Why is this important? (Important Because)
From The Embassy of Good Science
A description to provide more focus to the theme/resource (max. 200 words)
- ⧼SA Foundation Data Type⧽: Text
[
Duplicate publications are redundant because they present the same data. As readers are not made aware of this fact, they could consider the two (or more) publications to be separate studies. This may be harmful when such studies are included in evidence synthesis, such as systematic reviews and meta-analyses about health interventions, because they may skew the evidence.'"`UNIQ--ref-00000005-QINU`"' This can lead to erroneous recommendations for health practice, which may result in increased risk for patients.
'"`UNIQ--references-00000006-QINU`"' +
The term “funding effect” was formed in the 1980s, when it was discovered that researchers do not always report their results and conclusions honestly, but distort the findings to support the aims of the funding sources.'"`UNIQ--ref-0000000B-QINU`"' It has also shown that industry sponsored studies are more likely to publish positive results than those sponsored by not-for-profit and independent organisations.'"`UNIQ--ref-0000000C-QINU`"' One of the reasons is that funders may prevent publishing of the manuscript containing unfavourable results or even take some legal acts against researchers.'"`UNIQ--ref-0000000D-QINU`"' For example, in one case a sponsor sued a haematologist for breach of contract because she reported “concerns over the safety of a drug she was evaluating”.'"`UNIQ--ref-0000000E-QINU`"'
This distortion of reporting of the study results and conclusions occurs more often in studies funded by the pharmaceutical industry.'"`UNIQ--ref-0000000F-QINU`"' This can cause serious consequences because it directly affects the practice of medicine.'"`UNIQ--ref-00000010-QINU`"' Nevertheless, every researcher with a funded study could be pressured to report results and conclusions that are more favourable to the funders, regardless of the topic or research area of the study.
'"`UNIQ--references-00000011-QINU`"' +
In recent years, several academic disciplines – perhaps most strongly psychology – have witnessed a replication crisis, casting doubt on the validity of seemingly well-established findings and theories. The prevalence of HARKing, evidenced by empirical studies of research misconduct and research misbehavior, is considered one of several driving factors behind the replication crisis. This is due to two reasons: Firstly, data that were used to generate a hypothesis cannot be used to test that same hypothesis in any meaningful way. Portraying an empirically inspired post hoc hypothesis as a priori violates the falsification principle crucial for hypothesis-driven (that is, confirmatory) empirical research. Secondly, suppressing unsupported a priori hypotheses and the attendant failure to report null effects throws away opportunities to cast doubt on the validity of hypotheses derived from extant knowledge. Consequently, multiple disconfirmations of a hypothesis may go unnoticed when researchers HARK because disconfirmations are not reported.
While pure HARKing is unquestionably a detrimental research practice, the same is not necessarily true for transparent forms of HARKing. Transparent HARKing (THARKing) occurs when researchers develop post hoc hypotheses (that is, hypothesize after the results are known), but do so transparently and based on theory. If done transparently and inspired not only by results but also by theory, post hoc hypothesizing does not misportray the research process because exploratory findings are clearly labeled as such. +
Improper data use undermines the ethos of science and the corresponding misleading results can misguide and distort the production of knowledge.
Examples of improper data use include:
*'''Massaging''': … extensive transformations or other maneuvers to make inconclusive data appear … conclusive
*'''Extrapolating''': … predicting future trends based on unsupported assumptions …
*'''Smoothing''': discarding data points too far removed from expected … values
*'''Slanting''': … selecting certain trends in the data, … discarding others which do not fit …
*'''Fudging''': creating data points to augment incomplete data sets …
*'''Manufacturing''': creating entire data sets de novo, … '"`UNIQ--ref-0000000A-QINU`"'
Data dredging is looking for too many possible associations in a dataset to see of any of them are statistically significant. Data dredging results in false positive results.
