What are the best practices? (Has Best Practice)
From The Embassy of Good Science
[
Digital Curation Centre (DCC) has issued a checklist containing relevant questions for researchers concerning criteria for their selection of repositories. However, it has emphasised that while it provides “a shortcut through relevant policies and standards”, it does not guarantee that these repositories comply with specific standards or funders’ policies. Therefore, this guide provides a framework to help researchers with what can be a somewhat difficult decision. By going through these questions, researchers should be able to make a well informed decision on whether their repositories are suited for their data. They should also always check funder and journal policies for compliance.
The checklist contains five questions to consider when looking for repositories:
-Is a repository of your choice reputed?
-Will it take the data you want to deposit?
-Will it be safe legally?
-Will it sustain the data value?
-Will it support analysis and tract data usage?'"`UNIQ--ref-0000012F-QINU`"'
This guide also contains information on where researchers can find repository registries, such as Re3data and Biosharing.'"`UNIQ--ref-00000130-QINU`"'
'"`UNIQ--references-00000131-QINU`"' +
'''Planning stage of the clinical trial'''
To reduce the risks of missing data in the panning of the clinical trial, statistical analyses should be specified, key data items should be identified, and the procedures to prevent missing data '"`UNIQ--ref-00000250-QINU`"'.
'''Analysis stage'''
There are different strategies to deal with missing data that will depend on the specific clinical trial and type of missing data:
1. Complete cases analysis could be used when the proportions of missing data are below 5%and the potential impact of the missing data is negligible '"`UNIQ--ref-00000251-QINU`"'.
2. Single imputation replaces missing values by a value defined by a certain rule. However, this method ignores the data variation and can potentially introduce bias and should be used with great caution '"`UNIQ--ref-00000252-QINU`"'.
3. When the missing data accomplish certain characteristics, multiple imputation may be used to minimize bias . Missing values are replaced by a random sample of plausible values imputations. There are several multiple imputation methodologies that must be chosen according to the variable with missing values '"`UNIQ--ref-00000253-QINU`"'.
To conclude, handling missing data validly is an important, yet difficult and complex, task. This theme showed different strategies to handle missing data but always statistical expertise’s advice is needed.
'"`UNIQ--references-00000254-QINU`"' +
To support wide data availability, authors should publish their data so it could provide inputs to new hypotheses, and innovation (5,6). Journals could increase awareness on data availability and request mandatory data deposition. Modern manuscript checklists should routinely include data availability part which should additionally emphasize its importance to the authors. Finally, all of this could improve the verification of result, and reduce the misconduction related to data fabrication and falsification, and could serve as training tools for junior researchers (6). +
Some journals, such as Nature, require depositing data to data repositories as part of the manuscript submission process. This is one of the prerequisites for publication.'"`UNIQ--ref-000000BB-QINU`"' Nature has set out certain criteria for data repositories. They should:
-provide long-term preservation of data (at least 5 years after publication)
-be supported by a research community or institution
-provide deposited datasets with stable and persistent identifiers
-allow open access
-provide terms of data use
-facilitate for anonymous reviewers to access data under embargo.'"`UNIQ--ref-000000BC-QINU`"'
The journal also offers a list of repositories across research areas for researchers who are not certain where to deposit their data:
-[https://www.nature.com/sdata/policies/repositories#:~:text=Generalist%20repositories-,Biological%20sciences,-%E2%A4%B4 Biological sciences]
-[https://www.nature.com/sdata/policies/repositories#:~:text=Generalist%20repositories-,Biological%20sciences,-%E2%A4%B4 Health sciences]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20FAIRshaing%20entry-,Chemistry%20and%20Chemical%20biology,-%E2%A4%B4 Chemistry and Chemical biology]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20FAIRsharing%20entry-,Earth%2C%20Environmental%20and%20Space%20sciences,-%E2%A4%B4 Earth, Environmental and Space sciences]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20FAIRsharing%20entry-,Physics,-%E2%A4%B4 Physics]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20re3data%20entry-,Materials%20science,-%E2%A4%B4 Material science]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20FAIRsharing%20entry-,Social%20sciences,-%E2%A4%B4 Social sciences]
-[https://www.nature.com/sdata/policies/repositories#:~:text=view%20re3data%20entry-,Generalist%20repositories,-%E2%A4%B4 Generalist repositories].'"`UNIQ--ref-000000BD-QINU`"'
'"`UNIQ--references-000000BE-QINU`"'
AI has been applied in diverse research areas, from analyzing medical images and conducting statistical analyses to simulating complex climate change models. The transformative potential of AI is undeniable, but the responsible use of AI in research is already shaping best practices across disciplines to mitigate associated risks:
# Bias Mitigation: Ensuring fairness in AI models involves using diverse and representative datasets to minimize biases. Researchers can also adopt bias-detection tools to identify and address potential discriminatory outcomes, fostering more equitable results across various populations.
