Research on the Influence of Audit Interview on Non-Standard Audit Opinions

: The audit interview mechanism implemented by CICPA in 2011 and information disclosure in 2012 aims to enhance the audit quality of auditors and the quality of annual reports of listed companies. This paper empirically studies the impact of audit interviews on non-standard audit opinions. Based on the financial data of listed companies from 2011 to 2020 and the audit interview data from 2012 to 2021, this paper concludes that audit interviews have a significant negative impact on non-standard audit opinions. According to the actual situation, suggestions are made: the audit interview mechanism needs to be improved, and the audit market supply and demand should be balanced.


Introduction
In 1983, accounting firms began to be established nationwide, and began to formulate relevant legal norms for non-standard audit work. In 1988, the "Regulations of the People's Republic of China on Auditing" was issued. Beginning in 1994, the Audit Law of the People's Republic of China, its implementing regulations, and basic criteria have been continuously revised and improved with the progress of socialist development. Up to now, China's audit supervision model has been dominated by administration and supplemented by the industry. The government is responsible for punishing and supervising illegal behaviors in accordance with laws and regulations, and the industry is supervised by the CICPA to supervise misconduct in audit work.
From the perspective of the development of audit interviews, since 1988, the CICPA has responded to the call of the state to continuously strengthen and improve the management and standardization of market audits. The trial system promulgated in 2002 put forward the prototype of "industry selfdiscipline supervision", and the CICPA began to remind certified public accountants of their auditing behavior through the conversation system. However, the effect of this system in the industry is not outstanding and obvious. In 2009, the State Council requested the CICPA to continuously enrich the industry self-discipline supervision means, with the purpose of improving the supervision of auditors and the prevention of violations. In order to improve the strength of industry self-discipline, the "Working Regulations on Auditing and Supervision of Annual Reports of Listed Companies" was issued and implemented in 2011 by the CICPA, the leader of industry self-discipline. The association began to implement the disclosure of interview information. Audit interview means that the person in charge of the audit institution or the chief partner is interviewed by the CICPA in various ways (written interview, telephone interview, etc.) before the end of the annual report audit work. Higher risks will issue reminders and warnings, and audit institutions will also need to submit instructions to the CICPA afterwards. This mechanism aims to improve and ensure the quality of auditors' auditing of listed companies' annual reports.
Based on the above research background, combined with the financial data of listed companies from 2011 to 2020 and the audit interview data from 2012 to 2021, this paper deeply studies the impact of audit interviews on non-standard audit opinions. This paper concludes that audit interviews have a significant negative correlation with non-standard audit opinions. According to the actual situation, suggestions are made: the audit interview mechanism needs to be improved, and the audit market supply and demand should be balanced.
From 2011 to 2020, the ratio of non-standard audit opinions to the total market audit opinions first decreased and then increased. In 2012 and 2013, the number of listed companies whose market financial reports were non-standard audit opinions reached the lowest value. Among the major relevant events experienced by the market during this period, the audit interview initiated by the China CICPA was included. Does the audit interview really affect the non-standard audit opinion in a statistical sense? Based on this question, this paper studies the role of audit interviews on non-standard audit opinions. The CICPA's interview method is different from the government's administrative supervision, and its role in the domestic audit market is still not as good as that of the foreign AICPA and PCAOB's industry self-regulation. The research in this paper will help to strengthen the public's knowledge and understanding of audit interviews, and arouse the public and government agencies' attention and attention to the follow-up of audit interviews.
Combined with existing research on audit interviews, it is different from the industry self-discipline of foreign AICPA and PCAOB. In the literature research of previous scholars in China, some scholars believe that the audit interview is an in-process supervision mechanism, which will improve the prudence of auditors, and the probability of a listed company being issued a qualified opinion will be reduced. Some scholars believe that audit interviews prompt auditors to increase audit fees in order to reduce risk losses and increase audit workload, but the probability of listed companies being issued unqualified opinions has not changed significantly. The theoretical research significance of this paper is to enrich the literature on the impact of audit interviews on the issuance of non-standard audit opinions through the latest audit interview data.

