Although considerable research has been devoted to avoiding unintentional bias in gathering data, much less has been written about avoiding intentional bias or fraud. This omission is curious given the serious effects that intentional bias can exert upon scientific research - and the difficulties of dealing with fraud, even where its existence is proven. Whilst these problems manifest themselves in almost infinite variety, fraud and intentional bias can be roughly subdivided into four classes.
- Data fabrication:
Data fabrication encompasses everything from reporting entire experiments (or surveys) that simply never happened, through to inventing observations (but not declaring the fact, or not allowing for it when analysing & interpreting them). Most of the famous (or rather infamous) cases of scientific fraud fall into this category. Making-up data is recognised by nearly all scientists as serious misconduct and, where proven, may result in dismissal. Unfortunately many surveys of fraud only cover data fabrication - and ignore the other categories listed below.
- Misrepresentation of results:
Misrepresentation of results covers what some regard as the 'acceptable' side of intentional bias and includes biased and misleading summaries, ignoring (redacting, abridging, or censoring) results, concealing crucial issues (or information) in favour of trivial (or irrelevant) issues, and using inappropriate statistics and analytical techniques (and disregarding their most crucial assumptions). This merges into data fabrication at one end of the spectrum and simple 'bad science' at the other end. Misrepresentation is widespread both in original scientific papers and (increasingly) in the subsequent interpretation of research by funding agencies (whether government or commercial) and the media.
- Plagiarism, multiple publication and unjustified authorship:
The first of these - plagiarism - is the most serious, and includes stealing data, or using someone else's ideas (without giving them due credit), or copyright infringement, or industrial espionage, or otherwise passing-off another's work as one's own. Multiple publication refers to publishing an identical (or near identical) paper in several journals, sometimes in different languages. Unjustified authorship is the practice of adding names of authors who have made no contribution to papers (sometimes without their consent), either to improve the chance of the paper being accepted, or to add to the publication list of senior researchers. This is widely practiced and even on occasions justified as being 'merely part of academia'.
- Paradigm dominance, intentional bias and suppression of dissent:
Paradigm dominance is a serious source of bias, although by its very nature it is seldom discussed. It produces a real-world versus official-world dichotomy and commonly emanates from the scientific establishment (including peer reviewers, research-grant committees and university authorities) who favour established paradigms and strongly discourage new ones. Some researchers exploit these attitudes by applying conventional methods to produce conventional conclusions simply because this is the easiest way to publish substandard work. This has a darker side when the intentional bias comes from political, commercial or religious bodies who propagate only their own views and (passively or actively) suppress dissenting views.
- Misconduct denial:
Misconduct denial is where researchers, administrators and journal editors ignore proven cases of fraud, refuse to investigate claims of fraud or intentional bias, or help conceal it. As you read about the various cases of scientific fraud, you will find misconduct denial is disturbingly common, and may explain why such fraud appears to be on the increase.
For obvious reasons, there is a strong tendency for the scientific establishment to deny or 'down-play' such problems. However, most studies on the matter, ranging from Savan (1988) to Fanelli, (2009) seem to indicate that scientific deception is a serious problem. It is very difficult to estimate the prevalence of fraud in science and, as might be expected in such a situation, estimates vary widely. Lock (1988) carried out one of the earliest (non-systematic) surveys by writing to professors of medicine and surgery in medical institutions in Britain, as well as other scientists and journal editors. He asked them if they could add any cases of fraud to those already documented in Britain. Results of the survey are shown below:
Overall the majority of people (58%) knew or one or more additional cases of misconduct. Lock considered that the problem was likely to be even larger than indicated by the survey, and had been informed by someone in the pharmaceutical industry that 5% of all drug trials were untrustworthy because of misconduct of some sort. He concluded that it was time that Britain established appropriate bodies to ensure scientific integrity. Fanelli (2009) found that about 2% of scientists admitted to have fabricated, falsified or modified data or results at least once and up to 33.7% admitted
other questionable research practices. When asked about the behaviour of colleagues, admission rates were 14%
for falsification, and up to 72% for other questionable research practices.
