How researchers perceive research misconduct in biomedicine and how they would prevent it: A qualitative study in a small scientific community

ABSTRACT

The aim of our study has been to use a qualitative approach to explore the potential motivations and drivers for unethical behaviors in biomedicine and determine the role of institutions regarding those issues in a small scientific community setting.

Three focus groups were held---two with doctoral students and one with active senior researchers. Purposive sampling was used to reach participants at different stages of their scientific careers. Participants in all three focus groups were asked the same questions regarding the characteristics and behaviors of ethical/unethical scientists, ethical climate, role, and responsibility of institutions; they were also asked to suggest ways to improve research integrity. The data analysis included coding of the transcripts, categorization of the initial codes, and identification of themes and patterns.

Three main topics were derived from the focus groups discussions. The first included different forms of unethical behaviors including increasing research “waste,” non-publication of negative results, authorship manipulation, data manipulation, and repression of collaborators. The second addressed the factors influencing unethical behavior, both external and internal, to the researchers. Two different definitions of ethics in science emerged; one from the categorical perspective and the other from the dimensional perspective. The third topic involved possible routes for improvement, one from within the institution through the research integrity education, research integrity bodies, and quality control, and the other from outside the institution through external supervision of institutions.

Based on the results of our study, research misconduct in a small scientific community is perceived to be the consequence of the interaction of several social and psychological factors, both general and specific, for small research communities. Possible improvements should be systematic, aiming both for improvements in work environment and personal awareness in research ethics, and the implementation of those changes should be institutional responsibility.

KEYWORDS:

• Academic institutions
• ethics
• qualitative research
• research integrity
• research misconduct

Introduction

Science has become a very competitive endeavor. Ideally, the scientists should be honest, independent in their work, have good relations with their colleagues, and be loyal to the organization where they work (Wong, Lim, and Quinlan 2016) However, the environment where scientific performance is measured by the number of publications, combined with constant pressure to obtain funds for research, creates the potential for the occurrence of unethical behaviors (Edwards and Roy 2017).

Previous research defined several positive predictors for research misconduct: lack of attention paid to research integrity by the institutions, monetary incentives, lack of mutual criticism in institutions, and younger age of researchers (Pupovac and Fanelli 2015). Furthermore, psychological factors like Machiavellian personality traits were also found to be the risk factor for scientific misconduct, especially for biomedical students (Tijdink et al. 2016). The interaction of several different factors may contribute to research misconduct or inappropriate research practices (Fanelli, Costas, and Ioannidis 2017). The hypothesis that individuals who work in an intensive “publish or perish” climate would behave more unethically has not been confirmed when observing the number of article retractions (Fanelli, Costas, and Lariviere 2015), probably because publication retraction may be a very crude measure of research misconduct and may not distinguish honest error from unethical behavior (Wager and Williams 2011). “Publish or perish” may also be relevant for small communities at the scientific periphery, defined by the lack of critical mass of researcher to produce sustainable research, lack of funding, and language barriers (Marušić and Marušić 2012).

Although the recommendations for increasing the reproducibility and transparency in research are already available (e.g., data sharing, trial registration, improvement in research training, and redefinition of excellence criteria) (Ioannidis 2014), currently there is little evidence that interventions aiming to resolve the issue of research misconduct are effective (Marusic et al. 2016). There is a recommendation from the World Conference on Research Integrity that research institutions should act in coordination with journals in order to detect and sanction research misconduct of their employees (Wager and Kleinert 2012).

Research misconduct, when exposed, may also create problems for the parent institutions and research groups directly related to the perpetrator (Krstic 2015; Wilson et al. 2007). The ineffectiveness of interventions indicates that scientific misconduct should be approached in a more comprehensive and systematic way, with academic institutions possibly having the central role in the process.

Whereas previous research has focused on research of risk factors for research misconduct (Krstic 2015), there is still a lack of qualitative research on the ways to improve responsible conduct of research. The aim of this study was to use a qualitative approach to explore the potential causes of unethical behaviors and determine the role of institution regarding those issues in a small scientific community setting.

Methods

Study design and participants

We held three focus groups---two with doctoral students and one with research experts. We used purposive sampling to reach participants at different stages of their scientific careers. The included participants were doctoral students currently in their third year of a biomedical Ph.D. program (2 groups) and experts–--researchers who already obtained their Ph.D. and had significant previous scientific experience (defined as 3 or more published articles) (1 group).