“When a large number of associations can be looked at in a dataset where only a few real associations exist, a P value of 0.05 is compatible with the large majority of findings still being false positives.” '"`UNIQ--ref-0000000B-QINU`"'
'''Origin of words'''
There are several terms describing the act of data dredging. These include:
<br />
*"Data Dredging"'"`UNIQ--ref-0000000C-QINU`"'
*"Data Fishing"'"`UNIQ--ref-0000000D-QINU`"'
*“Data Snooping,”
*“P-hacking”
'"`UNIQ--references-0000000E-QINU`"' +
The media is the most powerful tool for spreading new information about scientific studies. Therefore, researchers and academic institutions place a strong emphasis on communicating research findings to the media '"`UNIQ--ref-00000016-QINU`"'. However, sometimes the media distort research findings, which results in the spreading of misinformation '"`UNIQ--ref-00000017-QINU`"'. Frequent use of oversimplified language, exaggeration, sensationalist reporting and the avoidance of complex issues are some of the main reasons for the misrepresentation of researching findings by the media '"`UNIQ--ref-00000018-QINU`"''"`UNIQ--ref-00000019-QINU`"''"`UNIQ--ref-0000001A-QINU`"'. Furthermore, when follow-up studies undermine the results and conclusions of the initial study, the media usually does not correct or supplement its previous reports or provide a new report altogether '"`UNIQ--ref-0000001B-QINU`"'.
Sensationalist reporting of medical research is not rare and can have serious consequences. It can raise false hopes or generate needless fear '"`UNIQ--ref-0000001C-QINU`"'. However, the media isn’t solely responsible for misleading the public about research findings. In 2014, research showed that the majority of university press-releases, usually approved by the lead researcher, tend to exaggerate research findings. Because reporters rely on these press-releases, inaccurate information is more widely disseminated '"`UNIQ--ref-0000001D-QINU`"'. Therefore, in order to improve reports of research findings and their presentation to the public, close collaboration between researchers and journalists is essential '"`UNIQ--ref-0000001E-QINU`"'.
'"`UNIQ--references-0000001F-QINU`"' +
In early 20<sup>th</sup> century, the concept of a P-value was introduced, along with a decision rule that stated that if p<0.05, then the null hypothesis should be rejected '"`UNIQ--ref-00000018-QINU`"'. In other words, when a P-value is less than 0.05, the results are regarded as “statistically significant” '"`UNIQ--ref-00000019-QINU`"'.
Journals widely encourage the use of the method of inferring from P-values for publication, which puts researchers under a lot of pressure to publish “statistically significant” results '"`UNIQ--ref-0000001A-QINU`"'. According to recent findings, 96% of abstracts and full-text articles in the biomedical literature from 1990 to 2015 presented p<0.05, which is considered “too good to be true”, and indicates that there is a practice of selective reporting '"`UNIQ--ref-0000001B-QINU`"'.
Developments in decision theory, information theory, mathematical modelling and computing in the second half of the 20<sup>th</sup> century shed a completely different light on the use of P-values and statistical inference in general '"`UNIQ--ref-0000001C-QINU`"'. By 2016, mounting criticisms of the use and interpretation of P-values prompted the ASA to publish a policy statement '"`UNIQ--ref-0000001D-QINU`"'.
'"`UNIQ--references-0000001E-QINU`"' +
There is no such thing as perfect research. Every study, whether it is experimental or observational, has its limitations and deficiencies that can influence outcomes of the research '"`UNIQ--ref-00000017-QINU`"'. Not reporting them properly is dangerous, because it could lead to over-confidence in flawed findings '"`UNIQ--ref-00000018-QINU`"'. This, consequently, could reflect negatively on trust between scientists and decrease the public trust in science '"`UNIQ--ref-00000019-QINU`"'. Different types of limitations that can occur during the research process in the first place are related to sample size, methodology, lack of resources and time constraints '"`UNIQ--ref-0000001A-QINU`"'.
Since limitations are a natural part of the research process, they should be fully reported and described. Reporting flaws and limitations shows that a researcher fully understands the topic '"`UNIQ--ref-0000001B-QINU`"', informs and gives readers an opportunity to detect existing gaps in research, '"`UNIQ--ref-0000001C-QINU`"' and provides them with a possibility to identify new research questions '"`UNIQ--ref-0000001D-QINU`"'. Finally, presenting study limitations also shows a researcher’s honesty and transparency '"`UNIQ--ref-0000001E-QINU`"''"`UNIQ--ref-0000001F-QINU`"'. Knowing limitations and deficiencies of a certain study is an essential tool which helps to modify every decision based on considerations of risks and benefits.