We believe it is essential that scientists be fully aware of the present situation, especially when assessing others' work. It is now clear that intentional bias in science is not uncommon, and that political and commercial interests have seriously distorted scientific opinion in a number of fields. In addition, Grayson (1995) suggests that the public is loosing confidence in scientific research.
There are many 'players' in science - individual scientists, the scientific establishment, commercial and political interests, and the media - and they all can be involved in scientific deception. We now consider the different types of deception carried out by the different players. Many of the examples we use are covered in more depth in the Examples section at the end of this More Information page.
Hargreaves et al. (2003)
It is at the individual scientist level that the some of the more infamous cases of fraud have occurred. Entire experiments have been invented complete with fictitious results and fabrication of artefacts. More commonly data are altered or inconvenient data are suppressed. Less obvious deception results from designing experiments so the results are inevitable and do not test any hypothesis, and analysing experimental results so they appear to point in a predetermined direction. Experimental results may be interpreted in a way that supports a particular theory, without exploring alternative interpretations.
Another form of scientific deception is plagiarism where other's data, analysis or ideas are presented as one's own, without crediting the original author. This practice is, of course, not new:
"In comparing various authors with one another, I have discovered that some of the gravest and latest writers have transcribed, word for word, from former works, without making acknowledgement."
Pliny the Elder A.D. 23-79
Natural History, bk. I, dedication, sec. 22
Publication of multiple versions of the same research paper is another form of deception. One form of this is publication of one's own (or worse, others') papers in a different language from the original. A related problem is unjustifiable 'guest' authorship, whereby individuals are named as authors when they have had little or nothing to do with the research. This can give undeserved respectability to a body of work. Multiple authorship can itself result in problems. As Horton (2002) points out, the views expressed in a research paper seldom reflect the plurality of the contributors views, but tend instead to only represent the views of the main author.
Most scientific journals use scientists who are not on the staff of the journal to assess whether manuscripts are suitable for publication. This peer review process is also open to considerable abuse by unscrupulous scientists. There may be deliberate bias in a review or delay in reviewing a research publication - and data or ideas may be stolen from research proposals, or papers under review.
One other form of deception has arisen as a result of the increasing incidence of fraud in science. Malicious allegations of misconduct may be made in order to sabotage the work of a competitor.
The scientific establishment
Here we include journal editors, the senior administration in universities and research institutions, and funding agencies. Journal editors might be expected to be in the front line of preventing publication of fraudulent results, yet they have traditionally shied away from accepting any responsibility. On occasions they have even declined to publish letters of retraction of fraudulent papers. On this topic the editor of the Journal of Clinical Investigations, Majerus (1982) commented that they were "the JCI, not the FBI".
As with any establishment, the scientific establishment seeks to maintain the status quo. This may take the form of suppression of unwelcome projects, hypotheses, or findings, and unjustified rejection of manuscripts or grant applications. There have been failures to enforce regulations, in particular for ethical clearance. A worrying form of deception that has apparently arisen recently is the debasing of terms, ideas, or procedures such that earlier work becomes meaningless. Williams et al. (1995) noted that some problems had even been perpetuated in order to safeguard funding to study them.
Perhaps the most insidious fraud in this category is the perpetuation of (often discredited) paradigms under the umbrella of 'conventional wisdom'. Once a theory or approach becomes accepted, there is enormous peer pressure to conform to it, even if an alternative theory or approach is equally plausible. In this 'official world' there is only one valid way of looking at things, whilst in the 'real world' things may be very different. We define this as fraud because it is intentional bias, albeit by the scientific establishment (often working with commercial or political interests) rather than by an individual. The inertia associated with such paradigms not only makes it difficult to publish research that appears to contradict the conventional ideas, but also tends to provoke extreme counter theories - which may be equally invalid.
"All truth goes through three stages. First it is ridiculed. Then it is violently opposed. Finally, it is accepted as self-evident."
Arthur Schopenhauer (attributed - possibly wrongly!)
