Paper presentation

Team 5

Manivannan (a1812066)
Rifat (a1836604)
Lim (a1757976)
Yogeshvar (a1817369)

Ethical Issues in Empirical Studies of Software Engineering

Abstract

Emphasizes the significance of ethics in empirical software engineering studies.

Informed Consent

  • Challenges in consent due to power dynamics in classroom research.
  • Importance of obtaining valid consent in all research involving human subjects.
  • Dr. Gauthier experiment on Students violates primary ethical principle in human subject research
  • Disclosure, Comprehension, Competence, Voluntariness, the actual consent or decision, and the right to withdraw from the experiment.

Scientific Value

  • Amaro's research was not done properly and produced unreliable results (Selection Bias).
  • The study lacked value because it did not follow the right methods and consider potential risks and benefits.
  • Scientific value: Importance of research & Validity of experiment results.
  • “common sense approaches” were shown to provide unreliable and invalid data.

Beneficence - Human

  • The risk/benefit ratio in research must prioritize maximizing the benefits for society and subjects while minimizing potential harm.
  • In the case of reengineered engineers, no consent was obtained from software engineers as upper management had already made the decision to translate the code.
  • Engineers face a significant risk of unemployment as they must adapt to changes over which they have no control.
  • The organization is favored with benefits, but at the expense of degrading the quality of engineers' working lives.

Anonymity

  • Measures are taken to ensure the anonymity of participants in an experiment, preventing anyone from being identified.
  • Personal information or data that could potentially identify the subjects (such as names) is not collected.
  • The data set pertaining to the subjects is designed in a way that prevents reverse tracing back to individual participants. To minimize the risk of exposing the subjects, aggregated data is reported instead of raw data.
  • Subject numbers or aliases can be used to identify individual data sets.

Confidentiality

  • Maintaining complete anonymity can be challenging due to possible interactions between researchers and subjects being witnessed by coworkers.
  • Raw data should be securely stored, possibly under lock and key.
  • Inform subjects about confidentiality limits and implications during informed consent, especially in small-scale workplace studies.
  • Researchers must minimize harm to organizations by considering the potential consequences of their findings and publications.
  • Ethical considerations need to be maintained to minimize harm at the organizational level, ensuring that disclosing information that may put companies at risk, such as flaws in their processes.

Beneficence—Organizational

  • Researchers must minimize harm to organizations by considering the potential consequences of their findings and publications.
  • Ethical considerations need to be maintained to minimize harm at the organizational level, ensuring that disclosing information that may put companies at risk, such as flaws in their processes.

  • Ethical responsibility lies in offsetting these harms to the greatest extent possible.

Critical Review

  • Inadequate information regarding the procedure of each experiment. Overview of Empirical methods and its strength and weakness are missing
  • Ethical misuse of consent is observed. However, the proper procedure to deal with this issue is not mentioned.
  • Regarding Re-Engineering a program, upper management can always make decisions rather than the engineers who worked on the code. This may reduce the effect on the work environment, but it might be harmful to an individual.
  • Programmers do not want to go through metric research, however, the procedure for not putting them at risk by reviewing their code is not described.

Q & A

Thank You

RSMC

By Maggie

RSMC

  • 178