Why Some Emergency Alerts Fail — New Research Explains What’s Missing - Part II Discussion

WEBVTT
00:00:06.320 --> 00:00:07.759
Hi, I'm Jeanette Sutton.
00:00:08.080 --> 00:00:09.519
And I am Eddie Ritzolena.
00:00:10.000 --> 00:00:13.119
And we welcome you to another episode of the Alerting Authority.
00:00:13.199 --> 00:00:19.600
And as always, we encourage you to subscribe, follow, listen, and most importantly, participate in these podcasts.
00:00:19.839 --> 00:00:30.160
We want your questions, concerns, ideas, problems, pain points, and success stories so we as alerting authorities can better do our jobs and make our communities safer.
00:00:30.640 --> 00:00:30.960
All right.
00:00:31.039 --> 00:00:37.439
And for our sponsorship, I'm not going to use a real sponsorship for this one, folks, but I want to tell you it's my little soapbox.
00:00:37.679 --> 00:00:40.640
It's continue to get educated.
00:00:41.119 --> 00:00:41.439
Okay.
00:00:42.079 --> 00:00:44.479
And continue to be a student.
00:00:44.960 --> 00:00:55.840
And wherever you are, whether you've completed your high school diploma or whether you are, you know, already a PhD, just keep that student mindset going.
00:00:56.079 --> 00:01:02.640
And if you if you heard the first part of what we're going to talk about today, and I know this is the second, and Jeanette will talk about it.
00:01:02.880 --> 00:01:06.319
If anything, it should spark that interest in you, folks.
00:01:06.480 --> 00:01:09.359
So the sponsorship is sponsor yourself.
00:01:09.680 --> 00:01:14.159
Take that next class, take that next step, learn about these things.
00:01:14.319 --> 00:01:23.599
If you are confused about a single word in this, what we're talking about today or past episodes, use that as a goal just to try to learn a bit more.
00:01:24.000 --> 00:01:26.480
So the sponsor is you.
00:01:26.959 --> 00:01:27.280
Okay.
00:01:27.680 --> 00:01:29.040
Take that next step.
00:01:29.280 --> 00:01:38.799
Um, and I'm saying this for myself as well, because I literally'm doing this right now, and you guys are hearing this, but I have to go back.
00:01:38.879 --> 00:01:44.640
And it it really sparks interest, questions, and it gives me that opportunity to learn more.
00:01:44.799 --> 00:01:47.280
So that's our sponsorship for today.
00:01:47.599 --> 00:01:48.480
I love that.
00:01:48.719 --> 00:01:50.640
I am a lifelong learner.
00:01:50.959 --> 00:01:58.079
Every day I feel like I'm getting a new academic article and looking for information that's going to help me to do better in my work.
00:01:58.319 --> 00:02:02.560
And I know that people are cracking open books and reading articles all the time.
00:02:02.719 --> 00:02:06.799
So investing in yourself and your knowledge is always a great thing.
00:02:07.120 --> 00:02:13.439
So we want to welcome back today to the podcast, Hugh Walpole, Dr.
00:02:13.520 --> 00:02:24.560
Hugh Walpole, who, if you listened previously, you met him and you know how much, how interesting and how deep we can go on subjects about research.
00:02:24.719 --> 00:02:26.639
And we're going to return to that today.
00:02:26.800 --> 00:02:30.319
Um, Hugh is a consulting research scientist for the Warren Room.
00:02:30.719 --> 00:02:43.840
His research focuses on how people perceive risk and respond to risk communication associated with a wide variety of natural and technological hazards and psychological measurement of risk-related concepts.
00:02:44.159 --> 00:02:50.240
Hugh received his PhD in environmental social sciences from the Ohio State University.
00:02:50.639 --> 00:03:01.919
And he and I worked together at the University of Albany, where he was part of my research team within the emergency and risk communication message testing lab.
00:03:02.479 --> 00:03:09.599
And um, we just had so many studies we've conducted, analyzed, and wrote together.
00:03:09.919 --> 00:03:18.080
We wanted to spend some time in this episode talking about the new most recent publication that came out of that lab.
00:03:18.240 --> 00:03:22.080
And I am very excited because it's some groundbreaking information.
00:03:22.159 --> 00:03:24.400
So y'all are gonna be the first to hear about it.
00:03:24.639 --> 00:03:36.159
But we're gonna start by asking a question about risk perception because this is one of Hugh's um areas of research that he has published on.
00:03:36.400 --> 00:03:40.719
Um, and most probably most importantly, it was his dissertation research.
00:03:40.879 --> 00:03:43.439
So he knows a lot about risk perception.
00:03:43.680 --> 00:03:52.800
Um, can you introduce yourself real briefly and then tell us about the importance of risk perception and how it's related to alerts and warnings?
00:03:53.199 --> 00:03:54.000
Yeah, absolutely.
00:03:54.159 --> 00:03:55.199
And thanks for having me back.
00:03:55.360 --> 00:03:55.919
Great to be here.
00:03:56.080 --> 00:03:57.599
Um I'm Hugh Wobble.
00:03:57.680 --> 00:04:04.960
Um I conduct a lot of research, as Jeanette said, on uh how people think and feel about risk from a wide variety of different hazards.
00:04:05.039 --> 00:04:13.360
Um and a big part of that work, and and we're kind of the genesis of me coming into that work was through the concept of risk perception.
00:04:13.599 --> 00:04:20.800
So uh there's a lot of definitions for risk perception that kick around in the world, but um it's mostly exactly what it sounds like.
00:04:20.959 --> 00:04:23.759
It's how people perceive the risk associated with something.
00:04:23.920 --> 00:04:39.839
So if we wanted to think of that in general terms, um we might think of it as how bad someone thinks uh the outcomes associated with some hazard, whether that's a a wildfire or whether it's a power outage, um, how bad that is likely to be.
00:04:40.000 --> 00:04:54.079
Um generally speaking, we think of this as some combination of how likely something bad is to happen, uh, how bad that would be if it did happen, and then uh how bad it feels, how scary it is, that visceral sense of risk.
00:04:54.240 --> 00:04:59.519
Um and so we usually think of perceived risk as some kind of combination of those things.
00:05:00.399 --> 00:05:04.319
So there's like an emotional element to risk perception too, right?
00:05:04.720 --> 00:05:05.199
Absolutely.
00:05:05.360 --> 00:05:05.600
Yeah.
00:05:05.759 --> 00:05:25.439
So uh there's a lot of research on kind of the the relative position of where that belongs in uh in a process of perceiving risk and how important it is relative to that analytic sense of how likely is this and how bad would it be, that more um yeah, that more analytic sense of of how those go.
00:05:25.680 --> 00:05:34.959
But most theories uh in the modern era of risk perception uh include both of those elements and give them a fair amount of prominence um kind of each.
00:05:35.120 --> 00:05:39.120
They they one is not considered to be more impactful than the other.
00:05:41.040 --> 00:05:50.720
So when I think about like scary messages related to risk perception, if I give people a really scary message, is that gonna get people to take action?
00:05:51.199 --> 00:05:53.680
It's it's really tempting to do that, right?
00:05:53.759 --> 00:06:00.319
Because a lot of the time the thing that we're really worried about is giving somebody uh a message and uh having them say, like, ah, it's not that scary.
