor (Y) and the independent variables namely Accountability (X1), Knowledge (X2), and Ethics (X3) have a corrected value item-total or rcount is greater than rtable. Thus it can be concluded that all statement items used in this study are valid a. Reliability Test Results

The reliability test is used to measure a questionnaire which is an indicator of a variable or construct. A questionnaire is said to be reliable if it gives a Cronbach Alpha value> 0.60. The reliability test results can be seen in table 4.10 as follows:

Hasil Uji Reliabilitas Reliability Statistics Variabel Cronbach's Alpha N of Items Akuntabilitas .622 5 Pengetahuan .633 5 Etika .671 5 Kualitas Hasil Kerja .618 5 Source: ouput processed by the author, using SPSS v.23

The results of the validity test in table 4.10 above show that each item of each variable from each dependent variable is the Quality of Work of the Auditor (Y) and the independent variables namely Accountability (X1), Knowledge (X2), and Ethics (X3) have Cronbach values Alpha is greater than 0.60.

1. Classical Assumption Test Results a. Normality Test Results The normality test aims to test whether in the regression model, confounding or residual variables have a normal distribution. In this study the tool used to test normality is to use the Normal Probability Plot (P-P plot). The retrieval criteria to test whether the data are normally distributed or not are as follows: a) If the data spreads around the diagonal line and follows the diagonal direction, then the regression model fulfills the norm of normality. b) If the data spreads far from the diagonal line or does not follow the diagonal direction, then the regression model does not meet the assumption of normality.

Normality Test Results By looking at the normal P-Plot graph display in Figure 4.1, it can be concluded that the P-Plot graph is seen spreading around a diagonal line and the direction of its spread follows the direction of the diagonal line. This graph shows that the regression model is feasible because it meets the assumption of normality. a. Multicollinearity Test Results Multicollinearity test aims to test whether the regression model found a correlation between independent variables (independent). To detect the presence of multicollinearity can be seen from the value of tolerance or the value of Variance Inflaction Factor (VIF). If the tolerance value> 0.10, it means that there will be no multicollinearity. Conversely, if the tolerance value <0.10, it means that multicolonierity occurs. Whereas decision making if looking from the VIF value is if the VIF value <10 then it means that there is no multicollinearity. Conversely, if the VIF value> 10 then it means that there is multicollinearity

Hasil Uji Multikolonieritas Collinearity Statistics Tolerance VIF

AKUNTABILITAS .955 1.047 PENGETAHUAN .950 1.053 ETIKA .993 1.007 Sumber : ouput diolah penulis, menggunakan SPSS v.23 Based on table 4.11 shows the Accountability variable has a Tolerance value of 0.955 with a VIF value of 1.047. Knowledge variable has a tolerance of 0.950 with a VIF value of 1.053. Whereas the Ethics variable has a Tolerance value of 0.993 with a VIF value of 1.007. Based on the multicollinearity test results, it shows that all independent variables in the regression model have a Tolerance value> 0.10 and a VIF value <10. This can be said that the regression equation model does not occur multicollinearity.

a. Hasil Uji Heteroskedastisitas Heteroscedasticity test aims to test whether in the regression model there is an inequality of variance from the residuals of one observation to another. If the variance from one observation residual to another observation is fixed, then it is called homoscedasticity and if different, it is called heterokedasticity. To detect the presence or absence of heteroscedasticity can be seen by the presence or absence of certain patterns on the Scatterplot graph between the value of the dependent variable (ZPRED) and the residual (SRESID). The decision making criteria for testing heterocedasticity through Scatterplots charts, namely: a) If the Scatterplots nerves show a pattern of points such as points that are bubbled or widened and then narrowed, then it can be concluded that heterokedasticity has occurred. b) If the Scatterplots graph does not form a clear pattern then it can be said that the regression model does not occur heterokedastisitas.

Hasil Uji Heteroskedastisitas

a. Hasil Uji Autokorelasi According to (Ghozali, 2016: 107) autocorrelation test aims to test whether in the linear regression model there is a correlation between disturbance errors (disturbance term-ed.) In the period t and error in the previous period (t-1). Autocorrelation tests can be done with Durbin-Watson. Decisions about the presence or absence of autocorrelation are: 1) If dU (4-dL), then a negative autocorrelation occurs. 4) If (4-dU)

Hasil Uji Autokorelasi Model Summary b Durbin- R Square Adjusted R Square Std. Error of the Estimate Watson 1 .488 a .238 .219 1.490 2.230