# Transparent AI Use: Transparency in AI applications requires clear documentation of how AI is integrated into research processes. Providing detailed explanations promotes reproducibility, accountability, and trust among the scientific community and the public.
# AI-Assisted Writing: When using AI tools for drafting or creating content, it is essential to maintain human oversight, disclose the use of AI, and ensure that the work remains original. Responsible use of AI in writing helps uphold integrity and authenticity in research outputs.
# Data Privacy Protections: Protecting data privacy is critical when using sensitive or personal information to train AI models. Researchers should adhere to stringent data protection regulations, such as encryption techniques and compliance with privacy standards, to safeguard participant confidentiality and maintain ethical practices.
Despite these positive practices, AI misuse also occurs:
# AI-generated content has been submitted to journals without proper acknowledgment, raising concerns about plagiarism and authorship integrity.
# Researchers have unintentionally perpetuated bias by relying on poorly curated or non-representative datasets, which can lead to skewed conclusions and reinforce harmful stereotypes.
To address these issues, guidance documents offer researchers concrete recommendations for the responsible integration of AI into their work. Educational initiatives, such as workshops and online courses on AI ethics, are also gaining traction to ensure that researchers are equipped with the knowledge necessary to use AI responsibly and ethically.
In their 2022 report on the problem of paper mills, COPE and STM concluded that the challenge posed by paper mills can only be overcome through urgent collective effort from all stakeholders in research.'"`UNIQ--ref-00000102-QINU`"' They made a number of recommendations for long-term actions that can be undertaken by different stakeholders in research:
*'''For Editors:''' ensure editors and editorial staff are aware of the problem of paper mills, and train them to identify them as they are submitted.
*'''For Institutions and Funding organisations:''' address the perverse incentives involved in academic publication so that researchers do not feel the need to use paper mills. Avoid inappropriate criteria relating to publications for employees and funding recipients working in non-research roles.
*'''For Publishers:''' review the retraction process to take into account the unique features of paper mills, and investigate options for ensuring that retraction notices are added to all available versions of a paper. Continue investment into tools and systems to detect paper mill papers as they are submitted.
*'''For all:''' Investigate protocols that can be put in place to prevent paper mills from succeeding in their goals.'"`UNIQ--ref-00000103-QINU`"'
Since the publication of the COPE and STM report, there has been further action taken towards achieving the collective effort described above including the formation of United2Act - a joint venture between academic publishers, research organisations, and funders, like the European Research Council. In January 2024, the group released a statement describing 5 key areas of action they will take. These include:
#Improve education and awareness of the problem.
#Conduct detailed research into paper mills.
#Improve post-publication corrections.
#Support the development of tools to verify the identities of authors, editors and reviewers.
#Ensure that the groups across publishing that are tackling the issue communicate.'"`UNIQ--ref-00000104-QINU`"'
'''For Journals'''
Scientific journal Naunyn-Schmiedeberg’s Archives of Pharmacology was targeted with papers produced by paper mills in 2020, resulting in the retraction of 10 scientific papers in 2021.'"`UNIQ--ref-00000105-QINU`"' Following this experience, the journal will adopt the following measures:
*Request of institutional email addresses. The corresponding author should provide an institutional email address.
*Request of supplemental original source data (raw data, original data, individual data points) presented in tables and figures in a generally readable format. Excel files are preferred. Pdf and Prism files are acceptable as well. Supplemental data must be cited in the main text. These data will be made available to the reviewers and published if the paper is accepted.
*Request of supplemental immunoblot data. Specifically, full-length immunoblots with molecular mass markers are requested. Supplemental immunoblot data must be cited in the text. These data will be made available to the reviewers and published if the paper is accepted.
