Sahana: Hi everyone! Welcome to In Limbo Conversations. Today we have with us Nick Byrd. He is an Assistant Professor at the Stevens Institute of Technology and also is the co-editor of the Brains Blog. His central areas of interest are - Reasoning, Well-Being and Technology. For more details about his research, please click here.
Sahana: Thank you for joining us Nick!
Nick: Thanks for having me.
Sahana: So I thought I'll start with some questions specific to Nick's research and then I'll move on to some public philosophy ones. So coming to the research ones, Nick, I recently came across some briefs about chapters in your dissertation and you had mentioned some points, so I thought we could just talk about some of them in relation to the pandemic situation.
So, the first question is about how you understand reflection. Could you give us an instance of a situation and what would constitute a reflective and unreflective decision making in it?
Nick: Yeah, so reflective reasoning is something we're probably all somewhat familiar with, intuitively, right? So some people describe it as just stepping back from some initial impulse that we have and then deciding whether we should accept or reject that impulse. So during the pandemic, when I started wearing face masks and I was around other people, what would often happen is I would approach someone to talk to them and I would immediately find myself having this impulse that I should pull down my mask so we can talk easier, and normally that makes a lot of sense right? In normal, non-pandemic circumstances having a mask does impede conversation, so pulling down your mask or removing your mask makes perfect sense, but during a pandemic it's the opposite of what you want to do in a conversation. You want to protect the person you're talking to and maybe protect yourself from the person you're talking to. So I had this impulse that I had to stop and reflect and remember, "Oh I remember why I'm wearing this mask", and in this situation I need to keep it on because the mask is protecting both of us and so that stopping, and then consciously reasoning through something is what I what I consider reflection.
So it involves those two things, where you don't accept your initial impulse and instead, you sort of reason about it or consider alternative impulses.
Sahana: Okay. That makes quite a good foundation for us to move to the next question, which is about measuring reflection. So, if you have a bunch of people who have a variety of responses to the measure to wear a mask, some would say "Yes, it's important to wear a mask to reduce the probability of transmission" or else another would say "No! It restricts freedom." So, could you tell us if - only if it applies in this case, of course - could you tell us how reflection could be measured for these different people? Please feel free to use any other instance you want.
And specifically, how is reflection measured using verbal report protocols and process dissociation?
Nick: So, reflection testing goes back, I guess, almost a couple decades at this point, maybe actually a little further if you include some of the logic puzzles that people have created to test reflection. So those reflection tests, from a while back, the way they measured reflection was they came up with questions or like, stories that are designed to get people to have a particular impulse or response to the question, that's false or wrong or something like that and so I'll give you an example. It's a variation on a recent reflection test.
So it goes like this: “If it takes two nurses two minutes to test two people for COVID-19, how long would it take 200 nurses to test 200 people for COVID-19?”
So, because of the pattern of the question, a lot of people are lured into saying "Oh well the answer is 200 because it was 2, 2, and 2 in the first half of the question, and the second half gave us 200, 200" and so you actually want to fill in 200. So that's the response that we're “lured” into. It's the impulse that we have, but upon reflection you can actually realise the math doesn't work out that way. The math would actually give us the answer of 2. Questions like this are called "reflection tests" because they “lure” us into a particular response that, upon reflection, we can realise is incorrect.
And the key to these tests is they have demonstrably correct and incorrect answers. So we couldn't do this necessarily with questions about, I don't know, ethics, where maybe there's some disagreement about what the demonstrably correct answer is. And so we give people these questions and based on that their tendency to respond with a lured or reflective response, we can gauge how likely they are in general to reflect or not reflect when given ordinary questions.
And you might think that we could come up with pandemic versions of these questions, so I sort of changed this question to be about COVID-19 testing. The original question is about blood testing. And you might think there's other types of variations we could use to test these things and you might also think that the way people reason about these questions probably tells us something about the way they reason more generally, about other things.
So people who are more likely to stop and question their initial response might be different than people who just tend to accept their initial response even if it's demonstrably incorrect.
People do find some correlations between people's reasoning about, on reflection tests and their other beliefs or their other attitudes or their other behaviours or intentions. Some have found that, for instance, people who are more likely to reflect on these tests are better at determining true and false news stories about COVID-19.
