Causes and countermeasures of research-integrity problems in the new era: From the perspective of the culture of science

Abstract

Research integrity is now one of the most important issues in the governance of science and technology (S&T) ethics in China. Many research results have been produced in the academic circle on this topic. A series of regulations and normative documents has been issued by government authorities, and numerous measures have been taken by universities and scientific research institutions. Despite this notable progress, the problems of research integrity have not been eradicated. Incidents such as rejections of papers by journals due to plagiarism and falsification of research results happen from time to time. Also, the awareness of research integrity among some S&T personnel and students has yet to be strengthened. The new challenge for research integrity in the new era is that of being able to locate the deep-seated causes of research-integrity problems and find effective ways to solve them. Exploring the causes of and countermeasures against those problems from the perspective of the culture of science is a research field that deserves greater attention.

1. Introduction

Research integrity, as a holistic concept, encompasses a range of qualities that researchers and research institutions must possess (Valkenburg et al., 2021). As such, research integrity is considered to primarily involve researchers and research institutions (Davies, 2019; Helgesson and Bülow, 2023; Horbach and Halffman, 2017; Shaw, 2019; Steneck, 2006). However, according to a review of the empirical research on research integrity conducted in the past decade, most empirical analyses focus on indicators involving individual researchers, while paying little attention to the impact of institutional policies and cultural practices (Aubert Bonn and Pinxten, 2019). Although there have also been some discussions of the culture of integrity, such as the ‘culture of peer review’ (Atkinson, 2001) and the ‘publish-or-perish culture’ (Génova and De la Vara, 2019), on the whole, discussions of research integrity from a culture-of-science perspective are still lacking in the current literature.

The culture of science is the way of life of the scientific community. It consists of a set of value systems, ways of thinking, institutional constraints, codes of behaviour and social norms formed around scientific activities (Meng, 2023). The culture-of-science perspective pays more attention to the underlying ideological and social roots of research-integrity problems, focuses on addressing both symptoms and root causes, and aims to gradually move the governance of research integrity from the operational level up to the conceptual level. Research-integrity governance is not just about ‘restriction’ but also about ‘facilitation’. The objective of research-integrity governance is to enable scientific and technological (S&T) personnel to put their talent to best use in a healthy and orderly research environment. Therefore, this paper seeks to examine the issue of research integrity in the overall social and cultural environment, explore the causes of research-integrity problems from three dimensions (the institutionalisation of scientific activities, talent-evaluation standards and technological means) and put forward targeted recommendations and solutions in relation to these three causes that can facilitate a unique path for the analysis of research-integrity governance.

2. The cause of research-integrity problems viewed in the context of the institutionalisation of scientific activities

On reviewing the history of the emergence and development of modern science, we can see that the problem of research integrity gradually revealed itself with the institutionalisation of scientific activities. At the beginning, scientific research activities were a kind of spontaneous, purely academic social practice without obvious utilitarian effects. Many scholars engaged in scientific research as a part-time job. The formal occupations of scientists, such as Copernicus and Galileo, were usually as teachers or clergy. They were engaging in theoretical research and observational experiments only in their spare time, which not only required great manpower and material resources but also involved huge risks. Until the end of the seventeenth century, they mainly relied on private correspondence to announce the results of their research, which caused many disputes over the priority of inventions, such as the Newton–Leibniz quarrel over the invention of calculus (Zhou, 1988). Although there were also controversies in the history of science about the rigour and reliability of the experimental data of certain modern scientists (mainly involving the modification of data by these scientists in order to obtain beautiful theoretical results) (Broad and Wade, 1983), research integrity was not yet an outstanding social issue at that time because research activities were not linked to any practical utilitarian purpose in society.

However, the institutionalisation of scientific activities changed the whole situation. A series of institutes dedicated to scientific research were established, such as the Royal Society in 1660, the Académie française in 1669, and the Académie des sciences de Berlin in 1700. Philosophical Transactions, founded by the Royal Society in 1665, is considered the world's first scientific journal. The first scientific award, the Copley Medal, was awarded in 1713 by the Royal Society. By the eighteenth century, scientific societies and science academies had been established across all European countries, and universities also started considering scientific research as their responsibility. Research activities were being institutionalised, the scientist profession was gradually established, more academic societies, journals and reward mechanisms were introduced, and the results of scientific research were being increasingly linked to honour, status and interests (Zhang, 2017). In 1834, William Whewell, a historian and philosopher at the University of Cambridge, coined the term ‘scientist’, which has been widely used ever since. It marks the beginning of viewing scientists as a relatively independent social group and scientific research as a professional social activity with its own publication, evaluation, appraisal, reward and punishment mechanisms (Ye, 2000).