00:06:00.399 --> 00:06:01.519
I'm not gonna do anything about it.
00:06:01.600 --> 00:06:09.120
And we're really worried about not reaching those people with information about how risky some of the hazards that they might be facing really are.
00:06:09.439 --> 00:06:12.399
Um and I think that makes it makes a lot of sense.
00:06:12.480 --> 00:06:14.399
And we do need to include that for a lot of people.
00:06:14.560 --> 00:06:24.480
A lot of the time, it's really critical to include those pieces of information that help people understand how they're at risk and how they could be at risk if they, even if they think that they're not.
00:06:24.720 --> 00:06:26.399
But that's not the whole story.
00:06:26.560 --> 00:06:37.360
We have a lot of great theory from the last uh 30 years or so that spends a lot of time thinking about um what other components of risk messages are required.
00:06:37.439 --> 00:06:41.680
So when do risk messages fail, even when they activate people's risk perception?
00:06:41.839 --> 00:06:43.279
They make them really scared.
00:06:43.519 --> 00:06:46.800
Why doesn't uh people kind of act on that information?
00:06:47.040 --> 00:06:55.519
And one of the kind of pieces of that theory is that if you give people something to be scared of, you also have to give them a way to get themselves out of danger.
00:06:55.680 --> 00:06:59.920
You have to provide them with information that helps them feel like it's not hopeless.
00:07:00.079 --> 00:07:06.000
There's a way for them to confront this risk, as scary as it might be, um, and that they're capable of doing that.
00:07:06.160 --> 00:07:13.439
So we talk about a couple of different concepts in that space, talk about self-efficacy, so a sense that there are things that I can do.
00:07:13.600 --> 00:07:17.759
I am capable of taking an action that will help reduce my risk.
00:07:17.920 --> 00:07:28.240
And then we talk also about response efficacy, um, which is essentially just how effective do I think recommended actions will be at helping me reduce my risk.
00:07:28.399 --> 00:07:43.680
So if you have high risk perception, you might feel motivated to do something, but unless you believe that you can do something, so you have self-efficacy, and you believe that that thing will be effective, so you have response efficacy, then you're not as likely to take an action on that behalf.
00:07:43.839 --> 00:07:56.000
It can make people fatalistic if you give them an enormous amount of risk perception, you boost their risk perceptions really high by telling them something is really scary, it's gonna impact them in a really serious way, and you don't give them a way out.
00:07:56.160 --> 00:08:01.519
You don't give them an action that they can take in order to kind of help them get out of that situation.
00:08:01.680 --> 00:08:12.399
Um, so that's one of the things that we're gonna be talking about, I assume, is is the importance of guidance, thinking about how important it is in risk messages to include information about what people can do to protect themselves.
00:08:13.199 --> 00:08:26.560
As we were talking about in the last part, they may not always do exactly that, but for the people who don't know what to do, it's absolutely crucial for them to have that sense of there's something I can do to reduce this big risk that I'm facing.
00:08:26.879 --> 00:08:27.120
Yeah.
00:08:27.279 --> 00:08:29.519
So we're gonna go off script for just a second.
00:08:29.600 --> 00:08:40.320
I'm gonna ask you a question about wildfire because this is one of those places where the messages that go out can be really scary, especially if there's a lot of environmental cues around you.
00:08:40.480 --> 00:08:56.000
But a lot of times there's there's jargon that's used to communicate about wildfire where the perception of risk has really increased, but the jargon is trying to do a lot of things in one statement, like evacuation warning.
00:08:56.639 --> 00:09:03.679
And for some people, evacuation warning might mean to them, okay, you're warning me that I need to evacuate right now.
00:09:04.960 --> 00:09:08.879
So the implied guidance is to leave.
00:09:09.440 --> 00:09:14.000
But we know that that jargon, evacuation warning, does not mean to leave.
00:09:14.159 --> 00:09:16.000
It actually means to prepare to leave.
00:09:16.240 --> 00:09:24.159
So it's it's raising people's perceptions of risk without if you don't give them that guidance that goes along with that jargon.
00:09:24.240 --> 00:09:27.440
It leaves people wondering, well, what do I do?
00:09:27.679 --> 00:09:30.000
And you know, we've we've written a lot about this.
00:09:30.080 --> 00:09:31.840
So what are your thoughts?
00:09:32.399 --> 00:09:45.679
Yeah, I think particularly in the case that you're describing, we've done work specifically looking at um evacuation warning versus evacuation order as ways of describing um kind of recommended guidance for a wildfire.
00:09:45.759 --> 00:09:56.399
And it can get pretty pernicious because a lot of the time the term evacuation warning or evacuation order are used kind of in lieu of describing what the risk actually is.
00:09:56.559 --> 00:10:11.600
So a lot of the time, the heavy lift that it's doing, the jargon, in terms of guidance, it's trying to tell you to either um leave if you need additional time, leave immediately, or prepare if you don't need additional time, versus leaving immediately for an evacuation order.
00:10:11.759 --> 00:10:19.759
You're asking it to do the lift of that guidance, but you're also not including information about whether what what kind of hazard people are facing.
00:10:19.919 --> 00:10:23.759
So in many cases for wildfire messages, the word wildfire isn't used.
00:10:23.919 --> 00:10:28.320
Um, they just describe, hey, if it's a it's an evacuation warning, it's an evacuation order.
00:10:28.559 --> 00:10:32.879
And I think it can be tempting to think like, oh, come on, what does it matter?
00:10:33.039 --> 00:10:35.200
Uh it's a hazard, we want people to evacuate.
00:10:35.360 --> 00:10:37.519
That's the thing that that we really want to focus on.
00:10:37.679 --> 00:10:44.799
And I can sympathize with that perspective because what you're asking them to do is not really dependent on what hazard they're facing.
00:10:44.960 --> 00:10:47.600
You just want them to get out of there to keep themselves safe.
00:10:47.759 --> 00:10:56.159
Um, but kind of pertinent to the conversation that we had last time about compliance, there's a lot of different things that people can do to keep themselves safe.
00:10:56.240 --> 00:10:58.799
And it's not always exactly what we tell them to do.
00:10:58.960 --> 00:11:14.240
And so if a person finds themselves in a situation where knowing that the threat that they're facing is a fire could help them take action more safely than if it was something like a windstorm, a flood, a landslide, or something like that, that you might ask people to evacuate for.
00:11:14.399 --> 00:11:16.320
Obviously not a windstorm, but you know what I mean.
00:11:16.639 --> 00:11:25.120
But knowing that a fire is in a given location, it might make traffic bad, it might block egress routes, it can behave in ways that are different from other hazards.
00:11:25.360 --> 00:11:30.559
It can help people contextualize the information that you're giving them about guidance and actually implement it.
00:11:30.799 --> 00:11:35.440
And unfortunately, that requires that they actually know what you're trying to tell them to do.
00:11:35.679 --> 00:11:46.639
And so in the case of evacuation warning versus evacuation order, we did some research and we developed some evidence that people don't really understand the difference between these terms very well.
