Emerging Ethical Breaches in Science
Emerging Ethical Breaches in Science
  • Reporter Kim Jin-Seong
  • 승인 2023.11.07 20:17
  • 댓글 0
이 기사를 공유합니다

▲Cherry picking in statistical analysis / GE HealthCare
▲Cherry picking in statistical analysis / GE HealthCare

  Can quantum computers shut down modern cryptogram systems? Can subsequent researchers reproduce the same results the pioneer researchers achieved? As both scientific research and industrial fields are discovering new things as well as achieving richer experiment results more rapidly, the competition is intense. In these circumstances where some researchers work hard under high pressure, there are serious issues emerging where researchers violate fundamental principles of scientific research including reproducibility, and exaggeration. 
  Reproducibility can be defined as “obtaining consistent results using the same input variables, methodological and computational steps, and analysis conditions” in the biological sciences. Glenn Begley contributed an article to Nature article, raising standards for preclinical cancer research. He proposed how methods, publications, and incentives must change. The article revealed that among 53 leading research papers about cancer, only six of them were capable of reproduction and the rest were impossible to reproduce. This article brought the reproducibility problem to the surface.
  On paper, experiments should obtain consistent results under the same conditions. However, research that lacks reproducibility frequently appears in the real world, and issues regarding reproducibility come and go. Some examples include Mendel’s experiment, the amyloid beta hypothesis, and the story of Dr. Hwang Woo-suk. 
  Mendel’s experiment and amyloid beta hypothesis are examples of research that are impossible to reproduce but still indicate a meaningful theory. Although they lack reproducibility, they are meaningful in terms of research, but just not as strictly tied together in practice. It is widely known that the result of the pea experiment by Mendel is exaggerated but Mendelian inheritance is still valid. The problem here is that Mendelian inheritance is not the only factor to consider regarding inheritance and there are other factors in the real world. Similarly, amyloid beta aggregates have a causality with neuron damage but not as much as suggested in the paper.
  Problems like reproducibility are critical since the impact of fake research is huge. Subsequent research that is based on the original research’s theory will also be meaningless. In this case, time and money are wasted. 
  It is fortunate that Mendelian inheritance and amyloid beta hypothesis still indicate valid findings. However, studies might not be reproducible when researchers write papers about the best outcome among hundreds of experiments, or when they change the data at their own will. 
  In various fields, cherry-picking is becoming more and more prevalent. Cherry-picking is the most common statistical fallacy that occurs in various fields. It is a phenomenon where pickers, in this case, researchers, make a biased selection of indicator data to support preconceived ideas. Just like where pickers are expected to select only the ripest and healthiest cherry on a farm. When cherry-picked, a significant amount of data that contradicts preconceived ideas might be ignored and cause significant changes in the performance of models. Like the case of reproducibility in bioscience, 
cherry-picked results are not necessarily incorrect, but the complete result is not suggested since data to support the other side of the complete picture is not factored in. This cherry-picking may occur when there is a large pressure to achieve rich experiment results, especially when the experiment setting is expensive, like in large-scale models.
  The spread of pseudo-science facts such as exaggerating the objective of the research is also a problem. Most physicists forecast that long term research is needed to develop a effective quantum computer for more than dozens of years. Some random people might come out and say some random pseudo-science facts, in a fun way, to the public, for example, “Quantum computers will be familiar to the public at least in ten years” or “Quantum computers will destroy the cryptogram system” although modern cryptogram systems will not collapse even if quantum computers are utilized. It is difficult to correct widespread spilled facts. Therefore, well-educated scientists should write books or be visible to the public, delivering exact knowledge instead of random titbits.
  Ethical breaches come and go, re-entering the public eye from time to time. These breaches occur under high pressure and competition. Researchers should do precise and delicate work while others should give them appropriate respect and support.