With the institutionalisation of scientific activities, scientists have gained stable working positions, places, equipment, communication channels and other necessary conditions. Outstanding scientists are respected by all sectors of the society and rewarded with various kinds of encouragement and material incentives for their research results. Institutionalised scientific activities require that scientific research conforms to certain academic norms and strict evaluation standards, and that members of the scientific community follow these requirements. However, the evaluation and recognition of research results have become increasingly complex and now include the difficult tasks of re-verifying experimental results in a timely manner, conducting peer review of interdisciplinary research results, and identifying whether certain papers or monographs are the original work of the authors. Because of these loopholes in research governance, some S&T workers, for utilitarian purposes, tend to seek illegitimate benefits by falsifying research results, adding to the problems in research integrity.

As an integral part of the culture of science and the scientific spirit, S&T workers are encouraged to make a selfless sacrifice, be indifferent to fame and fortune, and dedicate themselves to scientific research. That said, it is simply impossible not to think about the social rewards of research activities or the honour, status and interests of S&T workers. This change in attitude requires the creation of a cultural atmosphere that enables the healthy development of scientific endeavours and the continuous improvement of the academic evaluation and reward system to ensure fairness and impartiality. If the work of academic evaluation and reward is disrupted by non-academic factors (such as interpersonal relations, trading of benefits and canvassing), that will make people feel unfairly treated and promote academic misconduct.

One such example is the academic fraud case of Hwang Woo-Suk in the Republic of Korea (ROK), which shows the impact of certain cultural factors. The fact that Hwang Woo-Suk dared to publicise the results of his academic fraud, and that the results were reviewed and published by prestigious journals such as Science, indicates that the academic-review mechanism at that time was not sufficient to detect the fraud immediately. At that time, the ROK government honoured Hwang with awards of the highest class and provided him with huge financial support. Hwang himself was also seen as a national hero by many Koreans, reflecting how anxious they were to have Nobel Prize-winning research results in their country. However, the Korean academic circle and related organisations soon discovered Hwang's academic fraud. This shows that they were able to take timely and proactive actions against misconduct in research integrity, but the incident had already had a significant negative impact. Identifying the social and cultural factors that may induce such incidents and establishing the rules necessary to prevent them from happening are aspects that need to be addressed.

Similar incidents have also happened in other countries, such as Summerlin's ‘patchwork mouse’ in the United States and Haruko Obokata's thesis-falsification incident in Japan. In China, the ‘Hanxin’ chip fraud case of Shanghai Jiaotong University's Chen Jin was exposed in 2006. From Chen Jin announcing the creation of ‘Hanxin No. 1’ to the disclosure of the scandal, a total of three years had passed. During this period, Chen Jin received not just a large amount of research funding, but also honorary titles such as ‘Cheung Kong Scholar’ and ‘Shanghai High-tech Start-up Leader’.

The above incidents all have one thing in common: ‘time lags’, some longer and some shorter, between the beginning and exposure of the fraud. Perhaps some people's academic fraud will never be uncovered, and the investigations of some suspected academic frauds will remain inconclusive for a long time (such as the Baltimore incident in the United States) (Wang and Hu, 2016). Due to the time lag in dealing with academic fraud, people who seek to exploit the loopholes in the review and publication of academic results feel that they can do so, so they are willing to take their chances and think they can get away with it. If these people are able to get benefits through academic fraud and out-compete scholars or students who observe the bottom line of research integrity on thesis publication, promotion, projects and awards, over time, the social atmosphere will be poisoned.