So the good news is that people - this kind of easy thing that we can do by stopping and questioning our initial impulse and then reasoning about it - that simple thing can actually have some pretty big benefits, like maybe avoiding misinformation about really important things like a global pandemic.
You also have some other questions about thinking aloud and Processed Association, and these more advanced methods of measuring reflection. So, I gave you that original question- If it takes 2 nurses 2 minutes to test 2 people for COVID-19, then how long would it take 200 nurses to test 200 people for COVID-19? A lot of these questions are given to people online and so when somebody's on a phone or a computer or somewhere in the world and they're clicking buttons or typing and entering something on an online survey. And what the researchers receive is just their answers, whatever they typed whatever they clicked.
Maybe they'll get a little bit of information - like how long it took them to click and type or something like that. But what they're not getting is the entire thought process. From reading the question, to answering it and what we have found is some people can come to this so-called "correct" or "reflective" response on these test questions, without actually reflecting. Like it just immediately comes to mind, the correct answer immediately comes to mind without reflection. And then some other people sometimes come to the incorrect or "lured" response even though they spent a while reflecting on the problem. Which tells us that these reflection test questions might not always be measuring reflection. They might be measuring something else, like how familiar people are with whatever the question is about, right.
So people who are really familiar with blood testing or COVID-19 testing might be better at that question and they might have the answer correct. They might have the correct answer immediately as opposed to only after reflection.
One way you could try to avoid getting these kinds of false positive or false negative answers from people, is to just have them think aloud throughout the entire test. So we did a couple studies where we did this, both in person and online, where we basically said "Here's what's going to go on during the study. We're going to record the whole thing and we just need you to say everything you think, no matter whether you think it's related or unrelated. Like when you're reading the instructions, read them out loud. If you get distracted and see a squirrel and a tree outside just tell us about the squirrel. Like everything that comes to mind. Just say it.” And we have people complete the reflection test and what we find is a lot of people do come to the correct response without reflecting. It's a minority of people, but a substantial minority. And a lot of people do come to the incorrect "lured" response even though they thought through the problem multiple times and so, again it's a minority, but a substantial minority.
But, hearing them think out loud is what allows us to detect these false positives and these false negatives. So we can, in the data analysis, potentially control for these false positives and false negatives and therefore have a more accurate score for how whether or not people were reflecting and until we get people to think out loud and collect that audio data, it's really not clear how much we should be relying on reflection test scores that don't include that kind of information.
Process disassociation is another type of tool you can use where basically, you use two different types of reflection test. One type is the normal type. You get lured towards a particular response that's incorrect upon reflection and then the other type of reflection test is what we call a Congruent reflection test, where both the "lured" response that you're lured towards and the correct response is the same thing. And so we basically make sure we lure people toward the correct response on that other type.
And then you look at the probabilities of how people respond to each of those two types of questions - the normal incongruent type and this new congruent type and you can mathematically dissociate reflective from unreflective reasoning. This was actually originally used by cognitive psychologists to dissociate implicit from explicit memory but it's been used for lots of other things as well. There's those two methods - getting people to think out loud and using process dissociation are two methods for improving reflection testing.
Sahana: Okay. So, now that we have talked about measuring reflection, on a related note I think you had also mentioned and talked about the default Supplementarist Model of making decisions and the idea of Bounded Reflectivism. How would these relate in this situation?
I don't know how technical this is, so, please feel free to tell me in case we should be skipping this in case you feel like it’s going to be too much for our audience.
Nick: Good. So, we've already talked about how reflective reasoning could be related to lots of good things right. So, when I'm doing math I sometimes write out my first calculation and and then stop and reflect and
wonder if I did it correctly and then double check my work and then find sure enough, I did make a mistake, and it was that reflection that allowed
me to correct my mistakes. There's probably tons of examples in life where reflection leads to better reasoning. And so philosophers have taken this thought and run with it. They think reflective reasoning is probably going to be crucial to all sorts of good thinking and philosophy. The more likely we are to reflect individually and socially together like this, the more likely we are to come to better agreements about what we should think about ethics or politics or something like that. That sounds intuitively plausible and maybe you even have some anecdotes of this happening, but some social psychologists actually find that the more reflective people are the more polarized their answers are about certain things, like in politics. So it's the most reflective people that are on the extremes of when it comes to their answers on certain political questions. That suggests that actually, reflection doesn't always lead to this kind of progress towards, and convergence on some sort of particular answer. And you might even think some types of polarization are actually bad and so reflection isn't always this panacea. It might actually be part of something we don't like. In response to this, these conflicting intuitions about reflection; on the one hand reflection being great and on the other hand reflection maybe sometimes having some shadow sides, I came up with this concept I'm calling Bounded Reflectivism, which basically just tries to respond to two different views about reflection and philosophy.