Based on the table above shows that the value of Durbin-Watson is 2.230 while from the Durbin- Watson table with a significant level of 0.05 with the amount of data (n) = 122, and k = 3 obtained dU as large as 1.755, thus that the DW value obtained is 2.230 This is between dU of 1.755 and (4-dU) of 4 - 1.755 = 2.245. This shows that the regression model in this study did not occur autocorrelation. c. Multiple Linear Regression Analysis Test Results In this section, accountability, knowledge, ethics and quality of work are estimated using multiple linear regression models. The form of the multiple linear regression equation is as follows:

Y = α + β 1X1 + β 2X2 + β3X3 + e Information: Y = Quality of Auditor's Work α = Constant 1 = Regression coefficient  X1 = Auditor Accountability 2 = Regression coefficient  X2 = Auditor Knowledge 3 = Regression coefficient  X3 = Auditor Ethics e = error / term interference

Multiple Linear Regression Analysis Test Results

Coefficients a

Unstandardized CoefficientsStandardized Coefficients Std. Error Beta Sig. (Constant) 7.675 2.426 3.164 .002 AKUNTABILITAS .286 .075 .313 3.812 .000 PENGETAHUAN .225 .075 .248 3.012 .003 ETIKA .163 .071 .184 2.285 .024 Source: ouput processed by the author, using SPSS v.23

Based on the data processing in table 4.13 we get the multiple linear regression equation model as follows:

Y = 7,675 + 0,286X 1 + 0,225X 2 + 0,163X 3 The information from the multiple linear regression equation above is: a) The constant value (α) is 7,675 which means that the accountability, knowledge and ethics variables get a cash value (fixed) then the auditor's ability in the quality of work results is 7,675. b) The coefficient regression value of the accountability variable (X1) is 0.286 and is positive or unidirectional which indicates the higher the accountability the auditor has, the better the quality of the work of the auditor. c) The regression coefficient value of the knowledge variable (X2) is 0.225 and is positive or unidirectional which indicates the higher the knowledge the auditor has, the better the quality of the work of the auditor. d) Ethical variable regression coefficient value (X3) of 0.163 and positive or unidirectional value indicating the higher the ethics of the auditor, the better the quality of the work of the auditor.

4. Hypothesis Testing a. Coefficient of Determination (R2) The results of the coefficient of determination are analyzed to measure how far the model's ability to explain the variation of the dependent variable. The coefficient of determination between zero and one, the value of the coefficient of determination is small (near zero) means the ability of independent variables in explaining the variation of the dependent variable is very limited. While the low coefficient of determination (close to one) means that the independent variables provide almost all the information needed to predict variations in the dependent variable.

Results Coefficient of Determination (R2)

Model Summary b

Model R SquareAdjusted R SquareStd. Error of the Estimate 1 .488 a .238 .219 1.490

source: ouput processed by the author, using SPSS v.23

Based on table 4.14 the coefficient of determination test results can be seen that the value of R Square in this study amounted to 0.238, so it can be concluded that Accountability (X1), Knowledge (X2), and Ethics (X3) affect the Quality of Auditor's work (Y) of 23.8 %, while the remaining 0.762 or 76.2% is influenced by other variables not included in this study b. Model Feasibility Test (F Test) The statistical test f basically shows whether all the independent or independent variables included in the model have a joint influence on the dependent or dependent variable. To do the F-Test, you can compare the calculated F value with the F value according to the table. In this study, the numbers used are figures from table F with n = 122 and α = 5%. If the value of Fcount> Ftable, then H0 is rejected and Ha is accepted

Tabel 4.15 Hasil Uji Kelayakan Model (Uji-F)

ANOVA a Model Sum of Squares Mean Square Regression 81.895 3 27.298 12.296 .000 b Residual 261.973 118 2.220

Total 343.869 121

source: ouput processed by the author, using SPSS v.23

Based on table 4.15 the results of the model feasibility test (Test F) are known for a significant value of 0,000 <0.005 and for a value of Fcount> F table of 12.296> 2.68, it can be said that the regression coefficient is significant and the tested model is feasible to proceed.

Test Results t The t-test basically shows how far the independent or individual explanatory variables are in explaining the variation of the dependent variable. How to do the t-test can compare the t statistical value with a critical point according to the table. In this study using n = 122 and α = 5%. If the value of tcount> ttable states that an independent variable individually affects the dependent variable

Hasil Uji Hipotesis ( Uji-T)

Coefficients a

Unstandardized CoefficientsStandardized Coefficients Model Std. Error Beta Sig. (Constant) 7.675 2.426 3.164 .002 AKUNTABILITAS .286 .075 .313 3.812 .000 PENGETAHUAN .225 .075 .248 3.012 .003

ETIKA .163 .071 .184 2.285 .024

source: ouput processed by the author, using SPSS v.23

Hypothesis 1 H1: Accountability has a significant effect on the quality of the auditor's work. Based on the SPSS calculation, it can be seen that the significant value of 0,000 <0.05 with tcount for the regression coefficient X1 is 3.812 with the ttable for t with n = 122, k = 3 and e = 0.05 is 1.980. So from the results of the comparison of the significant level of 0,000 <0.05 and the ratio between tcount and ttable is 3,812> 1,980. So it can be concluded that H1 is accepted, namely accountability has a significant effect on the quality of the work of the auditor. The results of this study are consistent with research conducted by Diani Mardisar and Ria Nelly Sari (2007) which states that Accountability has a significant effect on the Quality of Auditor's Work.

