Data Ethics Ethical issues can arise when we collect data from people Difficulties can be more severe for experiments – Impose treatments – Can possibly harm as well as help their subjects Sample surveys just gather data – However, sensitive data/topics can still cause ethical issues

Basic Data Ethics Review by an institutional review board – Review planned studies in advance – Aim to protect the rights and welfare of human subjects – May require changes of plan – Monitor studies while ongoing

Basic Data Ethics Subjects in a study must give informed consent before data are collected – Must be told about the nature of the study and any risk of harm it will bring Physical harm or possible mental harm – For sample surveys, subjects should be told about the nature of the questions and about how much of their time Subjects must consent (usually in writing)

Basic Data Ethics There are many open questions about who can give consent – Children (usually have to obtain consent from parents) – Prisoners – Mentally incapacitated people – Severely injured people in emergency situations

Basic Data Ethics Confidentiality – All data about the individuals must be kept confidential – Only summary statistics are allowed, not individual responses – Different from anonymity—where subjects are anonymous Very rare in statistical studies Does not allow for follow up

Basic Data Ethics Advances in technology are complicating confidentiality protocols – Interconnected data bases – Hacking – Social media

Clinical Trials Clinical trials are experiments that study the effectiveness of medical treatments on actual patients Treatments can possibly harm as well as heal Trials can produce great benefits (cures, etc), but also pose risks – Most benefits go to future patients – Most risks are borne by subjects in the trial

Clinical Trials The challenge is to balance the risks and benefits Ethics and human rights standards say “the interest of the subject must always prevail over the interests of science and society” Medical treatments can only be tested when there is reason to hope the subjects in the trial can be helped. – Future benefits are not enough reason to justify a trial

Clinical trials If there is already significant evidence that a treatment works, then it is unethical to not give the treatment It’s a balancing act as to when to conduct a randomized trial – Must be sufficient belief that it will work – Must be sufficient doubt that it will work to justify withholding it from the other half of the trial subjects

Clinical Trials Question for Class Discussion: Why is it ethical to give a control group of patients a placebo?

Behavioral and social science experiments Physical risk to experimental subjects is less than in clinical trials However, possible benefits to subjects are also less Risks to consider: emotional harm, embarrassment, invasion of privacy, etc Informed consent is usually more vague than in medical applications – Subjects might change behavior if they knew exact purpose Many famous earlier studies would no longer be considered ethical by today’s standards