Participants were contacted via an e-mail invitation and asked to participate in a focus group discussion regarding research integrity.

Setting and data collection

Focus groups were held and recorded at the University of Split School of Medicine (USSM) in the after-work hours during April and May 2017. The USSM is located in Split, the second largest city in Croatia (metropolitan area population of about 350,000). It is one of the four medical schools in Croatia. Science and research at the School of Medicine in Split is high above the Croatian average, where 187 employees (80.3 full time equivalents) at the School of Medicine annually publish between 180 and 190 articles in journals indexed in the Web of Science. In the last five years, USSM participated in 48 projects–-21 international, 2 bilateral, and 25 national projects (Vidak et al. 2017). During the focus group, nobody but the lead researcher (IB) and participants were present in the room. The duration of each focus group was around 50 minutes. The conversations were recorded with an audio recording device and transcribed to Microsoft Word by the principal investigator. The transcripts were coded using Microsoft Excel. The questions for the focus group were prepared in advance, and all three focus groups were asked the same questions as follows:

How would you describe an ethical scientist?

How would you describe an unethical scientist?

Why does someone behave unethically?

How may the ethical climate influence scientific misconduct?

What is the role of the institution in the shaping of ethical climate?

For which behaviors is the responsibility institution’s and for which is the individual’s?

How may research integrity be improved?

Participants were also permitted to discuss issues among themselves and expand on the themes from the questions and to support their statements with examples.

Research team and data analysis

Most participants were familiar with the research team before the beginning of the study and the interviewer was the lead investigator of the study (male doctoral student, currently employed by the same institution as the participants and working on the project focused on professionalism and research integrity). The data analysis included the coding of the transcripts, categorization of the initial codes, and generation of themes and patterns. The participant’s identities were anonymized by the lead investigator (IB) (coded as P1 to P13). Two researchers (IB, LB) provided the initial coding of the themes based on the participants’ comments. Preliminary themes were derived during the analysis based on the focus group questions and then iteratively refined using constant comparative approach (Boeije 2002), but also allowing the identification of new themes to the existing ones. In the initial coding, a large number of codes was defined (>20) by one of the researchers (IB) and then revised by two researchers (IB and LB) to reach the final consensus. Each code was placed in a single theme category only. The saturation of data was reached with three focus groups as the analysis of the third focus group did not add any new themes to the previous groups. After the derivation of the themes, the summaries of the identified themes during focus groups were sent to the participants to confirm their validity. The quotations of participants’ statements were translated from Croatian into English by one of the authors (IB) and checked for accuracy by another (AM). A set of quotations from the participants are presented here to illustrate the theme categories.

Ethical approval and consent

The study was approved by the Ethical Committee of University of Split School of Medicine as a part of the research grant “Professionalism in Health Care,” funded by the Croatian Science Foundation (Grant No. IP-2014-09-7672). All participants received the standard information and signed the standard consent form required by the Ethics Committee (available in Croatian at http://neuron.mefst.hr/docs/znanost/ured%20za%20znanost/Eticko%20povjerenstvo/Obavijest_informirani%20pristanak%20za%20ispitanike_Eti%C4%8Dko%20povjerenstvo.doc). At the beginning of each focus group they were additionally informed about the confidentiality of any information disclosed during the meeting, and were asked to avoid the use of identifying characteristics when describing their experiences. They were also informed that only the lead investigator, who performed all focus groups, will deal with identifying data, and that data will be anonymized for analysis.

Results

Fifteen individuals were contacted and agreed to participate but 2 subsequently cancelled due to work obligations, leaving 13 focus groups participants. The participants were doctoral students from two different doctoral program at Split School of Medicine (n = 7) or experts who had significant previous experience with research (n = 6). Six participants were professors at the USSM who already obtained their PhD degrees, while 3 participants were medical doctors teaching at the USSM and worked at the same time on their PhD thesis. There were 5 male and 8 female participants.

Three main topics could be derived from the focus groups discussions, providing the thematic framework for research integrity in a small scientific community (Figure 1): 1) different forms of unethical behaviors, 2) factors that influence unethical behavior, and 3) possible improvements.

Figure 1. Perceptions of research ethics related factors and ways for improvement.