00:11:46.879 --> 00:12:12.720
And in some cases, this may not matter a great deal, but there are edge cases where if you have uh a group of people that you want to stage for an evacuation, but your egress is limited enough, you have a small town that has one road that goes in and out, and you have to be able to get everybody out along that one road, you're gonna want to stage that evacuation process so that you can get people out without choking the roads with traffic or causing accidents along the way.
00:12:12.879 --> 00:12:17.120
And so being able to stage people to get them ready to go on a moment's notice is really important.
00:12:17.279 --> 00:12:26.000
But if those people are also interpreting that guidance as leave immediately, then you're gonna end up with a more chaotic evacuation that's harder for you to manage and stage.
00:12:26.159 --> 00:12:32.399
Um, again, I'm not a wildfire expert and I haven't been in the situation rooms doing those logistical decisions.
00:12:32.480 --> 00:12:35.519
So this is a presumption that I'm making about a challenge there.
00:12:35.840 --> 00:13:01.759
But there are many other kind of terms that we could use if we move past using jargon, we move into using plain language that allow us to say what we mean and allow us to give people the best chance to interpret the guidance that they're given and then make a judgment on that basis, both in terms of knowing what hazard it is that they're actually facing, but also what kinds of behaviors are, in our view, kind of optimal for keeping them safe in the event that they don't already know.
00:13:02.720 --> 00:13:10.320
Well, and it touches on something that we we didn't talk about in our last episode that I do want to bring up is the concept of milling.
00:13:11.279 --> 00:13:22.879
And milling is one of those things that the warning response model calls out as being present for in response to every alert and warning message that someone receives.
00:13:23.039 --> 00:13:41.679
It is the process that we as social individuals go through to confirm the information, to confirm what is happening around us, to confirm that we should be paying attention to the message, um, to figuring out what we should actually be doing in response to the message.
00:13:41.840 --> 00:13:46.559
And milling is it delays action.
00:13:46.960 --> 00:14:09.279
And when a message is incomplete and it simply says something like evacuation warning, zone 25A, like really incomplete, not giving people a lot of information that's really easily consumed if you don't have pre-education and awareness of the jargon and of the location.
00:14:10.320 --> 00:14:15.440
That process of looking for information to confirm is gonna delay action.
00:14:15.519 --> 00:14:20.799
And that's where these incomplete jargony messages become really dangerous.
00:14:20.879 --> 00:14:23.360
Like you've increased people's perceptions of risk.
00:14:23.919 --> 00:14:26.960
And now you want people to take action immediately.
00:14:27.679 --> 00:14:31.600
Like you gotta give them a complete message so that they can actually do that.
00:14:31.759 --> 00:14:41.840
Um even we know milling's always gonna happen, but reducing the amount of time people spend looking for information is like a real key here to helping people to go.
00:14:42.480 --> 00:14:55.279
Yeah, I think it's it's the most natural thing in the world to when you're trying to figure out what I'm supposed to do in this situation, just to look at what other people are doing, um, and then to look for information on whether or not there's guidance about what you're supposed to be doing.
00:14:55.600 --> 00:15:05.919
So knowing that that's the most natural thing in the world, but also knowing that delays can be really costly, it really behooves us to try to reduce that delay, that decision delay, as much as you can.
00:15:06.159 --> 00:15:15.039
Um and that can uh that can best be facilitated, as far as we know, by including that complete information in those messages.
00:15:15.360 --> 00:15:25.840
Okay, well then kind of going back and incorporating what we just did and talking about the future, and this is kind of an all-around can we go over biases for a second or biases, right?
00:15:26.320 --> 00:15:29.279
Um we we talked about risk perception.
00:15:30.159 --> 00:15:35.360
And and I I know those two things are not the same, right?
00:15:36.559 --> 00:15:46.799
But but it seems like the biases that that can exist can play a pretty important factor and can influence a lot of that, the actions or inactions.
00:15:47.279 --> 00:15:48.879
Um so I guess so.
00:15:48.960 --> 00:15:55.919
My question is can you help identify the differences between those two things of risk perception and then the biases?
00:15:56.000 --> 00:16:05.200
And then is there a way as practitioners by following the principles that we're talking about that we we can affect people's biases?
00:16:06.080 --> 00:16:08.080
Yeah, so I think this is a really good question.
00:16:08.240 --> 00:16:13.440
Um we talk about biases a lot in risk research, decision research in general.
00:16:13.519 --> 00:16:17.519
So just thinking about how people make decisions and how they kind of focus on risky things.
00:16:17.759 --> 00:16:19.919
Bias is something that that turns up a lot.
00:16:20.159 --> 00:16:22.159
Um I think you're spot on.
00:16:22.320 --> 00:16:24.559
They're not quite the same thing as risk perception.
00:16:24.639 --> 00:16:32.080
We could think of relying on your risk perception entirely to make a decision might constitute a bias of some kind, but we'll talk about that in a sec.
00:16:32.639 --> 00:16:40.240
Um so they are different, but I would say people's biases affect their risk perception, and their risk perception can affect how their biases affect their actions.
00:16:40.559 --> 00:16:43.840
Um so we're gonna get super, super nerdy here for a minute.
00:16:43.919 --> 00:16:47.600
So um just derail me if I go off the rails a bit.
00:16:47.759 --> 00:16:51.039
But generally speaking, awesome.
00:16:51.600 --> 00:17:03.279
Generally speaking, we we we think about biases as coming about from an over-reliance on simplifying rules for decision making that we call heuristics.
00:17:03.440 --> 00:17:17.759
Um so in decision research, a lot of the time, when you're faced with a decision, making an objective decision to do something like should I evacuate or not as a result of this hurricane, answering that objectively is really, really hard.
00:17:17.920 --> 00:17:20.720
Um, you have to know a lot about hurricanes in general.
00:17:20.880 --> 00:17:27.519
You have to know a lot about this hurricane in particular, you need to know a lot about your home's resistance to hurricanes.
00:17:28.000 --> 00:17:32.240
You know, is it is it able to kind of withstand the wind speeds that you're gonna get from those kinds of things?
00:17:32.400 --> 00:17:37.920
Uh, traffic conditions, uh, the direction that the storm's gonna go, which requires you to be able to see into the future.
00:17:38.160 --> 00:17:41.759
You need to know all of those things if you want to make a rational decision about that.
00:17:42.000 --> 00:17:51.039
Um instead, you can rely on a simplifying question, a question that helps you get most of the way there that's much easier for you to answer.
00:17:51.200 --> 00:18:02.319
So, for instance, if you get a risk message from an organization that you trust, instead of saying, should I evacuate from this disaster, which is a really hard question to answer, you can ask an easier question.
00:18:02.559 --> 00:18:05.920
Do I trust the organization that's telling me that I should evacuate?
00:18:06.160 --> 00:18:11.440
If the answer to that is yes and they're telling me to evacuate, then none of that other stuff about the storm matters.
00:18:11.599 --> 00:18:13.839
I trust that they're going to try and keep me safe.
00:18:14.000 --> 00:18:15.759
That question's really easy for me to answer.
00:18:15.839 --> 00:18:18.559
And so I can make a decision on the basis of that.
00:18:19.119 --> 00:18:26.319
So that's a positive case, I think, when people rely on a heuristic and it helps them make a decision that's good enough.
00:18:26.559 --> 00:18:37.440
Um when those biases can kind of uh frequently lead us astray, or when those heuristics can frequently lead us astray, we we call it a bias.
