Books

HOME > Publications > Books

Manual for Research and Publication Ethics in Science and Engineering

Cho Eun Hee, Kim Young-Mog, Park Kibeom, Son Wha-Chul, Yoon Tae-Woong, Lim Jeong Mook, Hwang Eun Seong
ISBN-13: 978-89-5938-345-0-93190
Korean Federation of Science and Technology Societies
This is an open-access publication distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Appropriate Research Practices for Preventing Research Misconduct

Appropriate Research Practices for Preventing Research Misconduct

I. Scientific Thinking and Research Originality

Scientific thinking is based on a rationale and originality.

  • Rationale: Research must be conducted only when its scientific necessity is acknowledged. In modern society, which has advanced to a science- and knowledge-based community, scientific technique may become a method of earning profit and a great influence upon social development and the maintenance of stability. In other words, even if research simply originates from personal curiosity based on one's individual freedom of thought, the research may still impact the development of modern society; therefore, a meticulous deliberation upon the necessity of the research may help in eliminating the economic and social risks arising from scientific misconduct. A scientific rationale is the most essential factor for determining the range of research subjects and the kind of samples, thereby greatly impacting research on human or animal subjects to meet the requirements of bioethics.

  • Originality: Research originality may be achieved by proposing (1) a verification of fact that has not yet been proven or a process being tested for the first time or (2) a technique or method that overcomes current limitations. Research that has developed improved techniques instead of simply reproducing the results of past research is also acknowledged for its originality.

Characteristics of scientific research are (1) ability to self-correct, (2) objectivity, and (3) reproducibility.

  • Ability to self-correct: The ability to self-correct involves acknowledging that the method proposed by the researcher is not the absolute truth and may be modified in future research.

  • Objectivity: Objectivity is conducting research and analyzing its results in a rational, impartial manner without imposing one's personal beliefs, perception, values, or empathy.

  • Reproducibility: Reproducibility represents the validity of the results; if the research is re-conducted in the same conditions by another researcher, identical results must be achieved.

Research activity should be original and based on a rationale, and it must ensure self-correction, objectivity, and reproducibility for proper hypothesis verification.

Research topics and hypotheses are established based on rationale and originality; the relevant scientific activities must be limited to proving or rejecting the proposed research hypothesis. A research hypothesis is a sentence expressing facts or phenomena to be explored by the researcher. Research activity that determines the authenticity of the hypothesis is referred to as the process of ‘hypothesis verification.’

When the hypothesis is rejected, additional research can be conducted to determine the reason for hypothesis rejection, and through this process, the researcher discovers what he or she has not recognized previously. Therefore, rejection of the hypothesis is not a ‘research failure,’ and neither rejection nor acceptance of the hypothesis functions as an obstacle in research practice; in fact, this is an important process in discovering new facts or phenomena in nature.

II. Research Design

Research design must be based on the standard operating protocol (SOP) and bioethics.

1. Standard operating protocol

The SOP is a document detailing the procedure and methods of a certain task in order to achieve consistency by standardizing all lab techniques conducted in a laboratory.

Not only researchers affiliated with their own research lab, but also external researchers working in the corresponding lab must follow the protocol.

In order to test a hypothesis, a researcher conducts an experiment with two different groups: a test group to which some special treatment is applied, and a control group to which nothing is applied. All procedures of the experiment being conducted in a laboratory must be conducted and standardized according to the SOP of that laboratory.

The SOP must be updated continuously and should be disclosed to all researchers conducting their experiments in the laboratory.

An SOP is available only for the laboratory establishing the SOP, and may not be suitable for other laboratories exposed to different environments.

2. Compliance with bioethics (refer to Chapter 6)

In the case of research involving human participants, the core basis of bioethics is protecting the subjects; in the case of research involving animals it is acknowledging the value of their lives.

An institutional review board (IRB) and institutional animal care and use committee (IACUC) are both established and managed in each university and institute, and they review various ethical issues related to studies using human subjects and animals.

Academic journals require the researchers to observe the standards of bioethics in all research procedures and also require that approval of the IRB or the IACUC be stated in the text of the final manuscript.