Different forms of unethical behaviors

Increasing research waste

The opinion of the participants was that some researchers choose their research topic based solely on the prospect for publishing an article, meaning that no matter how bad the article is, or how unimportant, researchers will try to publish it.

P3: … (Unethical scientist is) the one who publishes only for publication`s sake.

This results in high number of publications on research topics that are already well explored, without adding new insights into the current state of art in the particular research field.

Non-publication of negative results

The fact that there are journals unenthusiastic to publish negative results may result in non-publication of the results despite the fact that the study had been performed properly. Non-publication of research may result in unnecessary repetition of research and waste of time and resources necessary for acquiring results.

P5: I heard that someone had a theory, and when the results did not show the results the principal investigator predicted; and were publishable in a journal with impact factor around 5–6, but not in the one desired by the researchers (with impact factors 20–25) the research was not published at all by the decision of the principal investigators. The problem is that huge amounts of time and resources were spent, but the results were in contrast with someone`s idea and were not published.

Non-publication of research data may happen when researchers are rigid and reluctant to adjust their expectations to the evidence from research results.

Authorship misuse

Guest authorship occurs often; this is not surprising due the fact that research publications are the measure of research success. In that way, people who did not participate in research are given the same credit as those that came up with a research idea and went through the entire research process. Furthermore, people who give undeserved authorship to their colleagues, often expect the same treatment in return from their colleagues.

P4: Unethical scientist is the one who signs many people on one paper, even though they have not participated in research and the one who asks the same in return.

P7: In general, everyone gives (undeserved) authorship to each other because they are going to have the benefits out of it.

The pre-registration of research projects does not help to solve this issue because, during the process or at the end of the research, it is still possible to assign credit to the additional people who did not deserve authorship.

Data manipulation

Whereas data fabrication is very rare, data falsification is more common. In order to achieve the desired level of significance, one may change the initially planned statistical procedure.

P6: If we performed multiple testing and the subsequent statistical analysis showed a decrease of thep value, we will stick to the one test we performed that showed the satisfactory level of significance.

One does not have to manipulate the data on their own, but can ask the collaborators to use the statistical test which will give the results they want, or even ask them to change the data a bit in order to get the results they wanted.

Sabotage and repression of collaborators

Even when the research procedures and standards are followed, a researcher may have an adverse effect on other’s people work, either by refusing to help them or by deliberately sabotaging them. Moreover, the higher the position of a professor, the more he or she may oppress younger colleagues.

P3: For example, in my opinion, it is unethical for the scientist who holds the tenure position to blackmail young colleagues with his demands in order to work on something he perceives has a scientific value.

Also, senior professors who are involved in decision-making regarding the acceptance of one’s thesis and advancements, may sabotage students by stalling the approval and in that way repressing the students and making them make changes to their topic, results, and interpretation for the professor’s benefit.

Factors that influence unethical behavior

Different definitions of ethics in research

Categorical perspective

This perspective emphasizes that ethical and unethical behaviors are clearly distinct categories. However, this view is more common as a public opinion rather than an actual practice.

P8: … That is the logical question that was answered by professor XX in his last interview. He says that, when it comes to cheating there is no “minor” or “significant” level of cheating because something is either cheating or it is not …

By this definition, unethical behaviors are those opposite to ethical ones, meaning everything opposite to the finding of the “truth,” a scientist`s fundamental task, is perceived as unethical behavior.

Dimensional perspective

The acceptability of behaviors in research practice differs greatly depending on the severity of the outcomes. This perspective sees research integrity as dimensional, with presence of grey areas. Unethical behaviors range from relatively minor violations to severe ones. Minor violations are relatively common and most times go unpunished. On the other hand, severe violations, auch as data falsification and data fabrication, are punished almost every time.

P10: … So there are no such severe violations (like data falsification), but there are other things, like false authorship. Although, when you look at the framework of ethics in science, which (of those two) has more severe consequences, that (data falsification) would have more severe consequences than false authorship.

As the behavior moves away from the prescribed codes of conduct and policy documents into real world and everyday behaviors, the grey zone between the absolutely ethical and absolutely unethical behavior expands.

Environmental factors

Model learning

Researchers work in institutions, surrounded by other people. By observing the behavior of others they understand what to do in order to achieve their goals. If some of their colleagues behave unethically and it goes unpunished; they may start feeling frustrated and less motivated for work.