00:18:37.680 --> 00:18:43.920
And so there's a bazillion different biases that come up in risk research, decision research.
00:18:44.160 --> 00:18:51.039
It's not worth unpacking them all of them here, and I couldn't possibly if I tried, because I don't know a lot about all of them, but there's a ton.
00:18:51.279 --> 00:18:58.720
Some that I think are really, really interesting to kind of focus on in this space are um what's known as confirmation bias.
00:18:58.880 --> 00:19:16.240
It shows up a lot in kind of popular literature as well as other places, but confirmation bias is essentially an over-reliance on the heuristic that if information that I see out in the world is similar to things that I already believe, I already know to be true, then it's likely to be true.
00:19:16.400 --> 00:19:21.359
Um and if it is dissimilar from things that I already believe, then it's likely to be false.
00:19:21.759 --> 00:19:25.759
A lot of the time, you can imagine how this would serve you pretty well.
00:19:25.839 --> 00:19:28.559
Um, it basically gives you a nose for BS, right?
00:19:28.640 --> 00:19:36.720
It's like, I know a lot of things about the world, and you got this piece of information, and that doesn't conform with my understanding of how the world works.
00:19:36.799 --> 00:19:38.720
And so like, I suspect that's not true.
00:19:38.960 --> 00:19:40.640
Serves us well in a lot of cases.
00:19:40.799 --> 00:19:52.799
But when you're getting information about a hazard that's asking you to do something that doesn't conform with the way that you normally think about stuff, or the last four hurricane messages that you got, it turned out not to be a big deal.
00:19:52.960 --> 00:20:04.240
You think that they're overblown, it may not kind of allow you to make the best decision if you rely on the conformity of that new information to stuff you already believe.
00:20:04.480 --> 00:20:10.799
So that's a bias that can really affect how your risk perception develops as a result of the information that you're getting.
00:20:10.960 --> 00:20:13.599
Does it conform with my worldview or not?
00:20:14.160 --> 00:20:38.240
Um, another one that we see a lot and is the basis for a lot of the early research on building an emotional sense into risk perception is a bias that results from over-reliance on what's called the affect heuristic, um, which is a jargony term that basically is associated with the idea that if you uh if something feels good, it probably is good and not bad.
00:20:38.400 --> 00:20:42.079
And if it feels bad, it's probably bad and not good.
00:20:42.559 --> 00:20:46.960
Um, and so if you can imagine a lot of cases, this is really, really good.
00:20:47.039 --> 00:20:58.240
Um, things that you have negative experiences with in the past, things that are taboo in your society, or other things that feel bad, uh, you probably don't want to think that those things are actually really good for your society to do.
00:20:58.319 --> 00:21:02.000
Um, you might think that they're more risky, you might think they're less beneficial.
00:21:02.160 --> 00:21:10.480
Um, but it can run into some serious trouble in the space that we're in if it creates a bias against kinds of things that we're asking people to do.
00:21:10.799 --> 00:21:21.920
So you can imagine um it could negatively impact, say, risk perceptions of a hazard that you're talking about in a community where that hazard is generated by some kind of important activity.
00:21:22.079 --> 00:21:33.359
If you're thinking about greenhouse gas emissions, uh talking about those as a hazard in a coal mining community, you're asking people to take something that is really important to their livelihoods and see it as a bad thing.
00:21:33.519 --> 00:21:35.759
And that can be really, really hard for people to do.
00:21:35.839 --> 00:21:38.160
Um, you're asking them to kind of go against that bias.
00:21:38.319 --> 00:21:48.240
So if you think that they have a legitimate risk that's associated with that, that can be a barrier to you being able to talk about that risk in a way that will help elevate their perceptions.
00:21:48.559 --> 00:21:59.759
Uh similarly, if you're advocating for behaviors that have negative competitions in the community that you're talking to, you're saying you should do this, and they think about that, and they're like, ooh, that's icky, they're gonna think that that is.
00:22:00.240 --> 00:22:04.880
Likely riskier and less beneficial than you're trying to convey to them as a risk expert.
00:22:05.359 --> 00:22:10.240
And for those on the audio side of this, I'm I'm doing bunny ears around risk experts.
00:22:10.480 --> 00:22:12.640
Okay, no, you're doing quotes all the time.
00:22:12.799 --> 00:22:13.279
Yeah.
00:22:13.519 --> 00:22:14.480
I love it.
00:22:15.279 --> 00:22:20.160
Yeah, it occurs to me that this is a this is an audio medium, so I should probably make that clear.
00:22:21.680 --> 00:22:26.319
But yeah, so those are just a couple of examples of biases that can come about from those heuristics.
00:22:26.559 --> 00:22:30.160
Whether or not we can change them is really hard.
00:22:31.279 --> 00:22:36.480
Generally speaking, when we're writing messages, our touch points with our publics can vary.
00:22:36.559 --> 00:22:39.039
You can have a lot, or you can have not that many.
00:22:39.359 --> 00:22:42.960
So you don't necessarily have many opportunities to change people's biases.
00:22:43.200 --> 00:22:47.839
And even if you did, it's unclear whether those would be particularly effective.
00:22:48.000 --> 00:22:57.680
It's hard to shift people's beliefs over time, particularly when they're based on using these decision rules that people use below the level of consciousness.
00:22:57.759 --> 00:22:59.680
So they use them without thinking about it.
00:22:59.920 --> 00:23:05.599
This is just an easier question to ask myself that I can actually answer as opposed to one that's really hard.
00:23:05.920 --> 00:23:23.359
So usually when I think about shifting biases, it's more about trying to keep in mind what the biases of the group that you're going to be speaking to are and trying to communicate in a way that can help you, if not bypass them, then account for them.
00:23:24.000 --> 00:23:39.359
So for instance, if you're advocating for a hazard in a community that thinks of it as beneficial, or in the other case, if you're advocating for behavior that response behavior that people have negative connotations associated with, using trusted messagers.
00:23:39.519 --> 00:23:51.680
So people who are in that community, who are part of that community, who can help them uh contextualize what you're asking them to do or what you're asking them to see as a threat in a way that is less likely to be discounted.
00:23:51.759 --> 00:23:53.599
It's from a member of that interior community.
00:23:53.759 --> 00:24:01.359
You can kind of fight fire with fire, use a trust-based heuristic to help you fight against that instinctual affect heuristic.
00:24:01.440 --> 00:24:05.039
And they might say, like, oh, Bob's telling me I really need to worry about this.
00:24:05.200 --> 00:24:16.960
Like I I thought it was all BS, but now I'm I'm thinking like this person who's important, who's kind of um who I trust has my best interests at heart, is asking me to to focus on this.
00:24:17.039 --> 00:24:21.039
And you might be able to get past those biases in that kind of way.
00:24:22.240 --> 00:24:29.519
I th I really appreciate that because so what you're saying, we may not be able to directly impact the bias that exists, right?
00:24:30.000 --> 00:24:50.319
But just knowing that they exist, knowing your community and what some of those are should help you again structure either your message, your dissemination method, maybe the trusted sources that you're gonna use to act as a secondary um sender of information or to to push or help influence that.
00:24:50.640 --> 00:24:59.200
Um and again, just I think not admitting that they are out there or being ignorant of the fact, I think just is just bad.