One is the view that reflection is great it's crucial for good reasoning, and the other view - that you find in philosophy sometimes - is reflection doesn't do any of the stuff that these other philosophers think it can do. It gives us the sense that we're reasoning better but that's all it does. It's just apparent, it's not really making a reasoning any better.
Bounded Reflectivism is a middle way between those views. It says, "Look reflection is great sometimes but you're right there's some evidence that reflection has some we should be concerned about what reflective reasoning might do in certain situations" and so what we need to do is figure out precisely when and under what circumstances reflection has value and try to rely on reflection during those circumstances, while being aware of the circumstances in which reflection can lead us astray or lead to polarisation or maybe other undesirable outcomes and try not to rely on reflection in those circumstances. It's a more strategic use of reflection as opposed to a Yay reflection! or a Nay reflection!
Sahana: Okay. So what you just said brings me to the paper; I think it was also there in the dissertation post - about how differences in reflection predict differences in philosophical beliefs and judgements. So, in the pre-print that you had shared, you focused on how messages urging people to follow simple protocols portrayed victims and also the cognitive style and philosophical beliefs of the receiver. So could you please tell us a bit more about it about what you found in those experiments?
Nick: Right. Yeah, in a few experiments we've looked at how reflection test performance - the tests that I mentioned a little while ago - predict all sorts of things. So, among lay people they predict their philosophical judgements and inclinations. For instance, a lot of papers find that more reflective people are more likely to be atheists or agnostics, for example. There's actually some more recent evidence suggesting that that isn't true in every country, so there's maybe more to reflection and religiosity than than I just suggested. But there are some interesting correlations between reflection and religiosity. And we wanted to, during the COVID-19 pandemic, see if reflection test performance would tell us anything interesting about people's responses to the pandemic and so we thought, one way that this might be relevant to the pandemic is in people's responses to public health messaging, right? So public health officials are often pointing at these graphs, called "Flatten the Curve" graphs. They show two curves - a high curve and a low curve and they say "Look the high curve is a lot of deaths and that represents what happens if we, if governments don't intervene on people's lives and like limit travel and some liberties about you know facial masks and other things”. And the low curve is what is how many deaths we would expect if we do intervene on people's lives and and curtail people's liberties in certain ways and maybe interfere with economics and business and trade a bit. We were wondering if certain people are going to be more swayed by these flatten the curve graphs than others. So you might think that people who score well on reflection tests are the types of people who tend to do better on statistical reasoning tests or just general math tests and so you might think they'll like these graphical representations, these "Flatten the Curve" graphs and they'll be more swayed by them because they're "mathy" or something like that. Whereas another type of public messaging strategy that people use is appealing to individual victims. So, a journalist might tell a story about a particular person in a hospital and how this is impacting their family and their business and you know, or someone has died of COVID-19 and you look at the the survivors of that family and how they're coping with this death.
This is focusing on an individual victim as opposed to statistics about lots of victims, and you might think, those two different types of messaging will have a different impact on people. Social Psychology has theorised that they will so there's this this thing called the Identifiable Victim Effect, and the idea is people donate more money to causes when they find out about the cause through a story about an individual victim related to the cause,
as opposed to when you just say like you know 13,000 people were killed in some country uh you know at this time because of such and such.
Hearing about individual victims does tend to be more motivating to people than hearing about statistical victims. In the past at least that's what we found. When we ran these studies with COVID-19 we used Flatten the Curve graphs and stories about an individual victim in a hospital hooked up to a respirator and how this is going to impact their three children and the rest of their family. We didn't find that people's responses were different between the individual victim and on the respirator and the Flatten the Curve graphs. So what we were measuring is how likely people are to comply with public health recommendations around mask wearing, hand washing, disinfecting, sheltering in place and you know the four or five or six things that most public health officials think we should be doing. It turns out that people were just as likely to do those things, regardless of whether they got the Flatten the Curve graph or the Individual on a Respirator.