Related concepts
(1) Audit interview The audit supervision interview mechanism is simplified and called "audit interview". The purpose is to warn the person in charge of the accounting firm about certain risks in the audit work, so as to improve the market audit efficiency and audit quality. This mechanism is an "upgraded version" based on the talk mechanism adopted by the CICPA in 2002 and an "upgraded version" of the industry's selfdiscipline supervision at the request of the Ministry of Finance in 2009. In the past, many scholars have different understandings of the definition of audit interview. Huang Yixiong and Li Changai (2016) believe that audit interview is a means of industry self-regulation. However, judging from the time distribution of audit interviews, Wu Xi, Yang Yulong and Zhang Junsheng (2014) believe that audit interviews are preventive supervision measures that are different from post-event supervision before the annual report is disclosed.
(2) Non-standard audit opinions Audit opinions are commonly divided into standard and non-standard audit opinions, or a combination of unqualified and non-unqualified opinions. A non-standard audit opinion includes an unqualified opinion plus a matter paragraph in addition to a non-unqualified opinion. In the sample collected in this article, the unqualified opinion and the explanatory paragraph are a new type of audit opinion in the 2017 annual report. The negative opinion is not published many times in the domestic market, probably because the word negative has a greater impact on the domestic public.

Related theories
(1) Information asymmetry theory Information asymmetry theory is a basic theory that is widely used in various transactions in the audit industry, such as entrusted transactions between shareholders and enterprise managers, loan transactions between creditors and enterprise management, and financing transactions between investors and management. Under the background of inconsistency in the information obtained by both parties involved in the event, the beneficial party who has more information will safeguard its own interests and gain benefits, and the disadvantaged party who has less information does not necessarily mean that it will suffer losses. Therefore, the disadvantaged party can choose whether to cooperate with the advantageous party or not.
(2) Principal-agent theory The information on the status quo of the enterprise is different between the principal who owns the ownership of the enterprise and the agent who has the management right, which leads to the fact that the ownership of the company and the management right are in a discrete state in this theory. Agents who know more information tend to protect their own interests because they are inconsistent with the client's goals. However, clients with less information will also take measures such as increasing contract conditions and distributing shares to reduce the risks brought about by less information.
(3) Risk management theory Risk management theory is applied in the management process of taking measures to minimize future risk losses in a specific risk environment. Risk management is classified according to the development process of events and can be divided into pre-event, in-process and post-event management. The first measure is to prevent losses caused by the occurrence of risks by signing agreements, for example, the amount of loan transactions in the agreement signed between the creditor and the enterprise. The scope of use can avoid the risk of not recovering the principal and interest caused by the company's future investment in high-risk businesses. In-process management is to take measures for the risk loss that has occurred but not ended. When the company's shareholders find that the company is at risk of bankruptcy, they will choose to sell the company's stock to reduce the loss. Post-event management is to take measures for the risk losses that have ended. For example, listed companies punished by the China Securities Regulatory Commission will use announcements and other means to improve the confidence of the public and investors.
(4) Signal transmission theory Signal transmission means that different subjects send different signals to their objects. Because it is an "advanced version" of information asymmetry, the theory is classified and analyzed according to two types of asymmetry theory. In the first type of intra-enterprise transactions of the theory, in order to reduce potential losses due to less information about the enterprise, shareholders will entrust a thirdparty audit institution to review the status of the enterprise. In order to send a favorable signal to shareholders, corporate management will tend to make the profit in the financial report approximate to the expected value of the shareholders.
In the second type of enterprise internal and external transactions of this theory, in order to appease the unfavorable creditors who do not have more information about the enterprise, in addition to fulfilling the contractual obligations on time, the company will also use the solvency and other data calculated from the relevant financial information to convey to the creditors that the company is able to repay the debt. Investors are the unfavorable party to the capital injection of the company, and the more dividends the company pays out, the greater the probability of inferring that it is doing well. However, it is not that companies that pay more dividends are in good business conditions, and that companies that pay less dividends are not necessarily doing well. Therefore, it is also difficult for companies to convey the correct signal of good operation to investors. Wu Weirong and Liu Yawei (2015) studied the inhibitory effect of media supervision on the issuance of non-standard audit opinions. It reflects from the side that listed companies will try their best to avoid auditors issuing unqualified opinions in order to convey positive signals to the public.

