Dishonest Scientists Do Not Shy Away from ‘Faking’ Results

Approximately five hundred medical scientific articles are published each year reporting results from various studies; however, for instance, there is still no cure for cancer or other serious diseases. The problem is that scientists, after conducting an experiment, cannot obtain the same results upon its repetition. Therefore, the results of scientific research are unreliable in about 70 percent of cases. It has been observed multiple times that many research results in the fields of psychology, sociology, medicine, and economics are not sustainable when other researchers attempt to replicate them.

The replication crisis in science was first ‘diagnosed’ in psychology and then spread to almost all areas of science. John Ioannidis, a professor of medicine at Stanford University in the United States, first warned about it in 2005 in the article ‘Why Most Published Research Findings Are False?’. However, it cannot be said that this is a sudden phenomenon that has existed for just under two decades; it is merely a problem that has recently received more attention. Others, such as the magazine New Scientist, have also written about it.

– I must admit that some of the articles published in New Scientist, including those I wrote, are unreliable. We did not intentionally deceive; our reports were based on research by esteemed scientists from top universities, published in peer-reviewed journals. Despite meeting all standards of credibility, some results were false. Science is in a replication crisis; namely, once obtained results of scientific research are not the same in a second attempt – admitted Clare Wilson, a journalist and editor at New Scientist who covers medicine, health policy, neuroscience, and life sciences.

Systematic Deficiencies

Wilson is an award-winning journalist and editor of medical sections in New Scientist, and she is a biologist by profession. She also writes for other magazines, such as the health magazine Health Check for BBC World Service.

– Some areas, such as psychology, initially seemed more prone to creating such ‘fake news’ than others, and now almost every area of science is under suspicion. It has even been shown that the unreliable results of research span an entire area of genetics. The replication crisis arises from extensive deficiencies embedded in the systems and institutions that support scientific research, which not only allow poor scientific practices but also encourage them. This has worsened over the past few decades – claims Wilson.

Questionable Practice

The term replication crisis emerged just under two decades ago and indicates that discoveries in behavioral science often cannot be replicated because researchers do not obtain results comparable to the original, peer-reviewed study when they repeat that study using similar procedures. For this reason, many scientists question the accuracy of published findings and call for increased scrutiny of research practices in psychology. But what has caused the replication crisis in psychology? Some scientists have been warning for years that certain methods of data collection and analysis, as well as reporting them, which are often referred to as questionable research practices, increase the likelihood that results will appear statistically significant even when they are not.

One attempt in 2015 to reproduce one hundred psychological studies succeeded in replicating the same result in only thirty-nine of them. During 2018, in the replication of prominent scientific studies, it was found that only fourteen out of twenty-eight could be replicated, and an attempt to replicate studies whose results were published in top journals Nature and Science revealed that only thirteen out of twenty-one results could be reproduced.

Without Critical Experiments

However, in relation to psychology, much more detrimental false research results pertain to other areas, such as medicine and cancer drug research. Since many discoveries in cancer biology over the past two decades have failed to translate into clinically useful new therapies, replication has come to the forefront. For example, independent validation of the effectiveness of induced overexpression of genes through 242 experiments showed a success rate of only 39 percent. A total of 342 studies were processed, involving 242 different genes overexpressed in 184 cell lines. More than 60 percent of studies failed to provide reproducible overexpression.

Additionally, the cancer biology replication project was launched in 2013 as an ambitious attempt to closely examine key findings in fifty papers on cancer published in Nature, Science, Cell, and other influential scientific journals. Its goal was to determine what proportion of influential biological studies on cancer is likely correct. Researchers from the American biotechnology company Amgen reported that they were unable to replicate forty-seven out of fifty-three important cancer papers. This was widely reported, but Amgen did not disclose which studies were included in their research.

– Researchers usually take published results for granted and move on without reproducing critical experiments – stated Glenn Begley, author of Amgen’s report.

Deviations and Biases

Such deviations generally result in a loss of trust in research results and necessitate the need for independent validation studies. To achieve greater efficiency, a higher level of quality control of conducted research is required. For instance, studies containing genomic data, such as transcriptomic analyses using gene arrays, now require the publication of raw data.

What research practices have led to unreliable results? On one hand, scientific journals require exclusive content, leading to publication bias, and on the other hand, scientists are under constant pressure to conduct research as frequently as possible and achieve ‘revolutionary’ results to maintain their positions and advance in their careers. Additionally, they must be flexible in data analysis to achieve statistical significance.

Despite this, the true cause of the replication crisis in science should be sought in research methods, and flexibility in research methodology may help explain why researchers inadvertently, and sometimes intentionally, produce unreliable results. In conducting an experiment, the researcher starts from a predetermined hypothesis that is not always based on logic. It is claimed that the result of the research is statistically significant if it is assumed that the results of a study based on a certain sample of people will likely generalize to the broader population.

However, not all scientists are happy that the results of their research are being questioned, and they claim that the replication crisis in science is unrealistic and overstated. Nevertheless, many of them are now correcting their research methods. Psychologists, for example, have developed a range of strategies to ensure greater credibility of future discoveries. This includes conducting a greater number of replications of newly emerged findings, relying on larger sample sizes, and using thoroughly tested measures.

(The) Culture of the Academic Community

The immediate goals of reformers include greater transparency in study planning and data analysis, as well as more routine post-testing of results.

– We cannot just teach scientists how to write better papers. We also need to change the fact that better papers are not cited more often than poor papers, that poor papers are almost never retracted even when their errors are visible to lay readers, and that there are no consequences for poor research. In a way, the culture of the academic community actively selects poor research. The pressure to publish many papers favors those who can assemble them quickly – and one way to be quick is to be willing to take shortcuts. Over time, the most successful people will be those who can best exploit the system – said Paul Smaldino, a professor of cognitive science at the University of California Merced.

The journal Nature spoke with some researchers about the replication crisis in science, and while some praised the reproducibility project, others were concerned that it could unfairly discredit their work.

– Careers will be at stake if this turns out wrong – stated Atul Butte, a computational biologist at the University of California whose research yielded the same results in the repeated process.

Many scientists primarily conduct research just to secure funding for it, that is, to have work and publish scientific papers, which is one of the fundamental conditions for career advancement. Thus, the replication crisis in science, alongside all the aforementioned causes, is also a consequence of a system that encourages poor research.