Now, again, we thought reflection test performance might explain some of the differences in these responses as well. So, I thought maybe people who are more reflective will be more responsive to the Flatten the Curve graph than the individual victim messaging and we didn't really find that. More reflective people were just as likely as less reflective people to wear masks and shelter in place and all the rest. So this told us that it doesn't seem to be like ineffective messaging or ineffective reasoning that's leading to differences in compliance with public health recommendations. It's something else. So, in our studies, we actually included some questions about philosophy because that's kind of the game I'm playing in academia - studying people's philosophical judgments. What we found is some philosophical judgments definitely predicted the differences in compliance with these public health recommendations. So the best predictor of non-compliance with mask-wearing and sheltering in place and these other things is endorsing Libertarianism, or I guess, as we measured it, the best predictor of that non-compliance is valuing liberty more than equality. And the best predictor of compliance was endorsement of a particular principle you might call an Effective Altruist Principle. It's from a paper by Peter Singer in the 70s, called "Famine, Affluence, Immorality" and the principle just goes like this -
If we can prevent great harm, without incurring as much harm, we should.
So the more that people agree that we should prevent great harm whenever we won't incur as much harm by doing so, we should - the more people agree with that, the more likely they are to comply with public health recommendations. There were also some other correlations in there, but those were the two most potent philosophical predictors of compliance.
Sahana: Okay, I was just wondering. Like, how does the use of individualised victims.. So we had an interview with Christopher Tindale earlier and we were talking about rhetoric and could this just be like the use of a victimised individual because they feel connected to the narration, the story? Could this just be the use of rhetoric?
So how is this related to rhetoric? The statistics - if you had included individual stories along with the statistics - how do you see rhetoric really playing out in this case?
Nick: Yeah, it's a good question. There is some sense in which messaging is marketing and therefore rhetoric, right? So marketing is largely about persuasion and whatever it takes to get someone to act on something. And so there is a sense in which all messaging - at least all good messaging - is a form of rhetoric. We do want people to respond in a certain way uh and then the question becomes, How does the rhetoric work? And that's less clear to me. So there's some evidence, again, suggesting that people respond more to individual victims than they do to statistical victims. There's also some evidence suggesting that people respond more to victims that are close by than further away. In the U.S. whenever there's some sort of crisis that affects people here, lots of people will respond and give lots of money. When there was a hurricane - well there's often hurricanes here - when a hurricane impacts some country in some sort of negative way, lots of people give time and energy to those people who are suffering and dying here. Or who are just..maybe they're not even dying maybe. They're just without homes or something, like that. We do lots here to help those people. When the same or even worse happens to even more people in another country, you don't see as many Americans giving as much or or traveling there to donate their time and energy. But they're both on the news and so we're both getting these depictions from the news of individuals who are suffering and dying somewhere and we respond to the ones who are closer to us than the ones who are further away. So you might think rhetoric aside, there's something about proximity to us that that seems to have some motivational power. And it's not clear that that's a good thing, right? So there might be far more suffering and death outside the U S. and this actually wouldn't be surprising that there was far more suffering and death outside the U.S. than within the U.S. It would be a bummer if people in the U.S. were more likely to help other people in the U.S. than they were abroad
Sahana: That's true. So before we move on to the pandemic related public philosophy questions, I thought of asking you a quick question about Well-Being in a Digital Context today. You have written about Causal-Network account of well-being and how it applies to ill-being, particularly depression in digital context. So right now the larger slogan seems to be that we are all socially distanced but virtually connected. So, how do you think the Causal Network Account would sort of talk about our ability to stay connected virtually and how it positively contributes to our well-being during the lock-down especially.