Literature review
(1) Literature research on the influencing factors of non-standard audit opinions There is a lot of literature on the factors that influence auditors' decisions to be more inclined towards a non-standard audit opinion. Bao and Chen (1998) comprehensively studied the significant effects of 11 factors, such as asset-liability ratio, operating loss, and equity ratio, on the issuance of non-standard audit opinions from the operating status and ownership structure of listed companies. Blay (2010) studied the effect of the independence of auditors and the litigation risks involved in listed companies on the auditors' decision-making process, and believed that once the auditors found that the legal risks such as pending lawsuits of the audited units were high in the decision-making process, they would choose a non-standard audit opinion to reduce future losses. Liu and Cullinan (2020) studied whether the audit opinion types of auditors' decisions were affected by the debt characteristics of listed companies, and believed that the debt interest rates and long-term debt ratios that the listed companies need to pay both promote non-standard audit opinions.
The reasons why the audit institution's financial report is considered unqualified can be roughly divided into three aspects: the audited entity, the audit institution and the market environment. The operating status of a listed company directly affects the positive and negative net profit in the company's annual report. Lv Minrong (2011) studied the impact of the financial status of listed companies on non-standard audit opinions in 2011 and 2012, and obtained the degree of loss of listed companies. The more severe the conclusion, the greater the impact on the non-standard audit opinion. Fang Junxiong et al. (2004) obtained a stimulating conclusion after studying the asset-liability ratio of listed companies. From the perspective of the shareholding structure of listed companies, Yue Heng (2006) conducted an in-depth study on the proportion of major shareholders in the company's shares, and found that when the major shareholders accounted for more than 10% of the shares, the probability of the listed company's financial information being considered unqualified is less As high as 50%, Lv Minrong (2011) not only re-verified and reached the same conclusion, but also studied that the establishment of the audit committee has an inhibitory effect on issuing non-standard audit opinions. Li Lili (2017) shifted the research focus to the top ten listed companies. Based on the proportion of shares held by famous shareholders, it is concluded that the promotion effect of Yue Heng (2006) is just the opposite. Combined with Qiu Shuangyu (2019)'s research on four factors in corporate governance, the audited unit's board of directors and equity are larger or independent. The smaller the number of directors and the size of the supervisory board, the greater the probability that financial reports will be deemed untrue and unreliable. Lin Jing and Su Qiuyun (2019) conducted research from the perspectives of the internal control quality level and pending litigation of listed companies, and believed that the financial reports of listed companies with high-quality internal control are more credible and truer, once the company has pending litigation., the facilitation of high-quality levels of back control will be diminished.
From the perspective of audit institutions, Fang Junxiong et al. (2004) found that whether it is the size of the audit institution or the profit brought by a single client, the impact on the importance ranking of the clients audited by the audit institution is more significant than that of non-standard audit opinions. Lv Minrong (2011) re-studied and found that firm size does not play a significant role in the rate of issuing non-standard audit opinions. However, Li Lili (2017) came to a different conclusion, arguing that the size of audit institutions has a significant inhibitory effect on non-standard audit opinions.
From the analysis of the market environment, Wu Weirong et al. (2015) believe that media supervision and government supervision in the market environment have an inhibitory effect on the issuance of non-standard audit opinions. Li Lili (2017) used analyst attention to replace the impact of market environment, and found that analyst follow-up inhibited the issuance of non-standard audit opinions. The promotion effect of external supervision on the issuance of non-standard audit opinions just verifies the role of listed companies' operating status and ownership structure in issuing nonstandard audit opinions.