Hypothesis 2 H2: Knowledge has a significant effect on the quality of the auditor's work. Based on SPSS calculations, it can be seen that the significant value of 0.003 <0.05 with the tcount for the regression coefficient X2 is 3.012 with the ttable for t with n = 122, k = 3 and e = 0.05 is 1.980 ,. So from the results of the comparison of the significant level of 0.003 <0.05 and the ratio between tcount and ttable is 3, 012> 1.980. So it can be concluded that H2 is accepted, namely knowledge has a significant effect on the quality of the work of the auditor. The results of this study are consistent with research conducted by Hermina Sihombing (2012) which states that knowledge has a significant effect on the Quality of Auditor's Work.

Hypothesis 3 H3: Ethics has a significant effect on the quality of the auditor's work. Based on SPSS calculations, it can be seen that the significant value of 0.024 <0.05 with the tcount for the regression coefficient X3 is 2.285 with the ttable for t with n = 122, k = 3 and e = 0.05 is 1,980 ,. Then the results of the comparison of the significant level of 0.024 <0.05 and the ratio between tcount and ttable is 2.285> 1.980. So it can be concluded that H3 is accepted, that is ethics has a significant effect on the quality of the work of auditors. The results of this study are consistent with research conducted by Desi Wahyu Lestari (2017) which states that Auditor Ethics influences Audit Quality

5. Summary of Research Results The object of research in this study is the accountability, knowledge and ethics of auditors on the quality of the work of auditors at the South Jakarta Public Accountant Office. The sampling technique of this study used a nonprobability sampling technique, which is a sampling technique that does not provide equal opportunity or opportunity for each element (member) of the population to be selected as a sample member. Nonprobbility sampling used is Purposive Sampling, a data source sampling technique with certain considerations. The sample in this study amounted to 122 respondents in the South Jakarta Public Accountant Office. Based on the results of the classic assumption test carried out through several stages of testing, the results can be explained as follows:

1. The normality test results show that the P-plot graph is seen spreading around the diagonal line and the direction of its spread follows the direction of the diagonal line. The graph shows that the regression model is feasible because it meets the assumption of normality. 2. Multicollinearity test results show that all independent variables used in the regression equation model do not occur multicollinearity. This can be seen from the Tolerance value> 0.10 and VIT value <10. 3. Heteroscedasticity test results show that the Scatterplot graph shows scattered points with unclear patterns above and below the number 0 on the Y axis, then the regression model does not occur heteroscedasticity. 4. The autocorrelation test results show that the DW value obtained for 2.230 is between dU of 1.755 and (4-dU) of 4 - 1.755 = 2.245. This shows that the regression model in this research does not occur autocorrelation.

Based on the results of multiple linear analysis tests, the results of the coefficient of determination (R2), the results of the model feasibility test (F-test) and t-test results can be explained as follows:

1. Multiple linear regression test results The constant value (α) is 7,675 which means that the accountability, knowledge and ethics variables get a cash value (fixed) then the auditor's ability in the quality of work results is 7,675. 2. The coefficient regression value of accountability variable (X1) was 0.286, knowledge (X2) was 0.225 and ethics (X3) was 0.163 and all were positive or unidirectional which showed the higher accountability, knowledge and ethics owned by the auditor, the better the quality of the work of the auditor. Test results for the coefficient of determination. The results of the coefficient of determination test can be seen that the value of R Square in this study amounted to 0.238, so it can be concluded that Accountability (X1), Knowledge (X2), and Ethics (X3) affect the Quality of Auditor's work (Y) of 23.8%. 3. Model feasibility test results (F-Test) The results of the model feasibility test (Test F) are known for a significant value of 0,000 <0.005 and for the value of Fcount> Ftable of 12.296> 2.68, it can be said that the regression coefficient is significant and the model tested is feasible to be continued. 4. T-Test Results a. The accountability variable significantly influences the quality of the auditor's work. Where the significant level is 0,000 <0.05 and the ratio between tcount and ttable is 3.812> 1.980. b. Knowledge variables significantly influence the quality of the auditor's work. Where the significant level is 0.003 <0.05 and the ratio between tcount and ttable is 3, 012> 1.980. Ethical variables significantly influence the quality of the auditor's work. Where the significant level of 0.024 <0.05 and the ratio between tcount and table