00:24:59.359 --> 00:25:01.599
Like you need to have that understanding.
00:25:01.839 --> 00:25:06.960
And and I I appreciate the understanding of the difference between those.
00:25:07.200 --> 00:25:22.960
Um and like I said, as you guys talk and as you guys nerd out, what I am thinking about is the way I'm writing messages and maybe the way that I need to do it a little differently based on a hazard for a specific population.
00:25:23.359 --> 00:25:25.519
And we don't talk about that enough sometimes.
00:25:25.599 --> 00:25:39.920
I think we we send out a message to this group when in reality we have population subsets within that greater population that may have a bias, a very strong bias one way.
00:25:40.319 --> 00:25:45.119
And if I send out this message to everybody, it is not a one size fits all.
00:25:45.279 --> 00:26:03.440
And I may need to do a secondary reach out or a secondary effort to specifically target that that potential barrier that may, you know, impede someone from taking that that action when I have other groups of my population that are immediately gonna be like, okay, it's a trusted source, all these things.
00:26:03.599 --> 00:26:09.119
Where this other one is like, yeah, you know, it was sent by Red Rock Police Department.
00:26:09.200 --> 00:26:10.000
I don't trust them.
00:26:10.079 --> 00:26:12.400
They're, you know, a bunch of whatever.
00:26:13.200 --> 00:26:16.079
So we may need that additional source to break through that.
00:26:16.240 --> 00:26:20.400
So I I I honor I'm very grateful that you went through that.
00:26:20.880 --> 00:26:29.440
But I'm I'm really glad you bring that up as well, just um in terms of of the bias that we have, because because the question this brings up for me is how do you know?
00:26:29.680 --> 00:26:32.559
Kind of how do you know what biases are active in your community?
00:26:33.200 --> 00:26:37.359
How do you subset your community in a way that helps you get around those different biases?
00:26:37.839 --> 00:26:45.599
And I think turning that lens around and thinking like researchers are biased too, practitioners are biased too, right?
00:26:45.759 --> 00:26:52.160
Like we have beliefs about the populations that we serve, the populations that we work with that we don't always know about.
00:26:52.240 --> 00:27:01.920
Um, and it's really important, I think, to while sometimes these biases are gonna be super obvious, um, and that anecdotada will suffice, right?
00:27:02.079 --> 00:27:07.839
That sense from your friends, from other community members who tell you a thing once, you're like, oh, okay, I know something about this community now.
00:27:08.000 --> 00:27:24.640
But there's really no substitute if you're going to be relying on this to help keep people safe, there's no substitute for really trying to get into those communities, get to know them really well, and collect more than just a handful of anecdotes about that before you make a decision on the basis of the biases that they're likely to have.
00:27:24.799 --> 00:27:29.680
Um, I think it can be really easy to kind of make that judgment a blanket judgment.
00:27:29.759 --> 00:27:37.200
Um, and we we definitely want to make sure that when we're making a call about going around a bias, um, we're making the right call.
00:27:37.440 --> 00:27:37.759
Dang it.
00:27:37.839 --> 00:27:42.160
And I know, Jeanette, I'm sorry for saying one more thing and then we'll get to your question.
00:27:42.319 --> 00:27:47.920
It goes to that whole Craig Fugate Waffle House comment because that is one that is very real.
00:27:48.079 --> 00:27:50.559
And Jeanette reminded me I have to get them on the show.
00:27:50.720 --> 00:27:51.359
I do.
00:27:51.759 --> 00:28:04.720
When you send a message out here and people look at the phone, they look at, and I'm not kidding, they'll compare it to Waffle House and be like, well, they're still selling waffles and pancakes, or oh, the level is only pancakes now.
00:28:04.880 --> 00:28:11.440
Like, whatever it is, it it is a real thing that people just naturally go to that you have to fight.
00:28:11.519 --> 00:28:12.240
Okay, I'm done.
00:28:12.400 --> 00:28:14.799
We'll go to your go to the paper.
00:28:17.039 --> 00:28:21.680
I you you started to talk about message contents where you're talking about like trusted sources and such.
00:28:21.839 --> 00:28:28.319
And so I think this is a nice shift to to talk about this study that that I'm I'm pretty excited about.
00:28:28.400 --> 00:28:37.920
This is the last study that we got to do with a little bit of money that was left over from the work that we had done with the FEMA project.
00:28:38.079 --> 00:28:41.599
I think we were like, oh, we have like a couple thousand dollars left.
00:28:41.759 --> 00:28:48.559
We could do this one last experiment out in the community, and we have no money left to analyze it, but we're gonna collect the data.
00:28:49.519 --> 00:28:52.720
This is what happens when you're a researcher and you have too many questions.
00:28:52.799 --> 00:28:58.319
You're like, where can I find a couple thousand dollars in somebody who just like really wants to do this work?
00:28:58.559 --> 00:29:02.400
And uh, when I have great research partners, we figure out a way to do it.
00:29:02.559 --> 00:29:02.960
Yeah.
00:29:03.279 --> 00:29:24.000
So this paper um we called it content mapping, which is not the same as message mapping, which I know some PIOs know this idea of message mapping, which is uh I think it came out of Cavello's research, and it's it's developing a framework for like the three ideas that you're gonna communicate in a certain amount of time.
00:29:24.160 --> 00:29:25.599
That's that's message mapping.
00:29:25.680 --> 00:29:26.480
That's not the same.
00:29:26.640 --> 00:29:43.039
What we did was content mapping, because we took the contents of a warning message and we identified what contents matter the most in terms of the outcomes that are within the warning response model.
00:29:43.440 --> 00:30:00.799
Um, and if you leave a content out of the message, what does it do to people's understanding of the message, their belief about the message, their personalization of the message, their ability to decide, and if they need to look for more information to be able to take action.
00:30:01.119 --> 00:30:03.680
So I thought this was a fantastic study.
00:30:03.759 --> 00:30:04.319
I loved it.
00:30:04.480 --> 00:30:11.200
So, Hugh, you could you describe for us some of the methods that we use to design the study?
00:30:11.519 --> 00:30:15.599
I know it was like a year ago, but it takes that long for things to get published.
00:30:15.920 --> 00:30:16.319
It does.
00:30:16.640 --> 00:30:22.000
It was an experiment, and it was in San Diego, and I think uh wildfire and chemical spill.
00:30:22.240 --> 00:30:24.319
So that should kick us off.
00:30:24.799 --> 00:30:32.559
Yeah, so um again, uh for those of you coming back in from the the first part of this, uh, an experiment is my favorite thing in the world.
00:30:32.640 --> 00:30:36.960
So special thanks to Jeanette for letting me play in my experimental playgrounds.
00:30:37.039 --> 00:30:43.119
Um it's a really important study that I think we we as a group really, really needed to do.
00:30:43.279 --> 00:30:47.359
Um and hopefully I can kind of walk you through and convince you why.
00:30:47.519 --> 00:31:07.200
Um so what we've been kind of working on for uh the last few years, um building on decades of research on the warning response model, is thinking about um the contents of messages that people uh need for when they receive that message to feel motivated to take some kind of action to protect themselves.