As a basic ethical requirement for scholars, ‘observing the bottom line of research integrity’ requires not only self-discipline but also a robust regulatory mechanism to ensure that academic rewards and punishments are well targeted and effective. However, as S&T management becomes increasingly complex, it is becoming more and more difficult to achieve such results. The institutionalisation of scientific activities has led to an increasing specialisation of S&T management. Even though many scientists work in S&T management, their social functions are different. S&T management is like a production line, with different links dealing with different issues. Before a case of academic fraud is exposed and investigated, all officially published academic research results related to the case are submitted to the institutions for reviewing scientific research projects, academic awards and title promotion, yet those departments do not have the function of checking research integrity. The existing governance mechanism for research integrity has proven effective in formulating policies and regulations as well as investigating and dealing with exposed problems, but it is unable to detect and deal with malpractices in their embryonic state or ‘treat the disease before it shows symptoms’, as in traditional Chinese medicine.

From the perspective of the culture of science, there are various explicit or implicit organic links between the scientific community and the social environment (Cetina, 1999). The governance of research integrity requires not only dealing with the issue itself, but also considering the impact of relevant social and cultural environments, discovering the causative factors and finding the potential ideological breeding ground of academic misconduct. This brings us to the issues of talent-evaluation standards, technological means and the effectiveness of research-integrity governance.

3. The cause of research-integrity problems viewed in the context of talent-evaluation standards

The existence of research-integrity problems is closely related to the application of improper talent-evaluation standards. Nowadays, universities and scientific research institutions take cultivating and attracting talented people as their top priority, but there are some misunderstandings about the standards for evaluating research talent, such as ‘the more academic results the better’, ‘the faster the growth of the talent the better’, ‘the more comprehensive the visible indicators the better’, and ‘the more sophisticated the indicators the better’. Such an orientation will fuel the tendency towards quick and instant results, which may easily induce academic misconduct and undermine the integrity of scientific research.

3.1 The tendency of ‘the more the better’

When evaluating the performance and competence of S&T personnel, some institutions consider a specific number of academic results as the sole criterion and advocate that more is better. The more results published, the more the scientist is appreciated and rewarded. For example, if a person publishes five Science Citation Index (SCI) papers a year, he will be considered not as good as someone who publishes six. Such an unhealthy competition will soon exceed people's normal physical and mental carrying capacity and lead to only two possible results: either hurting the health of the scientist or turning him into an opportunist. A few years ago, some institutions used to link the number of SCI papers to the rewards one could receive. As a result, some people published dozens of SCI papers a year. If the number is within the reasonable range of people's physiological and psychological capacity, the publication of high-standard papers and other results can indeed reflect a person's academic level. However, if it is apparently beyond the reasonable range, then ‘the more the better’ would be abnormal. Of course, the definition of such a reasonable range should be analysed in light of the characteristics of different disciplinary fields.

In order to solve this problem, some universities and scientific research institutions are now implementing the ‘masterpiece evaluation’ system, which, to a certain extent, can eliminate the problem of ‘the more the better’. However, ‘masterpiece’ reflects mainly the quality and level of research results, and a comprehensive evaluation also needs to consider the number of research results. If the originality reflected in the research results is the main purpose of evaluation, a very professional peer review would be necessary, which is difficult to do in most universities and research institutions. After all, modern scientific research is becoming more and more subdivided and specialised, and it is very difficult to find appropriate reviewers for every result. For some highly specialised disciplines, there might be only a small number of peers, and it is difficult to organise these people for the evaluation of a particular research result. It may be even more difficult to find ideal peer reviewers for research results that transcend disciplines. In many cases, interdisciplinary peer review is still focused on quantity, and the perspective of ‘the more the better’ will continue to play its role.

3.2 The tendency of ‘the faster the better’

When evaluating the ability and potential of S&T workers, some institutions often focus on their age when they reach a certain academic level. These institutions also believe that the faster one grows in their career, the greater one's potential. In China, the most ideal trajectory for a scientist is to graduate with a doctorate in their 20 s, become a professor and receive the title of ‘outstanding young scientist’ in their early 30 s, be accredited as a ‘Cheung Kong Scholar’ at the age of 40, and then become an academician. It would be even better if one could win a major national or international award under the age of 50. Such an orientation pushes S&T workers to reach for higher levels as quickly as possible—the faster the better. Because of this, they may easily become impatient, pay too much attention to staying ahead of the game, and even undermine research integrity.