3. Methods of prospective study and retrospective study

Prospective study | Prospective study is a method of planning and conducting research for hypothesis verification (observational study). The researcher can intentionally design control and treatment groups and set up the experiment environment according to the research purpose and hypothesis and all experimentation should be conducted observing the SOP. Compared to the retrospective study which requires differentiation and collection of data on many different variables, a prospective study allows precise analysis of the effect of the experimental treatment. Laboratory research and typical study designs are generally prospective studies.

Retrospective study | Retrospective study is the interpretation of previously collected data (a review of the medical records) on subjects of investigation from a period prior to when the study is carried out. This method requires examination and collection of data from the observation group with varying parameters, and therefore the various factors in the control group should be standardized in order to evaluate the hypothesis.

4. Principles of randomization, even distribution, and replication of experiments

Researchers must have a good understanding of the following principles for their research design.

1. Randomization

  • Randomization refers to a method of randomly selecting subjects in order to reduce bias stemming from experimental conditions and the environment. When selecting the control group and experimental group for hypothesis verification, the randomization process implies assigning humans, animals, or objects to each group on a random basis so that each group has identical characteristics and traits.

  • Due to the diversity of biological experiments, even if each group is selected on the basis of strict criteria, it is almost impossible to assign the experimental group and the control group to have identical characteristics. Consequently, subjects are randomly selected to prevent unintentional bias resulting from biological diversity.

2. Replication

  • Replication is repeating an identical treatment regularly in a single set of experiments for at least a certain number of times (usually 4 times).

  • To minimize any bias that may naturally occur in spite of the randomized selection of the control group, the experiment is repeated in an identical fashion and the data obtained from each experiment should be analyzed using the appropriate statistical methods.

  • This is essential for not only confirmation of data reproducibility but also proper statistical analysis.

☞ Confusion between replication and duplication of experiments

Researchers often confuse replication of experiments and duplication of measurements or observations. Replication of an experiment is repeating the whole experiment using the same experimental procedures but independently with different samples; duplication of measurement, however, is repeating the readings of experimental results from a single sample in order to reduce data fluctuation. Each replicated experiment is considered to be an independent experiment, while duplicate measurements, as merely repeat observations, should be considered statistically as one experiment. When an experiment is replicated and observed a couple of times to be able to obtain the necessary amount of samples, it should still be considered as a single replicate experiment. Likewise, when the obtained samples in an experiment are divided and measured a couple of times, it should still be considered a single replicate experiment. Duplication of measurement must be distinguished from replication of experiment, and when a disparity arises, it is considered to be an error in the data.

3. Even distribution

  • Unlike randomly selecting each experimental and control group to be identical in both quantity and quality, even distribution is selecting the experimental and the control group to be evenly constituted by establishing a special order or rule.

  • Randomization and even distribution are rules used complementarily to form a valid experimental group and control group (only one difference exists and the rest of the characteristics are the same).

  • An example of combining even distribution and randomization: In an experiment using newborn offspring (such as of mice), it is impossible to predict the number of newborns. When three separate experimental groups are designed, the mother must bear her offspring in multiples of three to achieve even distribution; however, it is impossible to artificially manipulate the subjects accordingly. Therefore, when the number of offspring is not in multiples of three, newborn offspring born on the same day must be randomly selected in multiples of 3 to achieve even distribution, and the remaining offspring must be excluded from the experimental treatment or be used in another experiment. When this type of distribution is ignored and if the offspring born from all the replicate experiments are randomly allotted, ignoring the differences in numbers, to each experimental group, the offspring cannot be considered evenly distributed.

III. Research Practice and Data Management

1. Laboratory notebook

Researchers must keep records of their research data in their lab notebooks in order to collect and assess the data of their results in an objective, unbiased manner.

Researchers must not only record their data directly in the notebook, but also enter all procedures performed during the experiments along with any environmental factors. The lab notebook is used effectively for research data quality management (to ensure originality and reproducibility) as well as to eliminate risk factors (to resolve problems).

In the case that the results are unexpected or difficult to understand, by closely scrutinizing the lab notebooks, the researchers can find the causes or discover new facts or phenomena that have not previously been discovered. In other words, by utilizing the information recorded in the lab notebook, researchers can design new experiments and troubleshoot problems from difficult experiments.