(2) Literature research on audit interviews and audit quality Hilary and Lennox (2005) studied the flawed audit opinion in the peer inspection report issued by AICPA, which will lead to the risk of losing some clients to the audit institution, so the audit institution will try to avoid negative comments in the report as much as possible. Casterella, Jensen and Knechel (2009) studied the impact of peer review reports in AICPA on audit quality, and believed that peer review reports could predict audit failures to a certain extent. Gramling, Krishnan, and Zhang (2011) argued that the impact of being examined by the PCAOB was not significant among the Big Four, while the degree of relevance of non-conformity opinions issued by the non-Big Four after being examined by the PCAOB was enhanced. Lamoreaux (2016) believes that under the pressure of audit institutions that may withdraw from the US Securities and Exchange Commission due to unqualified inspection results after being inspected by PCAOB, the audit institutions will increase the audit workload and process.
Liu Zeshan and Chen Yan (2002) studied the supervision method combining industry self-discipline and government supervision of the audit industry in the United States, and believed that my country's audit market supervision will inevitably develop into a supervision system with administration as the main industry and supplemented by the industry. There are not many literatures on domestic market research audit interviews, because the CICPA started in 2011 and only implemented information disclosure in 2012. Wu Xi, Yang Yulong, and Zhang Junsheng (2014) believe that audit interviews lead auditors to adopt stricter audit procedures, while audit results caused by audit interviews are not significantly different. Qiu Yuemin and Lin  found that the annual report audit supervision interview system will prompt audit institutions and audited entities to be more cautious about related audit risks and take measures to reduce their risk levels. Another paper by the same author argues that audit interviews can improve the quality of internal control of enterprises through the external motivation of audit fees and audit opinions and the internal motivation of corporate board governance. The research of Mo Dongyan and Zhao Min (2019) shows that audit interviews will affect auditors' decision-making, and auditors will choose to adjust audit fees to reduce audit risks within a controllable range. Zhang Mengting (2019) also studied the impact of audit interviews on auditors' decision-making. She believed that audit interviews had a positive and significant effect on the audit fees considered by auditors when making decisions. Liu Wenhuan, He Yanan, and Zhang Jixun (2017) used experimental research methods to explore the effect strength of three different ways of face-toface + written, face-to-face, and written on investors, while Zhang Mengting (2019) used three different methods: face-to-face, written, and telephone. The approach angle explored the conclusion that face-to-face interviews did not have a greater impact on auditors' decision-making than written and telephone interviews. Wu Xi, Yang Yulong and Zhang Junsheng (2014) found that listed companies and audit institutions that have not been interviewed will not strengthen audit efforts and improve audit quality because of their own audit risks mentioned by the CICPA.
Huang Yixiong and Dong Yujun (2015) took the lead in finding through professional issues that the audit institution that was interviewed charged the audited entity lower than the fees that were interviewed without an audit. Yang Hailin (2017) believes that the supervision effect of audit interviews is an inverted U shape, and there is a spillover effect of audit interviews for a period of time, and clearly points out that the audit supervision effect has shown a downward trend in 2014. Liu Danlan (2018) believes that audit interviews are an in-process supervision mechanism. In order to avoid being issued non-standard audit opinions, listed companies are more inclined to choose to change firms instead of accepting adjustments and rectifications required by audit institutions. Zheng Yi (2018) believes that audit interviews are a preventive supervision measure that can promote the improvement of audit quality. Zhao Ying and Zhai Huayun (2018) believe that audit interviews have a negative correlation and significant effect on the earnings management behavior of audited units, and that state-owned enterprises and high-risk investment companies such as real estate are more significantly inhibited by the audit interview mechanism.
(3) Review of domestic and foreign literature research Both AICPA and PCAOB inspections in foreign markets have a positive effect on audit quality, and audit interviews in the Chinese market will also affect auditors' decision-making through audit fees and external transmission mechanisms for audit opinions. It also has the same effect on the audit quality of the annual report by the auditor, and also promotes the improvement of the internal control system of enterprises with internal control defects.