*Authors must include the following statement in the section “Authors Contributions”: The authors declare that all data were generated in-house and that no paper mill was used.'"`UNIQ--ref-00000106-QINU`"'
Papers that do not meet these requirements are automatically rejected, but can be accepted later if the guidelines are met.
These represent just some of the ways that scientific journals can try to tackle this problem.
'"`UNIQ--references-00000107-QINU`"'
The Dutch funding agencies ZonMw and NWO have set up guidelines for dealing with COIs in the reviewing process. COI citeria include:
*personal interests
*professional interests
*interests arising out of other positions
*business (financial or economic) interests
The following personal interests always exclude participation in the process:
*being the applicant or joint applicant
*having written any part of an application without being an applicant or joint applicant
*having any of the following relations with the applicant or joint applicant
*actual or anticipated project manager or sub-project manager
*direct manager
*blood relation or affinity (up to and including the third degree
*contract of cohabitation (other than a civil partnership
*dean of the applicant’s institute'"`UNIQ--ref-00000000-QINU`"'
Both ZonMw and NWO ask reviewers to declare their COIs. Next, the head of the committee assesses the impact of the declared COIs. He or she can decide if the reviewer cannot assess a particular application, or if he or she should not partake in the review process at all. Subsequently, the committee states whether they agree or disagree with this decision. Moreover, to ensure impartiality, individual reviewers do not get to see each other’s scores. After the review process has concluded, the reviewers assess whether the process has raised new insights into COIs. In the worst case scenario, if the process has been influenced by a reviewer’s COI, the whole review process is redone.
'"`UNIQ--references-00000001-QINU`"' +
When research is sponsored by industry stakeholders this should be declared as a conflict of interest. Moreover, in a systematic scoping review Cullerton, Adams, Forouhi, Francis & White identified themes for good collaboration with industry stakeholders (specifically in the food industry).'"`UNIQ--ref-000000A7-QINU`"' In [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6767600/figure/obr12851-fig-0003/ this flowchart] the auhtors describe several steps and actions sthat can be taken to ensure good collaboration between researchers and industry stakeholders.
'"`UNIQ--references-000000A8-QINU`"' +
According to their policy on IP rights, the Dutch funder ZonMw states that when a contract is written, the IP rights derived from a project belong to the research performing institution. However, when the IP rights are not exploited adequately, ZonMw can interfere, in order to increase the impact of the research. Moreover, when considering public-private partnerships in funding, the research performing organization needs to have a strong judicial position. Therefore, research institutes are encouraged to seek legal expertise or request advice from ZonMw. This is to ensure that the IP is not exploited by other private, commericial and industrial institutions, as has happened in the past.
The European IPR Helpdesk has developed a fact sheet on IP management for Horizon 2020 projects.'"`UNIQ--ref-0000001B-QINU`"' The factsheet includes guidance on IP rights in a research proposal. It suggests the following:
*Proposals should describe how results will be protected from commercial and industrial exploitations
*Within multi-center research, the involved institutes need to organise ownership and access of rights and include the economic conditions of dividing the rights
*Proposals should state how joint ownership for expected jointly-owned results will be organized. For example, third-party licensing needs to be considered
*Proposals should consider the exploitation of research results (see Figure 1)
*Proposals should indicate the confidentiality measures that need to be put in place
*“When defining the management structure, applicants should consider the management of intellectual property rights. A committee for exploitation and intellectual property or a manager could be considered and even a consulting body of external experts from industry.”
'"`UNIQ--references-0000001C-QINU`"' +
==Bullying and harassment policies==
Bullying and harassment policies allow RFOs to stimulate positive research cultures. Such policies can improve research culture, and their existence “sends a signal that certain ethical standards must be met by researchers and organizations in exchange for funding”.'"`UNIQ--ref-000001D0-QINU`"' The US based funder National Science Foundation (NSF) requires RPOs receiving funding to inform the NSF about sexual harassment.'"`UNIQ--ref-000001D1-QINU`"' In addition, the Wellcome Trust has elaborate rules on what they expect from the organizations they fund (7):'"`UNIQ--ref-000001D2-QINU`"'
1. The funded organization requires policies that set out:
*standards of behavior from staff
*the procedure for responding to complaints
2. The funded organization should have an equivalent policy in place at sub-levels, where relevant.
3. The funded organization should investigate allegations of bullying and harassment in an impartial, fair and timely manner. It must:
*protect the rights of all employees involves
*take appropriate action.