Writing the lab notebook: Choosing a notebook

  • A hard copy notebook: For most purposes, a bound notebook rather than a spiral bound notebook is used with no set format other than the ruled lines.

  • E-notebook: An E-notebook uses the web and software as well as a database. Although it has the advantage of being able to set the format suitable for an SOP, unauthorized modifications must be prevented and risk factors such as data loss caused by technical problems must be eliminated. Researchers must manage their data by using a cloud-based system or an archive service, which requires establishing the required infrastructure. Although an E-notebook allows for the use of various devices such as touch pads to compensate for the lack of mobility compared to a paper lab notebook, the relevant software must also be developed at the same time.

Management of a lab notebook

  • Researcher must use a non-modifiable medium such as a PDF file or ink to prevent arbitrary modifications, deletion, or unexpected loss of data. The lab notebook must be kept in a safe, permanent location and should be retained securely during the required data retention period and as long as the laboratory is in operation.

  • In principle, a lab notebook may not be taken out of the laboratory and should be retained inside the laboratory permanently. When a researcher is going to transfer to another job, he or she may keep a copy of the lab notebook by obtaining an approval from the principal investigator or the head of the laboratory.

2. Statistical analysis of research data

The first thing to check in order to achieve an accurate statistical analysis of experimental data is the number of samples assigned to the groups of the experiment (control and treatment groups).

Different methods of analysis must be used depending on whether the number of samples in each experimental group is identical or different. Generally when the number of samples is identical in each group, a t-test can be used, but when the number of samples differs among groups, methods such as linear model analysis must be used.

When comparing the experimental effect among more than two groups, the model effect (used to evaluate the differences in normal distribution among all of the groups) must be first tested prior to conducting the comparison among each of the individual groups (paired test: comparing the three experimental groups—A, B, C— testing for a significant difference between A-B, B-C, and C-A). When analyzing the model effect, analysis of variance (ANOVA) is used, and if the model effect is not significant, there is no need for a paired test.

3. GLP, GMP, GCP

Standardization of both laboratory management and research techniques are criteria for achieving reliable results.

Applying the criteria of good laboratory practice (GLP), good manufacturing practice (GMP), and good clinical practice (GCP) are global trends in establishing standardization, and they are used as guidelines to ensure credibility of experimental results and industrial products. In Korea, the relevant regulations are established and applied in each field through agencies such as the Ministry of Food and Drug Safety, Ministry of Environment, and Ministry of Agriculture and Forestry.

1. Good laboratory practice

  • GLP provides guidelines for good laboratory management.

  • GLP is the regulation applied to non-clinical laboratory studies conducted for the assessment of the safety and efficacy of medicine, pesticide, chemicals, and commodities.

  • GLP is not simply limited to the standardization of experimental methods or facilities, but embodies a set of comprehensive principles that regulates the laboratory management system and the entire procedures and conditions of the research related to study design, experimental execution, data monitoring, recording, and reporting; for example, it includes the issues of securing the appropriate number of subjects, as well as education.

2. Good manufacturing practice

  • Expanding upon GLP, GMP provides the minimum requirement guidelines for facilities in manufacturing products with high quality assurance.

  • The facility standards of GMP may vary by the manufacturing facility but like GLP, GMP provides guidelines for the overall manufacturing process.

3. Good clinical practice

  • GCP provides management guidelines for clinical trials involving pharmaceutical products.

  • GCP describes the guidelines that must be observed in conducting clinical trials.

  • GCP is the ethical guideline that should be followed in designing, conducting, recording, and reporting clinical trials involving human subjects, and is the minimum detailed set of standards for scientific practice.

IV. Writing Reports and Research Papers

When writing a research paper or a report, the researcher must report the results of the study following the author guidelines. The general guidelines for writing a scientific research paper are as follows:

  • Write as concisely as possible.

  • Avoid unnecessary words and expressions.

  • Write appropriately for each section.

  • Write only about the relevant research, and avoid speculative expressions and interpretations.