Current status of audit interviews
The time period for the project leader, partner or auditor of the audit institution to be interviewed by the CICPA shall be before the end of the audit work. The audit interview really started in 2011, but since 2012, the information of the interview has been published on the official website of the CICPA. However, at present, domestic auditing is mainly regulated by the government, and listed companies do not pay enough attention to auditing interviews. Therefore, the public maintains a wait-and-see and skeptical attitude towards the positive effects of auditing interviews. In this paper, the specific situation of the interviewees and the number of each year from 2012 to 2021 is manually collected from the official website of the China Association of Chinese Acupuncturists, as shown in Table.1.

Influence mechanism and research hypothesis of audit interview on non-standard audit opinions
Starting from the basic concept of audit interview, the CICPA interviews the project leader or chief partner of the audit institution to warn about certain types of audit risks existing in the audit process and propose audit procedures to deal with them. The purpose is to improve the audit quality of the annual reports of listed companies by auditors.
Can audit interviews achieve their goal of significantly improving audit quality in the audit market? Huang Yixiong and Li Changai (2016) are the earliest studies on the relationship between audit interviews and audit quality. After studying the effects of audit interviews from 2012 to 2015, the two scholars affirmed the role of audit interviews in promoting audit quality. Among them, Huang Yixiong and Li Changai (2016) used the numerical value of manipulated accrual profits to judge the level of audit quality in their empirical research. However, Yu Laichao (2013) used the audit opinion as the substitute indicator of the dependent variable when studying the variables that affect the audit quality. According to the different choices for judging the audit quality level above, audit interviews can have a significant positive effect on the audit quality indexed by manipulative accruals. Can audit interviews have the same positive effect on audit quality with audit opinion as an indicator? This is also one of the research significances of this paper.
From the perspective of an accounting firm, listed companies with the same and similar risks after the chief partner or person in charge of the firm are interviewed by the CICPA will be dealt with more cautiously by the firm. On the one hand, the audit institution will adjust the members of the project team responsible for such clients. In detail, the firm will appoint auditors with stronger professional ability or experience in dealing with the risk to be responsible for the audit. On the other hand, the supervision of the audit process of listed companies with such risks will be strengthened by the firm, and the audit work papers are not qualified and the audit procedures are not in place. In the context of audit interviews, audit institutions will take more rigorous risk response measures, which will significantly increase the possibility of a listed company's annual report being presented with unqualified opinions.
Starting from the auditor's countermeasures, the auditor will conduct audit work more cautiously after being interviewed by the CICPA in writing, telephone, face-to-face, etc. to avoid the risk of future reputation damage or punishment. On the one hand, auditors who are re-assigned by the firm to deal with the risk more professionally or with more experience will be more sensitive to the violations prompted by the audit risk. On the other hand, the certified public accountant will conduct an in-depth examination of the financial data related to the audit risk, and then adopt a more rigorous audit workflow after reasonably measuring the size of the risk it brings. On the basis of audit interviews, the substantial increase in auditors' prudence also significantly affects the probability of non-standard audit opinions issued by them.
From the perspective of the influence mechanism of auditors' decision-making, listed companies will beautify their financial reports to a certain extent in order to send investors a signal of good operation and avoid delisting. At this time, listed companies have a strong incentive to choose to increase the pricing of audit services to affect the auditor's opinion. Mo Dongyan and Zhao Min (2019) found that auditors also reduce the risk of future penalties and reputation damage by increasing audit fees. But under any circumstances would auditors and their accountants agree to increase pricing for public company services and issue qualified audit opinions? the answer is negative. Although audit institutions are willing to express qualified opinions in order to attract and retain customers, they need to conduct risk assessment tests on listed companies. When the expected risk premium in audit fees is lower than the hidden losses in the future, they will face greater reputation damage and be in the face of greater damage to their reputation. Auditors tend to issue unqualified audit opinions under the motive of punishment. Therefore, this paper argues that audit interviews will lead to higher audit risks assessed by auditors, which makes them more inclined to issue non-standard audit opinions to reduce the risk of future penalties and reputation damage.
Based on the above, it is hypothesized that: Hypothesis: On the basis that other variables are controlled, the possibility of non-standard audit opinions being published increases significantly after the audit institution is interviewed by the CICPA.