'"`UNIQ--references-00000020-QINU`"' +
Research in biomedical sciences reveals that positive results have a higher chance of being published. Because of that, negative results (for example, the lack of effect of some therapy) might be unavailable to the scientific community. Consequently, when other researchers conduct systematic reviews and meta-analysis, the results are distorted in favor of the positive finding. Clinical trials with negative results, and those with reported serious side effects, often don’t get published, which is dangerous, unfair to participants, and a waste of resources. '"`UNIQ--ref-0000000B-QINU`"' Getting negative results in a costly project and after a lot of hard work can be very demotivating, disappointing and can negatively impact young researchers’ careers. Some supervisors may not be happy to publish negative results, and in that way add to the climate of positive-publications-only.
'"`UNIQ--references-0000000C-QINU`"' +
Failing to report all aspects of research leads to publishing incomplete or inaccurate results. It wastes valuable resources such as peer-reviewers’ and editors’ time, research funds, and collaborators’ efforts. Being biased in reporting results distorts the integrity of science and if not discovered, might influence future studies.
Outcome reporting bias affects the entire research ecosystem. From research subjects and collaborators to research administrators, funders, and also other researchers that might rely on the results of the study. It also affects society’s trust in science. +
Unfortunately, current practices in science show that journals that are considered of high quality (those with high impact factors) predominately publish statistically significant results. Researchers want to publish in such journals because it's important for their academic prestige and job.'"`UNIQ--ref-0000000C-QINU`"' This creates pressure on researchers, and can lead to P-value hacking. P-value hacking leads to false positive results, which can get published, and have a negative impact on future research in the field, secondary research and systematic reviews and human knowledge in general.'"`UNIQ--ref-0000000D-QINU`"'
'"`UNIQ--references-0000000E-QINU`"' +
Publication bias is defined as a conscious or unconscious decision to publish or distribute a manuscript based on the study results . '"`UNIQ--ref-00000006-QINU`"' Reviews of scientific literature point to the fact that papers with positive results are three times more likely to be published than the ones with negative results. Reasons for such practices are multiple. The most common one is that scientists simply do not submit studies with negative findings to journals. Unfortunately, a lot of negative findings are usually defined as those that do not reach the usual threshold of statistical significance of p<0.05. This issue was recently addressed in the American Statistical Association’s Statement on p value, which aims to steer future research away from p value as the most significant indicator. '"`UNIQ--ref-00000007-QINU`"'Also, a part of responsibility lies with editors and scientometric criteria, which, by their nature, favor studies with positive results. Two dire consequences of publication bias are resource waste and negative impact on meta-analyses. The former is the result of scientists doing experiments which have already been conducted by others but never published (because they had no positive results), and the latter simply means that meta-analyses will be oversaturated with positive results and skew the conclusions . '"`UNIQ--ref-00000008-QINU`"' '"`UNIQ--ref-00000009-QINU`"' '"`UNIQ--ref-0000000A-QINU`"'
'"`UNIQ--references-0000000B-QINU`"' +
Salami publication is a concept that is difficult to define, therefore making detection and prevention difficult, but it is generally considered to be a form of redundant publication and self-plagiarism characterized by the spreading of study results over more papers than necessary despite the same, or very similar, hypothesis, methodology, dataset or results. '"`UNIQ--ref-00000007-QINU`"''"`UNIQ--ref-00000008-QINU`"' The negative consequences of salami publication are multiple, and can be divided into two groups. The first is of a scientometric nature – scientists with more papers are likely to get more citations and probably more funding. The second, more serious, consequence is that results will be over-represented in meta-analyses, which are considered to be the highest level of evidence for any question '"`UNIQ--ref-00000009-QINU`"' – salami publication skews the results of meta-analyses because the same data is unknowingly analyzed twice.
'''Examples'''
The most blunt example of salami publication is publishing the same paper twice, with slightly different conclusions '"`UNIQ--ref-0000000A-QINU`"'.This type of salami publication was much more likely to occur in the age before online databases – nowadays, salami publication is much more subtle. For example, studies which investigate levels of biomarkers in different phases of a disease end up being followed up by a different paper investigating diagnostic characteristics of those very same markers on the same datasets.