4. The funded organization should contact the Wellcome Trust when an investigation has been opened.
5. The funded organization should contact the Wellcome Trust when the investigation has been completed.'"`UNIQ--references-000001D3-QINU`"' +
Managing people, projects, and budgets, as well as setting workplace procedures that foster efficiency, safety ethics, and high staff morale, are all issues that come with running a laboratory. Lab safety and ethics are two of the most important components of running a lab, yet they are often overlooked in favor of more fundamental lab management skills. It is critical to work hard to build a laboratory culture that prioritizes safety. +
To manage conflict of interests, which is inevitable and must always be declared, it is important to maximize transparency in industry-funded research. Some of the proposed approaches include a carefully constructed advisory board with specific responsibilities as well as strict guidelines to protect research quality.'"`UNIQ--ref-0000028C-QINU`"' It is important to fully disclose financial, personal, and professional interests. Oversight by an independent third party is recommended as well as pre-registration of the study. The role of commercial partners must be defined and the relationship has to be continuously monitored. '"`UNIQ--ref-0000028D-QINU`"' The scientific community has improved recently in regards to some of the key indicators of research reproducibility and transparency. This includes information on funding, conflict of interests statements, and an increase in protocol and data sharing. There is also a growing number of fully available texts, although the majority of texts are still not fully available. Nevertheless, confusion about research transparency still exists, and even though there has been some improvement, the scientific community is far from perfect in terms of research transparency and reproducibility.'"`UNIQ--ref-0000028E-QINU`"'
'"`UNIQ--references-0000028F-QINU`"' +
'''Best Practices for Ethical Research with Children'''
# '''Use child-friendly language''' – Explain the study in a way the child can understand.
# '''Provide visual aids''' – Diagrams, videos, or interactive materials can help comprehension.
# '''Offer the right to say no''' – Children should know they can refuse participation without consequences.
# '''Train researchers''' – Ensure that study personnel understand ethical considerations in pediatric research.
# '''Ensure ongoing consent''' – Regularly check with both parents and children throughout the study. +
The best practices is the first step in the targeting the right population of the patients that will be included in research. Also, it is important to give them the informed consent at the beginning of the research, before any procedure. +
'''How it works''' The session takes place annually in Alzheimer Center Amsterdam. All researchers are obliged to attend. Some preparation from the senior researchers in advance of the session is expected: they are asked to share an example of ‘sloppy science’ or a research integrity dilemma drawn from their personal experience. The session starts with a short introduction and is opened by the head of research. Senior researchers share their stories first, to show that issues happen to everybody, and it’s okay to talk about it. This is key to create an environment where it feels safe to speak about issues. Hearing about the issues and mistakes from the seniors stimulates junior researchers to talk more openly about the obstacles they have encountered in their research projects. The goal is not to discuss all issues during these particular sessions, but rather to keep researchers from feeling afraid or embarrassed to discuss issues during everyday research practice all year round. The session ends with educating the junior researchers about who they can talk to during the year if issues arise, as well as sharing the contact info of the hospital ombudsman.
'''Evaluation''' Whereas there has not yet been an empirical evaluation of the initiative, so far, the responses from researchers are very positive. The experience is that issues within the center are discussed timely. Researchers tend to speak first to people in their close working environment, whom they trust. Since the ‘Billen Bloot’ meetings are embedded in weekly scientific sessions, are costless, take place in the center itself and are mandatory to attend, the feasibility of the ‘Billen Bloot’ sessions is very high. This feasibility allows for a smooth implementation and therefore, it makes it a very attractive initiative.
The success of the formula in Alzheimer Center Amsterdam can partly be explained by the already open environment. It is possible that if this initiative is implemented in a less open work environment, it may be harder to make it a success.
'''What’s next?''' The ‘Billen Bloot’ sessions were initiated within the scope of the Alzheimer Center Amsterdam. However, in the last two years the initiative has been copied by other departments and institutions. Finally, researchers of Alzheimer Center Amsterdam have given presentations on the initiative to share the idea with others.
In the United States, the Office for Research Integrity (ORI) is developing policies, procedures, and regulations related to the responsible conduct of research and research misconduct. Moreover, ORI develops activities and programs aimed to promote research integrity and foster good research practices.
