This is Jessica. My collaborators and I have been exploring decision-theoretic approaches to fine-tuning language model agents to support expert decision-making, and we’re now seeking clinicians or other medical professionals for a brief (~30 min) session to get feedback. We are particularly interested in talking to clinicians with some experience in diagnosing cardiac dysfunction. We’ll share a small number of cases and preview what our method does to get your thoughts on how it aligns (or doesn’t align) with your domain knowledge and how you think it would affect the diagnostic process.
If you are available or would like more information, please contact me or Ph.D. student Ziyang Guo, who is leading this work ([email protected]). We’ll provide gift cards in appreciation of your time.
I can’t pay my rent with a gift card. Imagine the bias in the sample we’re getting by offering gift cards to MDs who earn 2,3,400k/year. I’ll leave it as a thought exercise to think about what kind of docs would do this.
I agree, you probably won’t have much success paying your rent looking for gift card on stats blogs. Maybe if you comment on enough of them, though? Worth a try!
The framing might be different though. One is a thank you in exchange for a good deed (possibly accompanied by a warm glow effect?); the other is payment for services. Not always, but sometimes these things do matter!
I was told that people often prefer lotteries (one big prize with a chance of winning), but I think the most reasonable thing is to consider the expectations that people have. If the social norm is to give a gift card, then giving a gift card makes the most sense.
Any time you go out and recruit people to talk to, it’s important to consider who you might be leaving out and what effect that might have on what you end up hearing. What you offer as an incentive to participate is only one potential factor among many.
The kind of docs who take part in studies like this are generally those with an interest in the subject. Some are motivated by the chance to speak as the expert in the room; others by what they see as an opportunity to contribute to their field, or to take a peek at the research that others are doing that may one day help their patients. I’ll take people with these motivations over people who are showing up for the cash any day of the week.
Few MDs are motivated primarily by the honorarium, whether it’s cash or a gift card, especially since the dollar amount for a study like this one won’t be large enough to register as a major windfall for someone making 2, 3, 400k/year. (But also, you *can* buy food with a gift card, so if there were a month where cash was tight because the payments on both the yacht and the private jet came due the same week, there’d be that.)
I agree that the gift card is a way of showing the appreciation, not a wage, and I think that a larger payment would introduce its own issues.
Here is a suggestion that builds on this: what if you offered volunteers a choice? For example, they could select either the gift card or have a donation made in their name. The real value here is not just the choice itself, but the data it provides. You could then analyze if the feedback from the “gift card” cohort differs from the “donation” cohort, giving you an empirical way to explore the very bias being discussed.
Yes, we do it as a token of appreciation. The choice of gift card or donation is a nice idea.
I’ve offered the donation option to doctors who were reluctant to accept an honorarium, whether because of the Sunshine Act (I often worked with med device companies) or just a feeling that it would be improper. It’s a good practice. However, while I don’t know anything about Jessica’s study beyond what’s written above, it sounds like it’s aiming at getting qualitative feedback from a relatively small number of people, which would make that analysis hard to pull off.
There is some research from the literature on survey incentives suggesting that people prefer cash to equivalent gifts. But that’s gifts, not gift cards. Gift cards are functionally almost equivalent to cash, so I doubt there’d be much of a difference. That is, I don’t see payment using a gift card as feeling more like an “appreciation” than payment using cash; it just might a bit more convenient, in terms of paperwork, for an organization to send out gift cards rather than cash.
> it just might a bit more convenient, in terms of paperwork, for an organization to send out gift cards rather than cash.
Yep, cash payments at Northwestern require significantly more paperwork, so they encourage VISA giftcards instead.
Visa Gift Cards are a nice solution if that’s a concern
In seriousness, compensation for research participation has important ethical dimensions; for a good overview, check out:
https://research.uoregon.edu/manage/integrity-compliance/human-subjects-research/guidance-library/compensation-participation-research
One issue is that compensation cannot be such that it acts as a form of coercion or bribery. At that point, potential participants may feel pressured to take on greater risks than they otherwise would in order to receive that compensation. In other words, they cease being “volunteers” and become more like servants, with researchers using compensation to exert control over participants without necessarily accepting any responsibility for participants’ well-being.
That said, the level/form of compensation that would reach the level of “coercion” will vary between people, between geographical regions, and between different types of research (e.g., it would be reasonable to expect greater compensation for a 6 week study than a 30 minute survey). Thus, striking the right balance between fairness and coerciveness is tricky and is one reason why ethical review boards are valuable institutions to keep around!
Gec:
I don’t buy that “coercion” argument; indeed, I think it can be unethical to not pay participants in a study; see here and here.
I agree that the coercion argument can be taken too far and should not be the only factor considered. As you point out in the linked posts, it is also unethical to expect free labor from participants, thus the need to strike the balance I mentioned in my earlier comment.
But speaking of going too far, I’ve also had the experience of IRB admonishing me for having a jar of candy in my lab available for participants.
Maybe that was a mistake if you were doing a study of diabetics.
We’ll turn this into a data lesson. I go to a dealership, but the allante is rare, so the only
thing the employee sees are allantes that are being sold by poor people for parts for $1000. Mine was pretty good condition (‘93), and I had some work done, which was expected since it was old and sitting in a driveway, and haven’t had to take it back since.
But the employee, not being data savvy, offered me $1,000. I even brought a list of repairs to show the dealer it at had been overhauled.
The only reason I wanted to sell it was for better gas milage (oldy that takes premium). Not a long range car.
I digress, problems interpreting data with small
sample size, a non-representative sample, latency, and over extrapolating. *sigh* given that he works at a dealership I would put $50 on the fact he’s never come near beginning the calculus sequence
Please keep future comments on topic. Thank you.