The professional growth of S&T workers is a gradual process, and the development of original research results requires long-term accumulation. Overemphasis on ‘the faster the better’ is like ‘pulling up seedlings to help them grow’, but the truth is that haste makes waste. Tu Youyou won the Nobel Prize in Medicine at the age of 85. If the evaluation standard of ‘the faster the better’ had been applied to her, it would have been difficult for her to receive adequate attention and support. Of course, if it is within a reasonable range, S&T workers who make a faster career progression do have stronger academic strength and growth potential, but ‘the faster the better’ should not be taken as an absolute standard for measuring S&T talent.

3.3 The tendency of ‘the more comprehensive the better’

Some institutions like to put all the indicators of the academic competence and influence of S&T personnel, such as research projects, research results, citation rates, awards, honours and academic part-time jobs, in their evaluation indicator matrix. In colleges and universities, indicators such as teaching performance, education reform projects and talent training are also added. When evaluating one's competence, these indicators are often scored separately and added up linearly to produce the total score. This means that one must try to score in as many visible indicators as possible. From the perspective of all-round talent development, such a requirement does make some sense, but it may encourage the unnecessary pursuit of perfection. As a result, some people will have to engage with all the elements in order not to miss any item, resulting in a preference for quantity over quality; that is to say, high-quality results will be replaced with a large number of less valuable results.

Of course, these visible indicators are necessary for the overall development of research institutions and universities, and they are also the indicators being assessed in some university rankings and discipline rankings. However, any individual will always have strengths and limitations. It is neither reasonable nor realistic to ask a person to contribute equally to all the indicators of an institution. In particular, forcing those whose work focuses on application, such as clinicians and front-line engineers, to publish high-standard papers only for the purpose of ‘not missing any item’ may well affect the reasonable evaluation of their actual ability and even lead some people to issue false papers.

3.4 The tendency of ‘the higher the better’

Nowadays, when evaluating S&T talent, some organisations prefer using hierarchical indicators, such as academic qualifications, job titles, achievement levels and so on—the higher the better. In many cases, these indicators can even replace the indicators for evaluating academic achievements. If a person graduated from a prestigious university with a superior degree, as long as they have one academic achievement, project or award that has reached a relatively high level, even if they have not many academic achievements on the whole, they will still be able to get a good evaluation result. On the other hand, a person who graduated from an ordinary university would find it difficult to receive extra attention, unless they have produced outstanding research results. Such an orientation does make sense in certain aspects, but, once it is adopted as the sole criterion, it may affect the development of S&T workers who have lower starting points and less prestigious educational backgrounds.

Opting for ‘the higher the better’ may strengthen the ‘Matthew effect’ in the scientific community, making it easy for scientists who have already scored on high-level indicators to get scores on more high-level indicators and thus receive higher evaluations (Hao et al., 2019). By contrast, the average scientists will need to spend a lot of time and energy to get scores on these high-level indicators, such as trying all kinds of ways to publish papers in authoritative journals, competing for big projects and competing for high-level awards. In such cases, it will be difficult, if not impossible, to prevent academic misconduct from happening.

There are many reasons why these misunderstandings exist in talent-evaluation standards. First, at present, many institutions choose the approach of ‘general peer review’ when evaluating S&T talent. Experts from different disciplines make their own judgement based on the performance of S&T workers in various aspects. Due to the differences in specific disciplines, such an evaluation can examine only the values listed in the indicator matrix and cannot accurately judge the specific academic competence and ability reflected in those figures. Some institutions, out of the need for disciplinary and workforce development, try to include as many high-level indicators as possible when designing the indicator matrix. In effect, they are mixing certain functions of scientific research administration and education administration with the design of talent-evaluation indicators, expecting to cultivate ‘generalists’ and ‘well-rounded specialists’. However, such an expectation is not consistent with the trend towards the development of highly specialised yet cross-cutting disciplines. If general peer review is only used to compare the scores of various indicators in the evaluation system, even a staff member, rather than an expert, could do the job.

Second, at present, many scientific research institutions and universities are eager to reach top-class status at home and globally as soon as possible. The evaluation indicators of these institutions are often directly related to the allocation of financial support and academic resources, and the evaluation results of S&T personnel are also directly related to their personal interests. Therefore, the pursuit of more academic results and faster growth of talent is driven by a strong interest. In the absence of rational reflection, the potential emotional factors may, though not by design, guide the evaluation of S&T talent. The blind pursuit of the number of SCI, Engineering Index (EI) and other high-level papers does not necessarily lead to the improvement of S&T innovation capacity, and many scholars are now already reflecting on this trend. The evaluation of S&T talent is a very rational practice and must not be influenced by irrational emotional factors.