In Europe, the [https://lari.lu/ Luxembourg Agency for Research Integrity] and [https://oeawi.at/en/ Austrian Agency for Research Integrity] are good examples of institutions providing all kind of resources for the promotion of research integrity. These organizations have developed guidelines and recommendations that are implemented by research institutions in their countries. Moreover, they offer training and workshops for researchers in different stages of their career and deal with cases of research misconduct. +
The aim of all eight scenarios is to allow researchers, research ethics committees ('RECs'), research integrity offices ('RIOs') and research administrators to focus their reflection on core principles and research contexts that enshrine good research practice as well as their local rules and practices. +
The European Commission funded VIRT2UE project has designed a set of practical tools for recognizing and dealing with RI dilemmas in practice. One of these tools is the so called ‘Virtues, Values and Norms in Perspective’ exercise. In this exercise, inspired by a [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5236068/ virtue ethics approach to RI education], '"`UNIQ--ref-00000099-QINU`"' participants are asked to reflect on a concrete dilemma experienced by one of the participants. By engaging in a dialogue, trainees are asked to put themselves in the shoes of the people involved in the case at stake and reflect on which values would be important for each of them. Then they are asked to reflect on which virtues are required to deal with the dilemma at stake: are they in conflict with each other? How are they related to salient norms? Which virtue is the most important in this situation? What is needed (concretely) to act in accordance with that virtue?
This exercise aims at fostering reflection in oneself by means of understanding the concepts of virtues/values and norms and their relationship with practice.
Another initiative is the [https://www.eur.nl/sites/corporate/files/24708_integriteitsspel_interactief_2016.pdf Rotterdam Dilemma Game], '"`UNIQ--ref-0000009A-QINU`"' which focuses less on virtue ethics but nonetheless provides structure to discuss RI dilemmas.
'"`UNIQ--references-0000009B-QINU`"' +
To address the issues stated above, the REWARD campaign offers five set of recommendations. First, research priorities need to be selected. This means that prior to engaging in any experimentation or observation, a systematic review of literature should be done, funders should make their criteria for funding transparent, and research funders should develop sources of information about ongoing research.
Second, the REWARD campaign supports rigor in research design, conduct and analysis. This includes improving protocols and making them public, engaging professional staff and non-conflicted stakeholders, as well as rewarding quality and reproducibility of the research, rather than novelty.
Third, research waste can be reduced through research regulation and management. It emphasizes the important role and influence of research regulators who, along with policy makers, should collaborate with researchers, patients and health professionals. Their cooperation should simplify and coordinate laws and regulations that control or guide research. Furthermore, researchers should be able to improve the efficiency of their research using high quality designs that reduce inefficiencies of recruitment, retention, data monitoring, and data sharing. Finally, the integration of research findings in daily clinical practice should be promoted.
Fourth, all information on research methods and findings should be accessible. Academic institutions and funders should reward research that is publicly available and disseminated. REWARD recommends the standardization of protocols and data sharing, as well as the release of complete study reports. Finally, journals, funders, sponsors, research ethics committees, regulators and legislators should support and enforce study registration and availability of complete information regarding research.
Fifth, research reports should be complete and usable. With that aim, the REWARD campaign suggests that the focus of research regulations and rewards should be to encourage complete research reporting. To facilitate this, there is a need for good reporting infrastructure. Authors, editors and reviewers should be trained about reporting guidelines, publication ethics, and research integrity. Funders should support and monitor training.
Researchers should be up-to-date in their field of interest and, when they notice a retraction of an article that they have previously cited, correct the article. The easiest way to be regularly updated on retractions is by following Retraction Watch and their database '"`UNIQ--ref-000000A4-QINU`"'. Zotero citation manager has established a partnership with Retraction Watch and has implemented retraction notifications that pop-up when an article from the users’ database has been retracted. Hopefully other citation managers will follow this practice.
An initiative to stimulate this kind of behavior could result in more corrected articles. In practice, taking into account the number of articles that are published every day, it is hard to expect an individual to notice everything. The ideal practice would be that the journal which has retracted the article, notifies authors which have cited the retracted article. However, that is hard to be expected, especially for older articles. Alternately, authors of the retracted article could inform all the authors who have cited their article. This may be expected from authors whose article is retracted due to unintentional mistake and have initiated the retraction, but it might be illusory to expect this from authors who have committed fabrication, plagiarism, or similar misconduct.
'"`UNIQ--references-000000A5-QINU`"' +