Data source, variable selection and model design
(1) Data sources Since 2012, the CICPA has published the time, object, subject and response procedures of audit interviews on its official website. According to the research content, this paper collects the relevant content such as the number, batches and methods of the CICPA negotiating firms from 2012 to 2021. From 2011 to 2020, the report opinions, audit fees and relevant information of the corresponding firms of listed companies are collected from Cathay Pacific. And financial data such as the total assets at the end of the period, the total liabilities at the end of the period, net profit, accounts receivable, inventory and other financial data of the listed company for the subsequent calculation of the ratio. Reasons for excluding listed companies in the financial industry and listed companies with special treatment: The financial statement requirements of financial enterprises such as banks are slightly different from the requirements of non-financial enterprises. The special treatment of listed companies is due to the special treatment of listed companies with financial and other anomalies when the securities are located to protect the interests of investors, and these two types of unstable listed companies are filtered out to avoid affecting the results of empirical analysis. In addition, some data with vacancies were deleted, and continuous data such as asset-liability ratio, return on assets, accounts receivable ratio, and inventory ratio were abbreviated from 0.01 to 0.99.
(2) Variable selection Based on the research on the impact of audit interviews on non-standard audit opinions, Table.2. below is a summary of information on all the variables in this study. Opinions is set as a dummy variable. When the annual report of the listed company is unqualified by the auditor, the indicator value is assigned 1, otherwise it is 0.

Independent Variable Interview
Interview is designed as a dummy variable. When the audit institution is warned by the CICPA about a certain type of risk, the index value is 1, otherwise it is 0.

LnAseet
The value of this indicator is the logarithmic data of the total assets in the annual report of the listed company.

Loss
Loss is judged according to the net profit in the annual report of the listed company. The loss status if the net profit is less than 0 is 1, otherwise it is 0.

Lev
In this indicator, total liabilities are the dividend and total assets are the divisor. When the ratio exceeds 1, it means that the company is insolvent.

ROA
Return on assets, also known as return on assets, or ROA for short. Calculated by dividing net profit by the average of total assets at the beginning and end of the year.

Rece
Among them, accounts receivable is the dividend of Rece and total assets is the divisor of Rece. The larger the value, the more aggressive the sales policy of the listed company.
Inven Inven=inventory/total assets, the larger the value, the more serious the inventory backlog of the listed company Year Virtual variable.

Big4
Virtual variable. The value is assigned 1 when the firm responsible for the auditee belongs to the Big Four (Deloitte, KPMG, Ernst & Young, PricewaterhouseCoopers), otherwise it is 0.

Last
Virtual variable. When the audit opinion of the listed company in the previous year is a non-standard audit opinion, the indicator value is assigned 1, otherwise it is 0. (1) Dependent variable. The explained variables are non-standard audit opinions or unqualified opinions (Opinions). Non-standard audit opinions (Opinions) are assigned a value of 1 when a listed company's annual report is issued an unqualified opinion by an auditor, and 0 otherwise.
(2) Independent variables. The explanatory variable is the audit interview, which is 1 when the audit institution is warned about a certain type of risk by the CICPA, and 0 otherwise. (