  • Always cite the source of other researchers' work, and do not include their data in your own data.

  • Use a consistent scientific style and format throughout the paper.

  • Try to achieve unity and coherence and maintain continuity in your writing.

  • Follow the standard organizational style for scientific publications: title, abstract, introduction, materials & methods, results, discussion, references, and acknowledgements.

The following are the detailed guidelines for writing each section.

1. Title

The title must specify the basis or the hypothesis of the study in a sentence or a phrase.

2. Abstract

The abstract is a summary of the key contents of the study, and each academic journal has its own word limit (around 200-250 words, in general, in English). The abstract generally describes the study in one sentence in the order of research purpose and hypothesis, research methods and design, and the samples or the animal subjects used in the study, followed by a description of each experiment method and its results. The last sentence in the abstract is the overall conclusion of the study, presenting the results of the hypothesis verification.

3. Introduction

The introduction should clearly state the rationale of the research.

  • The introduction is generally presented in the order of the present conditions of research development, current issues or the necessity of the research, and the hypothesis and design.

  • The present conditions of research development are succinctly described in the discussion so the readers can have a clear understanding of the research background.

4. Materials and Methods

Research ethics as well as the human and animal subjects of study are presented first, followed by basic experimental methods and a brief description of the specific experimental methods pertaining to the research process. Experimental design (experimental group, independent variables, parameters) and methods of data analysis are also presented; however, experimental design may be required to be presented at the beginning of the materials and methods or at the beginning of the statistical analysis methods depending on the academic journal.

5. Results

The results are a succinct description of experimental observations for hypothesis verification based on statistical analyses. Only the numerical data and the observed/measured results essential to the main points are presented, and their interpretation is presented in the discussion.

6. Discussion

The description of the discussion section should be restricted to the interpretation of the hypothesis and the experimental results of the study. The discussion should focus on the most important interpretation of the results, and should avoid lengthy or speculative explanations of other researchers' experimental results. Even when describing future study plans, it is best to compare only the experimental results directly associated with the experiments being reported.

A research proposal, unlike a research paper, must present a clear research design and must describe the preliminary experimental results with a relevant working hypothesis. Because the research plan and design conducted according to the research proposal may change after the start of the experiment, it is most efficient for the research proposal to focus only on the first experiment.

V. Lab Management and the Lab Environment

1. Communication

The most important virtue in the laboratory is communication.

Having effective communication among laboratory members is essential because the members, by being affiliated with the same laboratory not only use the same experimental devices and share the general SOP, but also share the same data and materials.

  • Communication among the lab members: Since a lab member's research directly affects the research results of his or her colleagues, communication among the researchers is essential for the quality management of the laboratory and the research results.

  • Communication between the principal investigator and the research staff: Research staff members who are responsible for conducting the experiment itself and the principal investigator responsible for supervising the overall experiment should complement each other, and failure to do so may easily cause errors in interpretation of the research results as well as manipulation of the research process. The principal investigator must refrain from giving authoritarian-style research guidance, and lab members must avoid omitting or distorting reports of research results.

Regular and irregular lab meetings

  • In addition to the regular laboratory meetings, lab members and the principal investigator should actively engage in informal meetings.

  • A forceful atmosphere in the management of the laboratory as well as a command-submission style of meetings should be avoided.

  • Laboratory meetings often take the ‘business meeting’ format, and therefore detailed reports of the research as well as members' review and discussion on the direction of research are critical. Lab members tend to place high importance only on presenting the data results and exaggeration in order to receive good evaluation by the principal investigator, which should be discouraged; in fact, the discussion of the assessment of the experimental results and the establishment of proper research strategy should be emphasized more.

Any topics including changes in the standard research technique, writing the lab notebook, and resolution of problems may be discussed. The minutes of all lab meetings must be recorded and be copied to all lab members who have attended the meetings.

2. Research practice

Research must be conducted providing equal status to the principal investigator and the other research staff members; agreement on the direction of research and consistent implementation of the experiments are essential for good research practice.

Research conducted based on effective communication and agreement among the principal investigator and the other research staff members is directly connected to the planning stage of the experimental design and greatly affects legal issues such as copyright issues.