00:31:07.359 --> 00:31:18.799
Um and so according to the model built on decades of research, there's five content areas that are really important for um kind of how people get motivated to do that.
00:31:18.880 --> 00:31:20.880
Uh so the first thing you need is uh a source.
00:31:21.039 --> 00:31:23.920
So your hazard needs uh a source ideally that is trusted.
00:31:24.079 --> 00:31:29.039
People need to know who the message is coming from before they'll kind of believe it's gonna happen.
00:31:29.519 --> 00:31:30.960
Um you need a hazard.
00:31:31.119 --> 00:31:34.400
So we were talking uh earlier on about uh wildfire jargon.
00:31:34.480 --> 00:31:39.920
A lot of the time um when you provide information in the form of just jargon, you're not including the name of the hazard.
00:31:40.000 --> 00:31:43.680
And knowing the nature of the hazard can really help contextualize the decisions that you make.
00:31:43.839 --> 00:31:45.200
So it needs to be in there.
00:31:45.359 --> 00:31:52.720
And an extension that we've made over time is thinking about well, I can tell you what the hazard is, but what if people don't really know what the hazard's gonna do?
00:31:52.880 --> 00:31:57.279
Uh and so we've been extending that by talking about the impacts that are associated with that hazard as well.
00:31:57.359 --> 00:32:00.880
And that's something that a lot of our weather colleagues have been doing for quite some time.
00:32:01.039 --> 00:32:07.920
So thinking about that and contextualizing how important that is is gonna be a big part of what this study is contributing.
00:32:08.160 --> 00:32:10.799
So you've got a source, you've got a hazard.
00:32:10.880 --> 00:32:12.880
Uh, you also need information about the location.
00:32:13.039 --> 00:32:14.799
So where that hazard is going to affect.
00:32:14.960 --> 00:32:20.640
This can really help people make a judgment about whether or not they're going to be, whether or not they need to pay attention to that message.
00:32:20.880 --> 00:32:26.160
Um, there are kind of limitations on our ability to hyper-target a message to a specific area.
00:32:26.319 --> 00:32:32.319
So being able to, in the message itself, delineate the area where that hazard's going to be particularly threatening is really useful.
00:32:32.559 --> 00:32:35.759
Also providing information about where people can respond to.
00:32:35.839 --> 00:32:38.079
So if we want them to evacuate, where do we want them to go?
00:32:38.240 --> 00:32:39.680
Where do we want them to not go?
00:32:39.839 --> 00:32:43.359
Um, it gives them an opportunity to think about that as well.
00:32:43.680 --> 00:32:45.519
Um, you need an element of time.
00:32:45.680 --> 00:32:50.559
So as important as where a hazard is going to strike, when it's going to strike is really important too.
00:32:50.720 --> 00:32:58.640
And if you see a hazard coming and you have time to prepare, being able to identify a time by when people should have done something is also really important too.
00:32:58.720 --> 00:33:08.240
So that sense of when a hazard is going to affect a group of people, uh, but also when they should be able to take action to make themselves the most safe.
00:33:08.480 --> 00:33:11.359
And finally, the other really important element is guidance.
00:33:11.519 --> 00:33:15.279
You need to provide information about what people should do to keep themselves safe.
00:33:15.440 --> 00:33:20.720
Now, we've talked a lot about people know in a lot of cases things that they could do.
00:33:20.880 --> 00:33:25.119
They may know better than we do in a particular circumstance what they can do to keep themselves safe.
00:33:25.359 --> 00:33:41.759
But that guidance information is really critical for those people who don't know or who just need a guidance in the moment of like, oh wow, I really need to take some kind of action, help them get past that milling that Jeanette was talking about earlier, give them a sense of this is my starting place, this is what I should do, and then adjust from there.
00:33:42.880 --> 00:33:56.240
Um so those content areas, uh, and I'm gonna stretch it from five into six because it builds into the way that we we talked about our message design, we differentiated between the name of the hazard itself and uh those impacts of that hazard.
00:33:56.319 --> 00:33:59.039
So including those as two separate pieces of information.
00:33:59.440 --> 00:34:06.799
So what we did is that we built uh seven versions of the same message for each of two hazards.
00:34:07.119 --> 00:34:12.320
And the first message that we built for each one was um a complete message.
00:34:12.480 --> 00:34:15.920
So it was one that included all of those content areas in it.
00:34:16.079 --> 00:34:25.920
It's the kind of best job that we could do in 360 characters of writing a message that includes all of the information that the warning response model would suggest that you have.
00:34:26.400 --> 00:34:30.000
So we started that, and that was our first treatment.
00:34:30.159 --> 00:34:38.800
And so thinking back to our discussion earlier on about experiments, a treatment is in this context just a different message, different set of information that you're giving people.
00:34:38.960 --> 00:34:57.039
And when you make comparisons across different treatments, when you do an experiment and you randomly assign people to those different conditions, um, then you can say with a reasonable degree of certainty that any differences that you see between people who are in those groups is a result of the different treatment that they got.
00:34:57.119 --> 00:34:59.760
So in this case, the different type of message that they got.
00:35:00.000 --> 00:35:02.559
So that first treatment was a complete message.
00:35:02.800 --> 00:35:09.920
They we also had six other versions of that message that eliminated one of those content areas one by one.
00:35:10.000 --> 00:35:14.800
So we'd get a message that was identical to the complete message, but didn't include a source.
00:35:15.039 --> 00:35:26.960
There was also an identical message that didn't include the name of the hazard, one that didn't include the hazard impacts, one that didn't include location, one that didn't include time, and one that didn't include guidance.
00:35:27.199 --> 00:35:43.039
And what that allowed us to do was to look across all of these different motivational outcomes and see when we compare them to the complete message, which messages did worse than that complete message at motivating people.
00:35:43.280 --> 00:35:54.639
We know from some research that we've done in the last couple of years, we have some pretty good evidence that including all of those versus omitting chunks of them at a time makes you more motivated to take some kind of action.
00:35:54.800 --> 00:36:02.960
But what this experiment let us do that was so exciting was to drill down into specifically what each of the contents was doing.
00:36:03.119 --> 00:36:07.760
How do each of those specific contents motivate people to take action?
00:36:08.239 --> 00:36:20.800
So in the warning response model, we've talked about this uh already, but just quick reminder generally speaking, what we're focused on is we want people to understand what the message is about and what it's asking them to do.
00:36:20.960 --> 00:36:24.559
We want them to believe the message, so believe that something is actually happening.
00:36:24.800 --> 00:36:27.039
We want them to personalize the risk from the hazard.
00:36:27.199 --> 00:36:30.960
We want them to believe that it's relevant to them, that they need to do something about it.
00:36:31.199 --> 00:36:35.280
And we want them to be able to decide based on the information in the message.
00:36:35.440 --> 00:36:41.599
If we give them all of this information and they look at it and they're like, I don't know what I'm supposed to do with this, then the message isn't as motivating.
00:36:41.920 --> 00:36:50.239
We also measured a couple of other constructs that I think are really useful to think about in terms of this, but aren't explicitly included in the warning response model.
00:36:50.480 --> 00:36:53.199
And uh the first ones are what we were talking about earlier.
00:36:53.280 --> 00:36:55.280
We were talking about self and response efficacy.