Finally, misunderstandings about how S&T talent should be evaluated also stem from the simplistic practice of quantitative analysis and linear thinking, which ignores the fact that the various characteristics of S&T personnel are an integral whole and can be accurately grasped only by using an interconnected and complex way of thinking (Wang and Li, 2016). There is an intrinsic logical link between the various characteristics of S&T talent: some characteristics are fundamental, and some are determined by or derived from those fundamental representations. The most important characteristic of S&T talent is academic innovation ability, and the factors directly associated with it are the knowledge base, academic competence and academic organisational ability. The prerequisites for its existence and development are academic platforms, financial support and physical and mental health, while academic achievements, personal growth, awards, academic part-time jobs, honours and tittles are all derivative factors. All these factors have complex relationships with each other, sometimes mutually reinforcing and sometimes mutually restricting. The index matrix for the evaluation of S&T workers can be seen as a collection of discrete numbers. When the evaluators simply compare them against each other or, even worse, when they just look at the grand total, the organic whole of the characteristics of S&T talent will be disintegrated, leading to nothing but one-sided or even wrong judgements. The purpose of evaluating S&T workers is to promote their healthy growth and sustainable development. We should avoid putting too much emphasis on some of the factors while ignoring the organic link between all the factors. Most importantly, we should not set unrealistic orientations that encourage people to focus on less important things while putting the most important things aside and to breach the code of research integrity to seek quick results.

In the evaluation of S&T workers, we often see such a phenomenon: those who advocate ‘the more the better’ for academic achievements are often non-professionals, not the evaluators themselves; those who advocate ‘the faster the better’ for the growth of talent often do not have the same or similar experience themselves; those who advocate ‘the more comprehensive the better’ for visible indicators are approaching the issue from the perspective of research management and education administration, rather than putting themselves in the shoes of the S&T personnel; and those who advocate ‘the higher the better’ for the hierarchical indicators are often experts who make condescending demands, with no regard for the specific conditions of each research institution or university and the actual situation faced by S&T personnel.

If we could get down to the actual research activities conducted by S&T personnel and put ourselves in their shoes to think about what they are allowed to do or capable of doing and how effective their work could be, then we will be able to connect all the indicators into an integral whole and make reasonable and dynamic evaluations of their performance, so that the evaluation mechanism could truly serve as an incentive for S&T talent.

4. The cause of research-integrity problems viewed in the context of technological means

‘Technological means’ in this context refers to the specific tools, methods and paths that are related to research integrity and used for dealing with the data, charts, literature, writing style, and expression and utilisation of research results in academic research. The development of artificial intelligence (AI) provides convenient tools for academic misconduct, such as plagiarising other people's research results by using searching and copy–paste tools, retouching pictures or tampering with data for the purpose of getting the best effect, and translating foreign literature with AI-powered translation software and presenting the translations as originals. The invention of generative AI software, such as ChatGPT, has enabled a new leap in generating academic results using technological means (Pournaras, 2023).

Plagiarism-detection software has played a significant role in detecting outright plagiarism in academic results, but it mainly utilises thesis databases and lacks more effective measures against direct plagiarism from masterpieces. Moreover, new technological means are now available to reduce repetition, and papers can easily pass the test by replacing quotations of the source literature with new formulations. Some products of academic misconduct are difficult to detect, identify and address. For example, it is difficult to distinguish whether an unexpected research result published by certain researcher or graduate student, which is apparently beyond their usual academic prowess, is a major breakthrough founded on a long period of accumulation or a product of plagiarism by skilful means. Even if their colleagues and classmates may have suspicions, very few of them will bother to follow the lead and verify the results.

Recently, some groups and individuals have turned academic misconduct into a profitable business, claiming that they can write all kinds of papers and monographs, publish academic results and apply for high-level research tasks on behalf of the author. The products of academic misconduct provided by these businesses can look very real. The problem is that there is still a big market for these kinds of products, especially among those who are weak in academic research but eager to publish high-level academic results. Recently, technological means based on ChatGPT and other generative AI have led to the rise of new tools for producing academic research results (Herbold et al., 2023). Papers generated using ChatGPT are now able to pass reviews for formal publication. Many prestigious journals and universities have made it clear that they do not accept papers generated directly by ChatGPT, but the adoption of this new tool is growing to varying degrees.