Descriptive statistical analysis
The first step in empirical analysis is to have a basic understanding of the collected sample data to test whether the collected sample can represent the population. Therefore, Table.3. lists the data obtained after descriptive processing of the samples using SPSS. This part analyzes the characteristics of the data based on the three major modules of the study: dependent variable, independent variable and control variable. The explained variable in this study is Opinions. The average level of the sample size of nonstandard audit opinions (Opinions) is 0.051, which is closer to the value of 0, and the value of the standard deviation of 0.219 is relatively small. It shows that there are few cases of unqualified financial reports of listed companies, and most of the financial report data of listed companies are evaluated as qualified and true by auditors.
The explanatory variable studied in this paper is the audit interview (Interview). The mean value of 0.551 for audit interviews is very close to the middle value of 1 and 0 but slightly skewed towards the presence of audit interviews, with a standard deviation of 0.497 compared to the standard deviation of non-standard audit opinions (Opinions). This shows that accounting firms account for more than half of the audit interviews each year.

Correlation analysis
After descriptive statistics before regression analysis, there is a preliminary test for the response of a single variable and dependent variable, that is, the effect of the explained variable due to the variable after excluding all other factors except the variable. Therefore, the analysis method in SPSS software is used to process the data. The following table.4. shows the coefficients and significant levels obtained after Pearson processing the data. The preliminary conclusion is that the audit interview has a significant negative correlation with the non-standard audit opinion. Next, the process of obtaining this conclusion is analyzed in detail combined with the tabular data.  (Note: *** refers to a significant level of 1%, ** refers to a significant level of 5%, the same as all tables in the empirical analysis) Focus on looking at the coefficient and significance level of the explanatory variable to the explained variable. Focusing on the analysis of the second column of non-standard audit opinions (Opinions) data, the correlation coefficient between audit interviews (Interview) and non-standard audit opinions (Opinions) is -0.022, and there is a significant negative correlation. It is initially found that the content of Hypothesis does not hold, that is to say, audit interviews have a significant negative correlation with non-standard audit opinions.

Regression analysis of the impact of audit interviews on non-standard audit opinions
After going through the first two stages of empirical analysis, the samples were then imported into SPSS for research using the OLS model. Considering that the dependent variable in this study is a dummy variable assigned a value of 1 or 0, in order to test the impact of audit interviews and nonstandard audit opinions, the Logit statistical method is used for the model. The regression results are listed in Table.5. Focusing on the observation that the regression coefficient of audit interviews (Interview) to nonstandard audit opinions (Opinions) is -0.203, which is significant at the 1% level, indicating that audit interviews significantly reduce the probability of non-standard audit opinions being issued by auditors. The conclusions of regression analysis and correlation analysis are unified, and both state that Hypothesis does not hold. This shows that after the person in charge of the firm was warned by the CICPA about a certain type of risk, the listed company took measures to reduce the expected risk of being punished by the CSRC, and the probability of auditors issuing non-standard audit opinions on the listed company decreased significantly.

Robustness check
When choosing the regression analysis method used in the empirical analysis, the first consideration is that the audit opinion of the dependent variable conforms to the scope of application of the logistic analysis method, but it does not mean that the sample size of this paper is not suitable for the OLS method analysis. Therefore, when testing the robustness of the conclusions of this paper, we choose to change the analysis method to verify.  Table.6. above. The regression coefficient of audit interviews to non-standard opinions is -0.008, and it is significant at the 1% level, which once again verifies the invalidity of Hypothesis. At the same time, it also verified the negative and significant impact of audit interviews on issuing non-standard audit opinions. In summary, the conclusions obtained by changing the regression method are consistent with the above conclusions.