Sometimes grossly over-simplified paradigms get absorbed into development policy, such as the 'tragedy of the commons' postulate that common land will not be managed as well as private land. This led to the blaming of groups such as pastoralists for overgrazing and subsequent environmental degradation, when in fact pastoralists strategies were generally far more sustainable than more intensive forms of agriculture. A similar thing happened when a paper was published in the journal Nature showing that non-timber forest products had potentially very high value. The development community converted this to a paradigm for sustainable forest conservation based on sale of such products by local communities.
|"An important scientific innovation rarely makes its way by gradually winning over and converting its opponents; it rarely happens that Saul becomes Paul. What does happen is that its opponents gradually die out and that the growing generation is familiarized with the idea from the beginning."|
The Philosophy of Physics
Just in case you think this problem is peculiar to the study of Physics, Huxley (1977) noted that the English Fabians, Beatrice and Sydney Webb made an historical study of the average time it took for an idea which at its first enunciation seemed revolutionary and revolting to be taken for granted and to be acted upon by the whole population. They concluded that the average time is twenty-eight years - roughly the length of a generation.
Commercial and political interests
In recent years this is the area where the most important and damaging fraud has occurred. Ong & Glantz (2000) alleged that at least one multinational tobacco company engaged in a misinformation campaign in order to undermine the conclusions of a report by the International Agency for Research on Cancer on passive smoking. In 1998 the Lancet medical journal was at the receiving end of this campaign, when it unknowingly published a pro-tobacco lobby letter about passive smoking.
Much has been written on the issue of commercial interests, primarily pharmaceutical companies, influencing the outcome of medical trials. Stelfox et al. (1998) reported that 96% of authors who supported the use of calcium-channel antagonists had financial relationships with manufacturers of calcium-channel antagonists, compared to 60% who were neutral on the matter and only 37% who were critical. Whilst Clifford (2002) did not find such a relationship, concern over the situation led the editors of 13 medical journals in 2001 to simultaneously publish an editorial highlighting the dangers. The former editor of the New England Journal of Medicine argued in a separate article that the editors did not go far enough:
"The entire system of clinical investigation is driven by profit. We are seeing the corruption of a system of research that used to have high ideals and be clearly in the public interest.
Arnold Relman (2001)
Trust me, I'm a scientist.
"Alternative" or "complementary" medicine is a rich source of fraudulent pseudoscience. There is of course some sound work being done in this area, but homeopathy stands out as beacon of quackery in an anyway murky area. Its practitioners constitute a multimillion dollar industry which regularly tries to undermine scientific analyses of its (lack of) efficacy. Weissman (2006) provides a recent review of the penetration of homeopathy into (amongst others) the British National Health Service and the US National Institute of Health.
Martin (1981) argues persuasively that in Australia commercial interests (mainly the forest industry and chemical companies) have joined forces with parts of the scientific establishment to suppress the views of dissenting environmental scientists. There is similar evidence of powerful commercial and political interests in the U.S.A. undermining scientific interpretation of data on global warming and rates of deforestation.
Governments are not averse to using 'spin' to promote or discredit results that support or conflict with their views. Berridge (2003) reported that the British Government was involved in a classic case of fraud and deception in the handling of the consequences of the great London smog of 1952.
If research results manage to make it intact through the interests so far mentioned, then they may have to run the gauntlet of media exposure. Some writers (especially those who have to deal with the media) have given the media a fairy clean bill of health when it comes to distorting scientific results. For example Krebs (2000) considered the attitude of the media only a minor factor leading to the change of public attitude towards genetically modified crops between 1996 and 1999. Others on the same topic have been less than complementary to the media:
"Since the media have as a priority a good story, and since scientific discovery is a gradual process of refinement of falsifiable hypotheses, it is almost inevitable that only the quirky and sensationalist science story gets aired...
The result is that the public are unduly panicked, insufficiently informed and view scientists as dysfunctional nerds or power-crazed robotic megalomaniacs with no normal emotions."
Professor Susan Greenfield (1999)
Science in the 21st Century (Millennium Lecture)
have provided an in-depth assessment of
of the media's role in the public (mis)understanding of science focusing on three contemporary issues: climate change, the MMR controversy and cloning research. Although written from the viewpoint of journalists, their report does provide some insight into how the public reaches its position on scientific issues of the day. Lowicki-Zucca et al. (2005)
focuses on the misleading and inaccurate HIV data and statements published by the media and humanitarian organisations on conflict-affected and displaced populations in Sudan, Uganda and Guinea.
Ethical principles for medical research