P2: When you see that someone achieved a goal with little or no effort or in a wrong way, then others do not feel the desire to fight against it, but only frustration because you have realized that someone could achieve the result in that way.

The possible consequence of that process is that departments could often stand divided between “ethical” and “unethical” research groups, who have difficulties in cooperation.

Cultural context

Early on in their education, students learn that cheating is a way to achieve their goals. Many students cheat on their tests, and that pattern may continue when they perform research, because it is perceived that it is acceptable to achieve goals by bypassing the recommended procedures. Finally, the wider cultural context, which emphasizes the examples of people in power who succeeded by “not going by the book” amplifies this process, because it implies the best way to climb up the hierarchy.

P5: And just for the end, I think that the basic problem in science in our area is that we are taught to cheat since we were kids. Everyone cheats on their exams, everyone learns from old test questions, the cheating is present in all areas of society.

From this perspective, unethical behavior in science is only the symptom of a wider societal problem and a logical consequence of upbringing that teaches that cheating is an acceptable way to achieve goals.

Academic institutions as big systems with surveillance difficulties

An academic institution is a system that consists of a vast number of people who often work on different problems, in different areas, and using different methods. The larger the institution is, the harder it gets to supervise its employees. For that reason, it is very hard for the institution to control the behaviors of their employees. In the cases of unethical behavior, although institutiond provide the resources and organizational support for research, the responsibility is almost always put on the individual who had behaved unethically.

P4: And if an individual is behaving unethically, then I think that there is no institutional, or any kind of surveillance who did what on a certain paper and who contributed the most … And the bigger the institution is, the harder it gets.

There is no real punishment for the institution if their employees behave unethically, although the institution is the one which employed them and gave them resources for the work.

Motivational factors

Lack of proper punishment and reward system

In the cases of misconduct (except in cases of data fabrication and falsification), the violators are rarely punished. The consequence is that other researchers are less motivated and feel that doing research in a proper way is pointless. Successful researchers should be rewarded for their work, but instead research is done only to satisfy the criteria necessary for promotion.

P7: Those who are successful should be rewarded in the end with something: whether by a better position or something similar but that would be some kind of reward for their efforts. That never happens.

Finally, there is no acknowledgements for behaving ethically and doing research according to the standards, which could promote responsible behaviors. In the end, social influences often play the key role in the decision regarding promotions rather than research merit.

External motivation

Although there is no developed system of rewards and punishments, publishing is a way to gain money and status. In order to get academic or research promotion, one must publish articles. In that sense, publishing articles becomes a self-serving activity to get the academic promotion. That may be the reason why one would engage in unethical behaviors.

P3: In our case, people often want to obtain PhD because that will help you later in your career, both clinical and academic. That is the way it starts, and later on you just want to be acknowledged by the society.

Although it may seem paradoxical, in a system where there is no established reward and punishment system for scientific work, publishing is the way to obtain money and academic status. The reason for that is in low promotion criteria, which are strongly influenced by social processes.

Decrease of productivity after achieving goals

When a scientist achieves the desired academic position, the motivation for research decreases and results in low publication output. It becomes very difficult to persuade professors to mentor young students if they do not have any benefits from it. They are not motivated because they have already achieved their status and do not need more publications.

P3: . . . Here, when you obtain a certain position in science, then you are there, until you die. You do not have to publish anything.

Low criteria additionally support these behaviors, which may lead to an inverted academic pyramid, with the higher number of teachers with tenures than younger academics.

Personal factors

Not caring for consequences of misconduct

Researchers who participate in data falsification and authorship manipulation may not feel the guilt for doing damage. The lack of education about the consequences of research misconduct may lead to unethical behavior of researchers.

P7: I think that those people are aware that they are behaving unethically, but they do not care.

Scientist’s self-interest

Ego of a scientist is an important characteristic that may lead to unethical behavior. Whereas a certain level of ego may be desirable to maintain self-esteem and motivation, it may be counterproductive. Ego may be the reason for not accepting different opinions or sabotaging other’s work.

P3: Someone who is egocentric and wants to be on the articles will do anything in order to achieve that goal.

Lack of experience

Younger researchers are perceived as a vulnerable population for unethical behavior, whether by the “publish or perish” pressure or by being oppressed by their senior colleagues. Their lack of knowledge and experience may present a problem when trying to determine whether certain behavior is intentional or not.