Research conclusions
On the basis of observing the 2012-2021 audit interview information, combined with the 2011-2020 annual report financial data of listed companies, the role of the CICPA interview behavior on the auditors' decision-making opinions is discussed in depth. On the basis that other variables are controlled, the CICPA negotiates with the accounting firm, which will significantly reduce the possibility of non-standard audit opinions being issued. That is to say, compared with firms that have not been interviewed by the CICPA in writing or face-to-face, firms that have been interviewed by the CICPA have significantly reduced non-standard audit opinions on their clients' annual reports. This may be closely related to the fact that listed companies and their audit institutions attach great importance to the supervision form led by the CICPA. The risk subject of the audit interview is highly pertinent, and both the listed company and its audit institution have made countermeasures and plans for this risk, and the accounting firm needs to submit a statement after the event. Therefore, audit interviews have a significant inhibitory effect on the issuance of non-standard audit opinions.

Related recommendations
a. Improve the audit interview mechanism Because audit interviews take place before the completion of the audit of the listed company's annual report, it can be classified as either pre-event management or preventive supervision, which can alert certified public accountants to adopt more robust audit procedures and improve the quality of listed companies' annual reports. Although the public is skeptical about the role of audit interviews, this paper uses data to support the important conclusion that audit interviews have a significant negative impact on non-standard audit opinions.
(1) The follow-up of audit interviews needs to be improved urgently. At present, CICPA pays more attention to conducting prior interviews with the principals of securities firms by means of written, face-to-face, written + face-to-face, telephone, etc., but the management during and after the event is relatively weak and lacking. At present, only the accounting firm needs to submit a job description after the audit work, and the measures taken by the audit institution on the risks reminded by the CICPA cannot be tracked in real time. This requires the CICPA to improve the follow-up mechanism during and after audit interviews, and to supervise listed companies and certified public accountants to improve and perfect certain types of audit risks and defects that have been warned.
(2) The subject of audit interviews needs to be improved. The risk topics currently involved in audit interviews are all common in the audit industry, and listed companies and firms will have relatively complete response measures to the risks, which will limit the role of audit interviews. Therefore, the subject of the audit interview can be changed to be more detailed or novel.
(3) The number of lists for audit interviews needs to be optimized urgently. In 2016, the CICPA interviewed the heads of 40 firms other than Zhonghuan Haihua Firm, which attracted media attention and made the public doubt the role of the mechanism. As a result, the number of firms interviewed by the CICPA gradually decreased in the following years. However, Wu Xi, Yang Yulong, and Zhang Junsheng (2014) found that audit interviews failed to have a significant impact on the accounting firms that were not interviewed and the listed companies they were responsible for, which means that the spillover effect of audit interviews was not obvious enough. Therefore, the list and number of firms interviewed by the CICPA should be more cautious, and updated in real time with the problems existing in the listed companies, so as not to be ignored because of the interviews with the audit institutions on the same risks in the previous issue.
(4) The way of audit interview can be diversified. Liu Wenhuan et al. (2017) explored the written, face-to-face, written + face-to-face methods adopted by the CICPA and found that the written + faceto-face method had a stronger impact. However, with the advancement of technology, in 2017, the CICPA began to join the form of telephone interviews. The interview method adopted by the CICPA may also need to add different elements. Initially, in 2012, the CICPA adopted the form of face-toface interviews with accounting firms in Beijing, and subsequently adopted the form of written interviews with accounting firms in remote locations. With the prevalence of various office software, CICPA can adopt more diversified forms of audit interviews, such as video conferences that break through time and space constraints.
b. Balance audit market supply and demand There are more than 9,000 accounting firms, including branches, in the internal market. The phenomenon of supply far exceeding demand has prompted audit institutions to reduce the pricing of audit services and issue standard opinions in order to retain and attract clients. Therefore, the government needs to take measures to balance the supply and demand in the audit market. The Chinese government can establish a quality scoring mechanism for audit institutions and auditors, and withdraw audit institutions and auditors with unqualified quality scores from the market. Under this pressure, both audit institutions and auditors will need to work harder and learn to improve their abilities in order to avoid being withdrawn from the market. In order to adapt to the development of the times, auditors need to continuously improve their professional knowledge while enhancing their professional ethics in work processing. On the other hand, audit institutions also need to continuously improve the audit workflow execution system and supervision system.