00:36:55.360 --> 00:37:05.360
So thinking about how capable do you feel of taking the action that's recommended in that message, then that response efficacy, how effective do you think that will be at keeping you safe?
00:37:05.599 --> 00:37:10.719
We also measured a couple of outcomes associated with trust in the organization sending the message.
00:37:10.960 --> 00:37:22.079
So we know that trust is important for how people uh kind of take in a message, and it can be really important to trust the organization that is the source of your message in order to help you kind of take that action.
00:37:22.320 --> 00:37:30.000
But we also think that getting a message that is in some way suboptimal can also have a deleterious effect.
00:37:30.079 --> 00:37:33.760
Um, it can be bad for trusting the organization in the future.
00:37:34.159 --> 00:37:36.239
I was about to say that was come on now.
00:37:36.480 --> 00:37:36.559
Sorry.
00:37:37.199 --> 00:37:38.480
Deleterious.
00:37:39.280 --> 00:37:39.679
Yeah.
00:37:39.840 --> 00:37:40.079
Yeah.
00:37:40.559 --> 00:37:41.440
Deleterious.
00:37:41.519 --> 00:37:41.679
Okay.
00:37:41.920 --> 00:37:43.360
I will look that one up later, folks.
00:37:43.519 --> 00:37:44.079
You're not alone.
00:37:44.239 --> 00:37:45.039
Delete.
00:37:45.440 --> 00:37:48.960
It could lead to it could lead to lack of trust or erosion of trust in the future, right?
00:37:49.039 --> 00:37:54.480
So if I get a bunch of bad messages from an organization, I might not trust them to give me good information in the future.
00:37:54.719 --> 00:37:59.679
So not only are we thinking about the effect for this hazard in particular, we're thinking about the next one.
00:37:59.760 --> 00:38:01.840
Um, what am I gonna have to do for that?
00:38:02.239 --> 00:38:07.519
Uh so those were the things that we were looking at across those different outcomes.
00:38:07.760 --> 00:38:20.159
And uh what we wanted to do is compare how each removal of each specific concept, content area, um, from those messages was gonna impact each of those outcomes.
00:38:20.400 --> 00:38:33.679
And by and large, the kind of fascinating stuff that we found is that when you emit certain kinds of information, it has a more dramatic effect across a wide range of those different motivational outcomes than other kinds.
00:38:33.920 --> 00:38:44.719
And in particular, what we found is that uh including hazard information, so information about the name of the hazard, the impacts of the hazard, and guidance were the most impactful pieces of information.
00:38:44.960 --> 00:38:51.360
I want to be very clear that we found that the most the best way that you can do this is to include all the information.
00:38:51.599 --> 00:39:04.079
Um, all of those things, the complete message outperformed every other message every single time on everything, to at least some extent, uh, largely significantly uh in a statistical sense.
00:39:04.320 --> 00:39:12.639
So the best message is still the complete message, but there are pieces of information that are really worth emphasizing in the messages that we put together.
00:39:12.800 --> 00:39:23.199
And that's if you if you have to omit something, make sure that it's not the name of the hazard, make sure it's not the impacts of that hazard, and not the guidance information.
00:39:23.440 --> 00:39:30.800
Those things really drove people's efficacy, understanding, belief, and a lot of those trust variables.
00:39:30.960 --> 00:39:45.360
So thinking about the future, when you want people to get your next message and believe it, believe that you're credible and trustworthy, then focusing on having those complete messages is really this this may seem like it's not a big deal to a lot of people who are listening.
00:39:45.840 --> 00:39:52.000
But before, when we measured hazard, we measured hazard and impact as one thing.
00:39:52.320 --> 00:40:11.519
And the so the academic process of getting something published, we write up a paper, we send it out to a journal for peer review, they send it out to people who are um who are familiar with the area of research, the methods, the theory, um, and can evaluate our study and give us feedback.
00:40:12.400 --> 00:40:14.639
And so we sent it out to the journal.
00:40:14.880 --> 00:40:21.840
They sent back the reviews to us, and we responded to the reviews by making some changes, because there are a few.
00:40:22.400 --> 00:40:24.480
But we got a second round of reviews.
00:40:24.960 --> 00:40:34.239
And the second round said, you need to emphasize the importance of this discovery that hazard and hazard impact are two different things.
00:40:35.840 --> 00:40:45.920
And I was like, I don't know how important that is, but clearly these are communication scholars who may or may not be in the warning space.
00:40:46.000 --> 00:40:46.880
They're always anonymous.
00:40:46.960 --> 00:40:47.920
We don't know who they are.
00:40:48.159 --> 00:40:51.840
But they were saying this is something that really needs to be emphasized.
00:40:52.159 --> 00:40:53.679
And it is very important.
00:40:53.840 --> 00:40:59.199
And it it it either extends the warning response model or it expands it.
00:40:59.360 --> 00:41:11.119
It's something where the way we were talking about these contents before, we now know a lot more about the importance of one piece that we hadn't previously emphasized.
00:41:11.440 --> 00:41:17.119
And I Hugh was talking about within the weather space, they've talked about impacts a lot in the past.
00:41:17.199 --> 00:41:26.559
And they they did a lot work on what they call impact-based warnings, which they truly are influential in people's understanding of the hazard.
00:41:26.800 --> 00:41:33.039
But impact-based warnings don't necessarily motivate people to take action because they're not including guidance.
00:41:33.599 --> 00:41:50.239
So a complete message has the impacts and the guidance and that partnering together, it's kind of like you increase people's fears about the impacts, you got to give them the guidance so they can be self-efficacious and take action.
00:41:50.559 --> 00:41:53.920
So I mean, this was a really big deal.
00:41:54.079 --> 00:41:59.840
And so I mean, theoretically, research-wise, like this this was super important.
00:42:00.000 --> 00:42:08.639
Um the things that didn't matter, like Hugh, you want to mention the, I mean, like the source and the time and the location, like what was that about?
00:42:09.119 --> 00:42:13.920
Yeah, so we did see smaller effects from source and time information.
00:42:14.079 --> 00:42:19.039
And again, I want to be super clear that that I don't think in this case means that those things aren't important.
00:42:19.199 --> 00:42:19.519
Right.
00:42:19.679 --> 00:42:31.440
Um, Jeanette made an allusion earlier on to when we do an experiment, we're talking to people in a space where they're very comfortable, uh, they're making different kinds of judgments than they always make.
00:42:31.519 --> 00:42:35.599
This is a weakness of all experimental research, is that the context is just a little bit different.
00:42:35.760 --> 00:42:39.119
So you don't always pick up on things that people will encounter in the moment.
00:42:39.360 --> 00:42:47.199
Um, so for instance, one of the things that we would hypothesize about a source is that if you don't have the source, people are probably less likely to trust the message.
00:42:47.440 --> 00:42:50.800
Um we're pretty confident that this is true in the field.
00:42:50.960 --> 00:43:02.960
We've seen some examples of this in other places, but we didn't see a strong effect of that in this circumstance where emitting source didn't have as big an impact as we thought on things other than trust.
00:43:03.119 --> 00:43:07.360
Um, so trust in the organization in the future, it is kind of like, well, it's an anonymous organization.
00:43:07.440 --> 00:43:08.320
What do I know?