P4: It depends how young you are. If you are a bit more experienced and you have already learned to observe people, it does not matter whether you encounter unethical scientists …

More experienced researchers may more easily resist the influence that is contrary to their previous work.

Suggestions for improvements

Within the institution

Research integrity education

Research integrity and responsible conduct of research should be a part of the curriculum from the early stage of higher education. Education should be oriented both towards theoretical and practical knowledge. Also, mentoring of student research would be important, because students, while learning about research, often test how far they may go, and mentors are usually those who define the boundaries.

P11: I think many young people will follow those (unethical) examples and, besides education, the important is the role of a mentor, too. Some of them will allow you to do anything, and some of them will control you for every little thing. The one who is told that something should not be done will not do it.

Young researchers should learn how to communicate with others in a responsible way, and not to be afraid to share their ideas with others.

Research integrity coordinator/officer/body

Academic institutions should appoint research integrity coordinators for research groups or establish a supervisory body at the institutional level. That person/body would supervise research processes within the institution, and give advice about ethical and legal issues researchers encounter during the research process and establish quality control check. Although certain procedures are already defined in the codes of conduct, they are not always followed. There should be a hierarchical system of responsibility in place, in order to decrease the number of violations. This system would need to be independent from direct academic hierarchy in order to avoid protecting those in the position of power and shifting blame to younger colleagues.

P7: Perhaps there should be someone at the institutional level. Maybe the institution should hire someone who will deal with (ethical) issues. In that way, all the articles that leave the institution could be checked before publication.

P8: Quality control system means that there is very precise description of procedures. If I report my colleague, then the person who holds the report “in the drawer” is in violation. So, everybody knows what needs to be done, whether to assemble the board or not. And the board will make the decision if that person deserves to be punished or not.

In order for this person or office to be effective, they should have real power, such as to stop manuscript submission to a journal in cases where there are serious problems with the manuscript. The quality assurance body should react to the allegations promptly, and the key is in the implementation part of the process, which should be fair and objective.

Data and process management

The institution should prescribe the standards in data and information management and make keeping the records of original data mandatory. Those procedures should be supervised on a regularly basis. The key participants in that process are young researchers who should learn how to perform research in a responsible and transparent manner.

P9: In general, from my own personal experience, you have to manage your data properly. If someone comes in tomorrow and wants to see your results, you must have them. I keep my results in electronic format, and I write my laboratory journal.

Transparency in research is something institutions could promote in their internal regulations, which must always be available to the integrity coordinator or the ethics board.

Outside the institution

External supervision of institutions

Institutions, i.e., their structures and processes for research integrity, should be supervised by independent international bodies. Those bodies should be independent and have power to advise and sanction the institutions who repeatedly violate ethical principles.

P6: And if the aim is to decrease unethical behavior, entire institution should be affected …

P13: … who would control that (unethical behavior)? Probably someone from the outside (of the state).

Accountability of institutions

The institution should take part in the accountability for unethical behaviors. If there are instances of unethical behavior in an institution, the institution should take a part of the blame for not putting enough effort to support their employees in responsible research conduct. Institutions give resources to their researchers, and expect from them to follow the code of conduct. However, in the cases of misconduct, the institution should be at least partially blamed for not protecting its assets.

P4: The affiliation of the scientist is important because research conducted in a respectable institution may be more valued because of the reputation of the institution itself.

Redefining excellence criteria

The main outcome measure of research productivity are scientific publications. Although there is a need for an objective measure of productivity, the downfall is that in most cases, the quantity outweighs quality. In that way, a very productive scientist does not necessarily have to be a good scientist.

P7: The number of published articles should not be considered more important than the quality of those articles. That is something what should be changed. Because it is not fair that someone has ten articles published in journals with an impact factor greater than 1, and someone who has one article on which he/she worked for two or three years and is published in a journal with a very high impact factor but will not be promoted because he/she has only one article. The biggest problem is that the focus is on quantity and not on the quality. And that should be regulated in a way that there is a decision whether one will have higher number of articles with lower IF, or higher number of articles with higher IF …

The quality criteria should also apply to researchers in tenure positions, because they are often not productive in research after getting the tenure.