'"`UNIQ--references-0000000B-QINU`"' +
Being selective in using previously published work results in biased and/or incomplete analyses and conclusions. This endangers the integrity of claims, and harms society’s trust in research because it creates unfounded authority. '"`UNIQ--ref-00000005-QINU`"' Selective citations affect authors of previously published work, whether they are cited or not. It also affects readers of research hoping that it is accurate and unbiased. Other parties that might be impacted by selective citations are researchers conducting meta-analyses that synthesize a body of published work, decision making agencies that rely on accurate research results, as well as regulatory/oversight bodies of the research landscape.
'"`UNIQ--references-00000006-QINU`"' +
Spin can distort the production of knowledge and mislead readers and misguide decision and policy makers. Being aware of spin is tremendously important for readers of scientific papers, for researchers, for editors, for information specialists (synthesizing knowledge), and for users of scientific evidence, such as policy makers. '"`UNIQ--ref-0000000D-QINU`"''"`UNIQ--ref-0000000E-QINU`"' Distorted presentation and interpretation of results have been revealed in the cardiovascular literature.'"`UNIQ--ref-0000000F-QINU`"' While professional disagreement drives scientific progress, spin hampers it, as it frequently becomes static and entrenched. Attention and awareness can be ways to reduce the problem. It is an interesting issue whether spin borders to misconduct, as it can involve misleading and manipulation although spin is not directly related to money or profit.
'"`UNIQ--references-00000010-QINU`"' +
Gaining consent from all authors before submitting a manuscript demonstrates honesty and respect for colleagues. A number of cases have drawn attention to this questionable research practice. For example, a co-author found out a conference paper was re-published in another journal by the first author, without getting consent from all authors.'"`UNIQ--ref-00000008-QINU`"' Moreover, that same first author later translated the conference paper to their native language and published it in a journal written in that language, still listing all authors without consent. In this way, the co-author self-plagiarized without being aware of it. This case led to a request to the journal to retract the article.
<br />
'"`UNIQ--references-00000009-QINU`"' +
Research collaborations can occur between different partners, such as different departments within the same institution, different institutions, teams from academia and industry and teams from different countries. At the outset of the partnership, there may be varying expectations of how the responsibilities and benefits of the research will be distributed. This is especially relevant in the case of academic-industrial partnerships. <sup>2</sup> Multicentre research based on collaborations between the Global North and South may also pose special challenges, because of the differences between the infrastructure, resources and negotiating power between the partners. <sup>3</sup> Not addressing the concerns of each partner could lead to misunderstandings and disagreements during later stages of the research.
Based on the principles of fairness and distributive justice, all collaborators should ensure that benefits and burdens are distributed proportionately. The benefits of the research may take various forms such access to a deliverable product, career advancement of researchers, authorships and acknowledgements, local capacity building and others. These should be anticipated beforehand, and a consensus reached on how each partner can share in the outcomes. Similarly, the burdens and risks involved, such as personal risks to participants and field researchers, institutional investments in research and risks associated with providing and exporting data should be distributed as fairly as possible, in order to avoid the exploitation of one or more partners. Not doing so, or deviating from initial agreements on benefit sharing, constitutes a questionable research practice. +
Peer reviewing is a pillar of the scientific process, improving the quality of published research. Within the article submission process, many journals allow authors to suggest peer reviewers. By providing fake email addresses generated by the author themselves, authors can manipulate the peer-review process. Several forms of ‘fake reviewing’ can be distinguished:
#Authors can refer to existing scientists, but provide fake email addresses;#Authors can refer to ficticious reviewers, supplemented with an email address;#Third parties offering services can also provide fake reviews. '"`UNIQ--ref-00000013-QINU`"'
If Journal editors are unable to detect the fake reviewers, articles might be published without the quality control of sufficient genuine peer review. All major publishers, including Elsevier, Springer, Taylor & Francis, SAGE, Wiley, and Informa have retracted papers because of fake reviews. '"`UNIQ--ref-00000014-QINU`"''"`UNIQ--ref-00000015-QINU`"''"`UNIQ--ref-00000016-QINU`"' As of 2020 over 600 articles have been retracted because of fake reviewing according to the Retraction Watch Database.'"`UNIQ--ref-00000017-QINU`"' COPE estimates that 10% of retractions are due to fake reviewing. '"`UNIQ--ref-00000018-QINU`"'
Below a list of ‘red flags’ in peer review can be found, which should raise suspicions about authors aiming to fake the review process.