Nick: So, the Causal Network Account of either well-being or ill-being is basically what you would think when you think of a causal network - there's going to be little dots or nodes that you might think of, with lines between them representing like correlations or you know causal effects or something between these nodes, right? So you might think Happiness is a node and motivation is a node and like Finances are a node and Social Support is a node and you might think all of these things are interrelated, right, like the more social support I have the happier I am, the more motivated I am, the more financial resources I have and the more of those two things that I have, the happier I am, and things like that. But there's going to be lots and lots of other nodes on this network that are going to have lots and lots of other complicated relationships. But the idea is, there's a lot of factors that are related in complicated ways that explain
our well-being and ill being. So now when you think about the pandemic, you rightly pointed out, the pandemic is having an impact on lots of these nodes, right? So my social support might be.. I might feel like I have less social support right now because I'm not going to my office anymore and seeing my colleagues. I'm just staying at home. I have great social support with my cat but not nearly as much social support with uh you know my colleagues and superiors. Or just friends, right? I don't go out really with friends either because of the risks related to that. I will call them or or do this with them .. and so my social support takes a hit and as a result of that might affect my well-being in other ways and make me less well-off.
And then obviously people are losing their jobs, and losing sources of income or businesses are surviving but on, agaiṇ, with less revenue, and so people are having to take cuts to the amount of money they have available to themselves. And so that also has an impact on people's well-being.
And so the causal network account is just going to, basically, point to the various nodes that are being affected during the pandemic and then explain differences in well-being and ill-being as a result of these changes in these nodes and how they affect other nodes in the network like happiness and social support and financial support and motivation, things like that.
Sahana: Okay! So, now let's move to some philosophical questions which are related to public practices. So, quite early on when the situation hadn't even been declared a pandemic you were quite proactive and had talked about online conferences; especially the Minds Online Conference results. So if we have academic philosophers, you know, hoping to do online conferences, who are watching us right now, maybe they are a bit sceptical. Could you please share how you feel that online conferences might be better than the traditional conferences and also if there are any easy to use or if Open-Source software apps that they can use for this purpose?
Nick: Yeah, so .. right up front, we organised a conference for three years. It was an international conference. It was quite a few speakers over the course of weeks and it cost us .. whatever the cost of a typical website would be so like maybe five ten bucks a month or something like that. And that was the .. the cost of the entire conference right, so .. all of the labor we put into organising it was volunteer. All of the reviewing was volunteer. All the keynotes speaking was volunteer. All the commenting was volunteer. Over thousands of people participated in the conference ..roughly every year. So- a massive international conference happened for about five dollars a month.So that alone I think suggests that .. reasons to at least take seriously the possibility that we should be doing more conferencing online.
But there's there's just way more benefits even than that. So we found that because people can submit a paper from anywhere regardless of their ability to get on a plane and leave their family, women are more likely to submit and therefore more likely to be accepted, because they're less, you know, they're less.. the people who are more involved in child care are less inhibited by an online conference because their kid can be in the other room or right next to them. Sometimes I talk to colleagues and their kids in their lap, but you know, I love it. There is not .. as much barrier to entry for certain people who are more impacted by things like child care. Also people from other countries can submit because you know people from every ..Not every country offers their universities' researchers as much money to travel around, and so lots of countries are just [inaudible] conferences in the world. But at a conference, free to attend or present at? That barrier is removed, right? So there's various ways in which inclusivity and diversity is increased by having a conference online. And so insofar as we think that's important we should probably prefer an online conference. And then you learn during the online conference, I think it was 2015/2016/2017 when we were doing this is, there was a hurricane every single year during the conference and so all of the organisers were like not at home. They were usually in a hotel or at a, you know, a family friend's house like a few hundred miles away, while their backyard is just being destroyed and it's fine because that's how the internet works you just need to go somewhere that has the internet and it turns out there's a lot of places that have internet .. and so as long as you can evacuate .. and so we ran the conference - Me, from usually a friend's bedroom, and it was it just worked out totally well and you might think"Wow a pandemic", that would be pretty handy to have that type of emergency resilience. So those are some of the major benefits and like i said it was only about five dollars a month and the five dollars basically goes towards hosting a website. You can even do this for free. So if [inaudible] WordPress has a free website service where it'll just be you know, your-web-address.wordpress.com and that'll be a free website and you could host a conference on that WordPress website.
Yeah, so that's the overview then.. all of the costs - which are almost nothing - and all of the major benefits - which seem drastic compared to a traditional in-person conference.
Sahana: Okay..that's pretty much the questions that I wanted to ask you. Thank you so much for joining us Nick! It's been great talking to you!
Nick: Thank you so much for having me- this has been a lot of fun.
Sahana: Thank you!