How can we encourage and ensure the originality of academic research? How can we make academic journals an avenue for introducing new ideas, rather than standard components mass-produced using technological means? These questions also require in-depth reflection. New technological means coupled with the commercialisation of the generation of academic results have made the governance of research integrity increasingly challenging. Of course, to contain the use of technology in manufacturing academic-misconduct products, we can also develop various technological means for detecting the traces of ChatGPT in academic results, invent more effective plagiarism-detection software, and develop tools that can identify data tampering and photo retouching in research results. However, this is like ‘applying medicines only to address the symptoms’ and cannot fundamentally solve the problem.

There are many typical cases of overhauling academic misconduct both at home and abroad, and significant progress has been achieved, but many institutions are only meting out light punishments for such incidents, making it difficult to fundamentally reverse the practice of violating research integrity for personal gain. This is because, for the person involved in academic misconduct, while the punishment is bearable, the potential gains are even more tempting. For a graduate student whose dissertation is found to be plagiarised, the most likely punishment is the revocation of an academic degree, which usually has little impact on his/her subsequent work and life. For university teachers and researchers involved in academic misconduct, the most common punishment is disqualification from promotion and the recruitment of postgraduates and applications for awards and honorary titles within a certain period of time; dismissal from public office is a rare occurrence. For some reported cases of suspected academic misconduct, the investigation might not produce any conclusive results. Of course, such investigation needs to be conducted with caution and according to different disciplines, and the responsibilities of the parties involved need to be clearly distinguished. Neither should we condone such incidents, nor should we overburden the parties of indirect responsibility with excessive social pressure. The crux of the matter lies in whether the investigation and punishment have produced any practical effect in reversing the unhealthy trend within and outside the academic community. If people think that it is just bad luck that one is penalised for academic misconduct, then such incidents are likely to continue to happen.

5. Countermeasures and future prospects of research-integrity governance from the perspective of the culture of science

From the perspective of the culture of science, the governance of research integrity requires us to foster a scientific cultural environment conducive to the governance of research integrity on the basis of existing laws, regulations and institutional measures (Valkenburg et al., 2021). The measures include deepening education on research integrity and avoiding the tendency of going for quick results in institutional design, so that researchers will no longer be bound by irrational evaluation indicators, and so that original research achievements will be duly valued and rewarded with the right incentives. Researchers with outstanding performance in research integrity should be identified and promoted as role models for others to follow.

In order to reverse the orientation of ‘the more academic results the better’, we should focus on the quality of the masterpieces of S&T personnel, especially the originality of those works. This requires a more rigorous specialised peer review, in which the authoritative experts in the discipline could give their objective and fair opinions. Academic results of original innovation, major theoretical breakthroughs and important practical applications should be given high scores in evaluations. Of course, specialised peer review may also have its drawbacks. For example, it is more likely to be affected by the interpersonal relationships between experts in the same circle or the subjective opinions and even conflicting views of the experts on the panel. Therefore, the reviewers should point out specifically how theoretical value and originality are reflected in the candidate's work, and where the candidate stands at the domestic and international academic frontiers, as well as assume the responsibility of guaranteeing the academic credibility for their own evaluation opinions.

To reverse the orientation of ‘the faster the talent grows the better’, we should focus on examining the academic potential and development momentum of S&T personnel and give them appropriate incentives based on reasonable predictions. If a person has a solid academic foundation, maintains strong development momentum and continuously produces high-quality results, even if he/she is advancing relatively slowly in career or at an older age, we still need to give them adequate attention. The biological age and academic age of S&T personnel are not exactly equivalent and need to be treated differently. For young S&T personnel, the emphasis should be placed on their basic S&T innovation ability, so as to avoid caring too much about the number of their academic achievements and putting them through too many academic appraisals. Researchers engaged in original innovation should be given more time and strong support. If we only encourage young scientists to seek prestigious titles as soon as possible and take this as the only criterion for allocating academic resources, it is likely that some original innovations that require long-term accumulation will slow down or even be cancelled prematurely due to the lack of necessary support.