Red flags in peer-review'"`UNIQ--ref-00000019-QINU`"':
<br />
*“The author asks to exclude some reviewers, then provides a list of almost every scientist in the field.
*The author recommends reviewers who are strangely difficult to find online.
*The author provides Gmail, Yahoo or other free e-mail addresses to contact suggested reviewers, rather than e-mail addresses from an academic institution.
*Within hours of being requested, the reviews come back. They are glowing.
*Even reviewer number three likes the paper.”
<br />
'"`UNIQ--references-0000001A-QINU`"' +
Expectations, in relation to behavior, attitude and competencies, are often influenced by traditional gender roles and stereotypes. These expectations influence how people approach you. For example, the competence people expect you to have or not to have, influences whether you are encouraged to pursue a career in science or not, whether they invite you for a job interview or not, and whether they develop and promote your career – or not.
When selecting candidates for assistant professorships, committee members tend to look for someone “who has the potential to survive in the competitive academic world by being productive, confident, committed to the profession, and internationally mobile.”'"`UNIQ--ref-00000002-QINU`"' Men are more often perceived by others to possess these characteristics than women. Gender bias can manifest itself in a number of situations: in working conditions (such as job security), in recruitment and career advancement, and when awarding grants.
For example, researchers from the University of California studied video recordings of job interviews across five engineering departments of research universities.'"`UNIQ--ref-00000003-QINU`"' They found that female candidates received more questions and were interrupted more often. Women had less time to deliver their message than men. This study illustrates the phenomenon of “stricter standards” of competence that are demanded of women compared to men when applying for what is perceived to be a masculine-type job.
It seems that the people who reach top positions in the current system are mainly ‘alpha males’. This means that there is also a group of men that is underrepresented in these positions. Keep in mind: even if all people in top positions are men, not all men are in top positions. Moreover, we speak of ‘men’ and ‘women’ here. This excludes yet another group of people who do not identify as either one of them.
Gender bias goes against the principles of research integrity, and is detrimental to good science.
'"`UNIQ--references-00000004-QINU`"'
Authorship has important academic implications and authors are accountable for published research '"`UNIQ--ref-00000009-QINU`"'. Research showed that naming as authors those who have not contributed significantly to the study is considered one of the most prevalent types of authorship frauds '"`UNIQ--ref-0000000A-QINU`"'. Honorary authorship is often related to established or famous senior researchers who are named authors just because they hold senior positions and can help junior or less established researchers obtain funding or enhance chances for publications, awards, and recognition in the research community. It is considered that honorary authorship is often given with the recipient’s knowledge or even asked or demanded by recipients '"`UNIQ--ref-0000000B-QINU`"'. This is especially an issue for junior researchers who may feel pressured to assign authorship to senior researchers or feel they own authorship in return for their advice or help '"`UNIQ--ref-0000000C-QINU`"''"`UNIQ--ref-0000000D-QINU`"'. Similarly, the gift authorship usually includes researchers adding one another as co-authors regardless of contributions to the study to enhance their publication profile and spread collaborative networks. In many cases, this type of authorship fraud leads to a false representation of research skills and expertise, which gives these researchers an unfair advantage in competing for career opportunities and awards '"`UNIQ--ref-0000000E-QINU`"'. +
Peer review comments offer a space for growth, improvement of articles and increase the quality of reporting.'"`UNIQ--ref-00000005-QINU`"' Constrictive criticism is one of the most valuable qualities of a good review. However, when they are hostile they don’t provide adequate feedback. This way, a researcher misses out on the opportunity to improve work.'"`UNIQ--ref-00000006-QINU`"' Moreover, this type of review often has a negative impact on self-esteem, especially for young and inexperienced researchers.
'"`UNIQ--references-00000007-QINU`"' +
