Hacker News Reader

Comment by microcolonel | Hacker News Reader

The IRB should not be capable of learning about research conducted on them which amounts to a survey, before said research can commence. It produces an effective tyranny, where the only people who can validate the work of the affiliated institution are at the same time prohibited from doing so.

In my humble opinion, I think the IRB process is backwards: researchers should determine whether their methodology is ethical (or within whatever other institutional regulations), and be tried only after a report is received, or an investigation is started. Not approaching the IRB alone should not be an offense.

View on HN

> I think the IRB process is backwards: researchers should determine whether their methodology is ethical (or within whatever other institutional regulations), and be tried only after a report is received, or an investigation is started

Err, I don't think that would work as well as you think. Plenty of studies fall into gray areas, especially in behavioral research, and it's not necessarily easy to reverse the damage once it's done.

The "Little Albert" study could very easily be argued as "ethical" or at least not harmful (almost certainly the behavior became extinct over time) by the researcher, but nonetheless the potential for lifelong harm is there and it would likely never be approved today.

Sometimes once you've done the damage, it's done. Research institutions would rather be safe than sorry with studies involving people, both for basic ethical considerations and reputation interests. Especially after the spat of very questionable studies done in the latter half of the 20th century.

> The "Little Albert" study could very easily be argued as "ethical" or at least not harmful (almost certainly the behavior became extinct over time) by the researcher, but nonetheless the potential for lifelong harm is there and it would likely never be approved today.

Sure, but those lessons are already learned. It's not like research ethics was materialized out of thin air by institutional review boards: it was synthesized from experience, and then executed by the boards. The ethics are the same whether the assessment is before or after the fact, and nothing stops researchers from consulting the board or anyone else on the topic beforehand, if they are concerned they may run afoul.

I think it's silly to think that an IRB has a better grasp before the fact of how ethical or appropriate a methodology is than the researchers do during the research. In this case, the IRB is being used as a bludgeon to suppress low-risk research which is simply politically inconvenient to the institution.

  and be tried only after a report is received, or
  an investigation is started

At that point it's too late to save the victims. Obviously unacceptable, right?

(edit: got two posts backwards, sorry, fixed.)

Presumption of innocence is a foundation of law, at least in the United States.

So yes, waiting until there is a victim to start a trial is a reasonable approach to scientific ethics. It may not be the best one, but it’s surely justifiable.

The majority of industries which involve the potential to cause harm operate with comparable licensing and oversight systems. Food manufacturing, aerospace engineering, civil engineering, factories with dangerous equipment like smelters or refineries, all of these strictly require licensed professionals to do design work, that plans be approved by various oversight bodies before operation commences, and in some cases that those oversight bodies conduct regular inspections. We mandate this oversight because, without it, people do horrible things and get away with it. They hide their crimes too well, they have pockets deep enough to win any court case, or they foist the blame off on someone else. So we prevent harm instead of trusting people and corporations to be good when we know they won't be. https://www.niehs.nih.gov/research/resources/bioethics/timel...

> all of these strictly require licensed professionals to do design work

Sure, but it's not clear that the licensure itself reduces accidents or abuse. Plenty of licensed electricians (at least here in Ontario) are deadly incompetent. I've heard at least two accounts of electricians assuring builders or homeowners that a circuit is disconnected, and those people having an unwelcome surprise when they go to cut/remove said circuit.

All of these things come with a cost as well as a benefit.

It is possible to deal with crime through an 'oversight system', this is called a surveillance state.

As I indicated, I'm uncommitted on which way is best for the case under discussion. For criminal matters, and for hairdressing, I prefer presumption of innocence; for aerospace engineering, licensure and oversight.

I think it is a justifiable approach and might be a better one. We chose IRBs after appalling cases like the Tuskegee Airmen experiment; perhaps we should have hanged the perpetrators instead.

Perhaps not. Having never been an academic, I am aware this subject invokes strong feelings, while having none myself.

The Tuskgee syphilis experiment resulted also in informed consent laws in clinical trials; which I would argue would have a better effect than trying to get a board within the institution which carried out this experiment to apply an ethical standard to the study.

Only criminal law and there's no reason to think that standard should be applied in all situations beyond criminal law. The criminal standard is so high because the consequence of guilt is so high (including incarceration or even death) and because it's applied by the State.

Should the consequences of harming test subjects be any lower?

I can make a solid case for a less oppressive IRB regime and criminal sanctions for those who harm their test subjects.

Prior restraint in the form of requiring approval by an IRB is better than harsher punishment. The point is to prevent any research from getting to the point of criminal charges.

A researcher operating under "better to ask forgiveness than permission" is in serious mad scientist, Frankenstein territory.

Reply on news.ycombinator.com