To reverse the orientation of ‘the more comprehensive the indicators the better’, we should make a distinction between different types of S&T talent, adopt a classified approach to their evaluation, and allow S&T workers to have relative weakness or ‘missing items’ in certain aspects of the indicator matrix. In the history of science and technology, there are countless examples of S&T personnel with outstanding performance in some aspects and relative weakness in others. For example, Robert Boyle was not good at socialising, Chen Jingrun was not good at teaching, and Tu Youyou did not have the title of academician or overseas educational experience. Some people who are strong in research might not have research projects in their hands, and some people who are highly localised may lack the experience of international academic exchange. We need to be tolerant when evaluating such S&T workers, focus on their academic innovation ability and development potential and open up ‘green channels’ for them, so as to support the growth of people with extraordinary talent.

To reverse the orientation of ‘the higher level the indicators the better’, we should take indicators such as academic qualifications, job titles and achievement levels as reference factors, focus on the real talent of S&T personnel, and identify talent based on true merits. Many academic leaders of key Chinese universities and research institutions studied in ordinary universities during their early years. Although, with the popularisation of higher education, talent cultivation of colleges and universities at different levels is diverging, some outstanding people are still able to change their academic fate through hard work. Therefore, there should be some special indicators for the evaluation of such people, such as examining where they stand among their domestic and international peers, how quickly they are improving their academic competence, their influence in the academic community and the prospects of their research topics, in order to create an environment for S&T personnel with real potential.

In the matrix of evaluation indicators for S&T workers, it is also necessary to add indicators that examine their academic integrity and social responsibility. They need to have noble academic ethics, consciously resist academic misconduct, strictly abide by the bottom line of research integrity, and prudently deal with the social risks or negative impacts that may be brought about by S&T innovation activities. If there is a problem with research integrity, no matter how impeccable the person's performance is on other indicators, the rule of ‘one-vote veto’ should apply.

Some S&T personnel may blame the misguided orientation of S&T evaluation for their breach of research integrity, but not all will cross the bottom line of research integrity in the face of misguided orientation. Therefore, S&T personnel need to constantly adjust their academic mind-set, refrain from seeking quick success and instant results, reject the speculative mentality, and be ready to endure loneliness or being underappreciated. This is the foundation for observing the bottom line of research integrity. Establishing a reasonable mechanism for the evaluation of S&T talent and encouraging S&T personnel to cultivate academic integrity are mutually reinforcing. Only when they are aligned with each other can we instil the power of innovation into China's S&T undertaking and produce a constant stream of S&T innovation talent to make greater contributions to the country's economic and social development.

To promote the governance of research integrity, it is necessary to strengthen positive guidance and prop up role models (Kretser et al., 2019). Drawing on the ideas and wisdom of traditional Chinese culture, the governance of research integrity should not only utilise the governance strategies of Legalism, but also adopt the moral indoctrination approach of Confucianism, with a combination of moral self-discipline and imposed disciplinary restrictions. In promoting the lofty deeds of scientists, such as the pursuit of truth, rigorous academic approach and meticulous academic style, we should not just talk about Copernicus, Galileo and Kepler, but also tell people how contemporary scientists deal with fame and fortune, observe the principle of research integrity and resist academic misconduct. We could consider including a ‘research credibility index’ in the talent-evaluation indicator matrix. If one always does a good job on research integrity, one should be given a high score on the research credibility index through serious evaluation and receive the corresponding honours. The index should serve as an important indicator for the overall talent evaluation. In cases of academic misconduct, a poor rating should be given, depending on the seriousness of the violation, so that the person concerned will realise that this is a stain that cannot be erased in their academic career.

We also need to underscore research integrity in the education of S&T ethics. Research-integrity education is not purely intellectual education. The problem cannot be resolved through policy lectures only. The idea of research integrity must seep into people's brains and hearts. Education on the basic principles of research integrity is not purely knowledge-based education, but an education on academic integrity that requires the combination of knowledge and action. Only through the cultivation of moral passion can that moral awareness be transformed into stable moral behaviour. Greater efforts are also needed to explore the methods and evaluation approaches in research-integrity education.

The governance of research integrity is itself an integral part of the endeavour to promote the culture of science. The construction of research integrity is a fundamental work of S&T ethics governance. We need to step up research on the causes of and countermeasures against research-integrity problems in the context of the culture of science and strive to take research-integrity governance to a higher level.