Discussion

The qualitative analysis of problems related to research integrity at a research-oriented university medical school in a small scientific community revealed that there is a wide range of behaviors that are considered as unethical and are perceived to be caused by the interaction of environmental, personal, and motivational factors. The suggestions for improvement primarily revolve around the academic institution as the main generator of change, which may affect the ethical climate by implementing integrity policies. These perceptions may serve as useful basis for future research and hypothesis generation on research integrity issues in small scientific communities.

Perceptions of scientific misconduct presented by the participants of focus groups of postgraduate students and researchers from a small scientific community is similar to that described for other research settings, where previously published studies show that over 50% trials never get published (Schmucker et al. 2014), standard authorship guidelines are not followed (Marusic et al. 2016), and inappropriate use of statistical methods leads to an increase in “research waste” (Macleod et al. 2014). Moreover, undeserved authorship still presents an acute problem (Kornhaber, McLean, and Baber 2015) and everyday relations among academics may play a fundamental role in the occurrence and prevalence of research misconduct practices (De Vries, Anderson, and Martinson 2006). All of these factors may have adverse influence on the credibility and mutual trust among scientists.

Qualitative approach enabled us to explore the possible factors leading to research misconduct in more depth and detail than previous quantitative research (Pupovac and Fanelli 2015). The participants in our study considered the lack of perception of severity of the consequences for perpetrators as the potential cause for research misconduct. Financial incentives are considered to be potential catalysts for unethical behaviors, but also the absence of a functioning reward and punishment system. The groups which are potentially most at risk and are most likely to commit scientific misconduct are young scientists who lack experience, and scientists with egoistic traits, characteristics which have been confirmed using quantitative approach (Pupovac and Fanelli 2015; Tijdink et al. 2016). Ethical climate in an institution has shown to be predictive of research misconduct (Pupovac and Fanelli 2015). Our recent study, which identified that the University of Split School of Medicine has an ethical climate predominantly driven by self-interest (Malički et al. 2017), supports, at least in part, the relationship of the institutional climate and the perception of research misconduct. Overall, the ethical climate at the University of Split was rules-based, which can be expected in a system where, until recently, all schools were public schools and have to comply to national regulations in all aspects of academic work (Malički et al. 2017). When the overall environment in a small postcommunist country is open to cheating (Kukolja Taradi et al. 2010), it is not surprising that these culture-related behaviors influence the perception of research misconduct in academic institutions.

Previous findings indicated that ethical and unethical behaviors are not two polar categories because there is a graduation of severity of violations in scientific research, ranging from honest errors, over inappropriate research practices to outright fraud (Marcovitch et al. 2010). Our study emphasized the position of the institutions in this grey spectrum. We could hypothesize that the institutional grey area begins in policy documents, because it is not possible to predict all questionable behaviors and situations which may occur, and under which conditions. Most institutions do not have explicitly stated research integrity policies in their professional codes of conduct and they prescribe good research policies in an aspirational rather than in a normative, i.e., prescriptive tone (Komić, Marušić, and Marušić 2015). Finally, due to the wide spectrum of human behaviors and situations in which they may occur, the grey area between ethical and unethical behavior is very stretched. The incentive should come from institutions where the research is performed, along with international bodies, such as Committee on Publications Ethics (COPE) or All European Academies (ALLEA). The institution should raise awareness about research integrity both through education and through specific rules within the institution to ease the identification of inappropriate research practices. This is often difficult as policies do not consistently define the individual vs. institutional responsibility. For example, Croatian national Code of Ethics, written by the national committee for ethics in research and higher education, prescribes the responsibility of the institutions, stipulating that institutions are obliged to promote ethics in all aspects and to educate all participants about the rules and their rights and obligations (Croatian Code of Ethics 2006). On the other hand, the Code of Ethics of the University of Split School of Medicine emphasizes the responsibility of individuals and presumes that scientists understand and follow the prescribed rules (Code of Ethics 2009). The opinion of the participants in our study was that institutions might not enforce the rules prescribed or promote/educate the scientists on matters regarding research integrity in practice enough. Based on the identified issues, the sole existence of codes may not be sufficient and there should be more efforts to increase awareness about responsible conduct of research through education.

Current evidence, both from this study and other research (Edwards and Roy 2017), suggest that excellence criteria in research and academia require additional attention. On the one hand, research performance is continuously measured by publications and citations, a practice that, although inducing motivation, also leads to unethical behavior and research misconduct (Krstic 2015), while on the other, as our study suggests, establishment of low criteria sometimes may lead to the decrease in motivation when the research position is acquired. Current findings reveal that research productivity actually decreases in later phases of a research career (Sinatra et al. 2016), for which the possible solution, emerged from our data, is that research excellence should be measured in a more flexible way, differently for different stages of career. Participants agreed that, after obtaining the tenure, some people tend to decrease their productivity because they have accomplished their personal goals, although in they have more experience that could be used in research from which the society would benefit. The suggestion is that excellence criteria for tenured professors should take into account primarily the quality and impact of their work, while, for the promotion of young scientists, both the number of articles and the quality of their work should be taken into account.

The most important suggestion that emerged from the focus groups is that improvements in research integrity should start within the institution where the research is conducted. One of the proposals for improvement was to have an “integrity coordinator” or other body in the institution in order to prevent misconduct and help researchers to address problems and questions which may arise during the research process. Some institutions already reported the potential benefit of having an integrity coordinator or officer as a part of the research process and mediation in potential conflicts (Binder, Friedli, and Fuentes-Afflick 2016). Other steps include improving the communication between researchers, implementing research integrity education in curriculum and promotion of research integrity (Bouter 2015; Fanelli, Costas, and Lariviere 2015). The final suggestion that emerged from the focus groups in our study was that the institution should be held accountable in the cases of research misconduct of their employees. Besides bad publicity, institutions hold no responsibility for misconduct of their employees (Cosentino and Picozzi 2013). If institutions were to be sanctioned or the best practices in research integrity promoted as examples of good research practice, this may be the topping point towards real changes in research integrity environment in academia. However, it is not clear which kind of the oversight body would take on this task, particularly having in view the academic autonomy (Altbach 2001). In other countries, there are national bodies for research integrity that also have the oversight role, but in others such bodies, if they exist, have only advisory functions (Godecharle, Nemery, and Dierickx 2013). In the current situation of great varieties of bodies and practices, perhaps a good way forward would be to have public information on the frameworks for research integrity at institutional or country level. The proposal of Research Integrity Country Report Cards, proposed at the World Conferences on Research Integrity (Kleinert and Marušić 2016), is an example of how the elements of research environment and efforts to promote research integrity could be captured and best experiences shared.

Strengths and limitations

Our study has several limitations. Small sample size (only 13 people participated in the discussions) may give a narrow perspective on issues that surround research integrity processes and potential solutions to decrease the number of unethical behaviors. We tried to eliminate that issue by recruiting participants with different levels of research experience and holding separate session for young and experiences researchers to allow free exchange of opinions. Also, the thematic richness of the data and the saturation of themes over the three focus groups allowed us to draw a composite picture of how research integrity is perceived and responsible research practiced in a research-intensive academic institution from the so-called scientific periphery (Marušić and Marušić 2012). We also had gender balance in the focus groups, which reflected the overall gender structure of the academic medicine in Croatia (Puljak, Kojundzic, and Sapunar 2008), and allowed us to explore any gender-specific issues related to research integrity. We did not include policy makers who might have provided a different perspective on solutions and barriers to research integrity promotion in institution. Finally, all focus group participants were members of the same institution, in a single medical school in Croatia, which might make our results not generalizable to other settings, but may be relevant to many small scientific communities in the region and globally. Our focus groups included participants who spent time doing research in other countries, so they were familiar with different approaches to ensuring research integrity and addressing unethical behaviors.

Implications for further research

Based on the results of our qualitative study, research misconduct in a small scientific community is perceived to be the consequence of the interaction of several social and psychological factors. The possible improvements should be systematic, aiming both for improvements in work environment and personal awareness in research ethics, and those changes should be implemented in research institutions.

For further research, the suggestion is to approach stakeholders other than young and experienced researchers in environments with high and low pressure for publishing, in order to gain insight about the experiences and perceptions for promoting responsible conduct of research and prevent misconduct. Furthermore, it is necessary to test our hypotheses that institutions with better defined regulations have fewer cases of research misconduct using quantitative approach.

Additional information

Funding

This research was funded by the Croatian Science Foundation, grant No. IP-2014-09-7672 “Professionalism in Health Care”). The funder had no role in the design of this study, during its execution and data interpretation; Hrvatska Zaklada za Znanost [IP-2014-09-7672].