00:43:08.960 --> 00:43:23.199
But I think in the the case where you're in a situation where you have to deal with this, um, in particular, source and time are things that you kind of make inferences about when you're taking a study online.
00:43:23.599 --> 00:43:28.800
If I don't tell you time information and I tell you this is the message, you're probably assuming that this is happening now.
00:43:28.960 --> 00:43:33.199
But the now doesn't feel as urgent when you're doing a message like this.
00:43:33.440 --> 00:43:48.559
Similarly, if you get something from a source, you're not expecting in an online survey environment where you're doing something for uh research purposes, you're not necessarily expecting to see the name of a real organization there that you know and trust and that you know members of.
00:43:48.719 --> 00:43:54.960
And so those kinds of things we didn't see uh the effects that we would anticipate in a research sense.
00:43:55.199 --> 00:43:59.599
But I think there's pretty plausible reasons for why we may be underestimating the impacts that.
00:44:00.159 --> 00:44:02.000
Those have in the real world.
00:44:02.159 --> 00:44:06.239
So I really don't want people to take away from that like source and time don't matter.
00:44:06.400 --> 00:44:07.440
I don't need to include those.
00:44:07.599 --> 00:44:09.920
We've got strong reason to believe that they do matter.
00:44:10.320 --> 00:44:13.599
It's just really difficult to pick those up in the experimental context.
00:44:13.840 --> 00:44:19.280
And you got me, I have to admit, experiments aren't good for absolutely everything.
00:44:19.360 --> 00:44:28.400
There's no substitute for doing some of that field research, going back to that qualitative space, getting back into just talking to people and having them tell you in words what matters.
00:44:28.719 --> 00:44:34.079
And that's where we get a lot of the information that we that makes us confident that the source is important.
00:44:34.400 --> 00:44:34.800
Yeah.
00:44:35.039 --> 00:44:47.039
I if I have, well, actually, I do, I have a little bit of funding that came back to me to my university position due to a grant that we're currently working on that's funded by USGS.
00:44:47.360 --> 00:44:58.639
And with that funding, I would love to design another little mini experiment where we learn more about the importance of source and also where source really belongs in the message.
00:44:58.800 --> 00:45:04.000
Um, because our early evidence said source should come first, and this is why.
00:45:04.239 --> 00:45:06.480
And I would love to replicate that.
00:45:06.719 --> 00:45:16.320
And I mean even if we did it with a student population, did an experiment or some sort of like card sorting thing where we had people move contents around.
00:45:16.559 --> 00:45:26.960
Um students, people who are between 18 and 25 are getting a hundred, two hundred alerts on their phones every day.
00:45:27.920 --> 00:45:37.679
That is a great population to talk to about how important is it to know who's giving who's sending you that message when you're getting so many things that you're just constantly swiping away.
00:45:37.920 --> 00:45:44.079
Um, so that's my next little mini experiment, and um hopefully we'll be able to do that.
00:45:44.480 --> 00:45:45.440
I can't wait.
00:45:45.679 --> 00:45:46.320
I know.
00:45:49.599 --> 00:45:52.400
I'm on I'm just like taking all this in here.
00:45:52.719 --> 00:45:58.480
Um and it's stuff that Jeanette and I, I mean, we've we've trained this for a while, and I've known this is coming.
00:45:58.719 --> 00:46:03.840
Um I I have a couple questions that a, how do people get it?
00:46:04.159 --> 00:46:05.119
Number one.
00:46:05.519 --> 00:46:16.400
Um, and again, because you talk about it's published, and for those not in the academic space, I I you talked about it a little bit, but like, does that mean it's available to everybody?
00:46:16.559 --> 00:46:20.719
Because if you're out there and you're like, I'm published too, self-publishing does not count.
00:46:20.960 --> 00:46:21.280
Okay?
00:46:21.920 --> 00:46:25.280
Putting something on Wikipedia does not mean you're a published author.
00:46:25.440 --> 00:46:27.920
Um, means you have too much time on your hands.
00:46:28.320 --> 00:46:38.480
But it's but for for this, like when they say it's it's been published in a journal, that that is very significant.
00:46:39.119 --> 00:46:43.920
And so then it gets to the question of uh how how can we apply it, right?
00:46:44.079 --> 00:46:46.159
Bringing it back to pract practitioners.
00:46:46.480 --> 00:46:52.320
I appreciate you saying that the complete message is is what we need to do.
00:46:52.559 --> 00:47:01.519
Um, but this is the first time we've been able to hear actual evidence that talks about why each of these things is important.
00:47:02.000 --> 00:47:10.800
And if you omit any one of these things, you are decreasing right the effectiveness of your message in in one form or another.
00:47:11.039 --> 00:47:19.760
And I I gotta say, one of the things that's probably most frustrating to me when we're out there training with people is when we go over portions of this.
00:47:19.840 --> 00:47:21.760
And granted, this is new.
00:47:21.920 --> 00:47:24.960
Like if you're hearing this, like this is brand new.
00:47:25.199 --> 00:47:26.559
So welcome.
00:47:26.800 --> 00:47:30.480
Um, so there's a little grace, I guess, for everybody else, but now we know.
00:47:30.639 --> 00:47:42.880
So if we're doing a training and we're talking about each of these components, and you omit one, you're just like, well, I'm not gonna do that one because I it just because we don't, because my current template says we don't.
00:47:43.920 --> 00:47:48.400
It doesn't make sense anymore how you can stand on those thoughts or principles, right?
00:47:48.639 --> 00:48:02.239
Because Hugh, from what you're saying from the experiments, from the research, from the data, is removing any one of those things, it I I keep going to the basic terms, right?
00:48:02.400 --> 00:48:04.400
Doesn't make it as effective as it could be.
00:48:04.960 --> 00:48:06.639
Am I oversimplifying?
00:48:07.280 --> 00:48:11.039
Well, I think there's I think that's fair to say.
00:48:11.199 --> 00:48:18.079
I think there are uh there is a metric for everything that if you didn't include that in the message, it makes it worse in some way.
00:48:18.239 --> 00:48:23.039
How important each of those are in order to achieve your goals as a message writer may vary.
00:48:23.119 --> 00:48:26.480
So your mileage may vary on exactly how important each of these things are.
00:48:26.639 --> 00:48:40.960
But what we do know is that out of those contents, among the most important for all of these different um kind of motivations that we know from a lot of research that we've done are likely to lead to people trying to take some kind of action to keep themselves safe.
00:48:41.119 --> 00:48:45.199
That hazard name, the hazard impact, and the guidance are the most important things.
00:48:45.440 --> 00:48:52.639
Um where this gets really important is that we see a lot of messages that don't include any information about impacts.
00:48:52.800 --> 00:49:02.079
And in fact, up until now, we haven't had as much evidence that as we do currently to advocate really strongly for including that impact information.
00:49:02.239 --> 00:49:16.079
But what we've got suggests that people are less likely to understand what they're gonna be facing, they're less likely to personalize that risk to themselves, and they're not gonna be able to make a decision with confidence on the basis of that when they don't have information about those impacts.
00:49:16.159 --> 00:49:33.519
So if they don't know how it's gonna affect them, it's gonna make it really hard for them to make that decision, which makes them more vulnerable to things like biases that we were talking about earlier, or to milling or information seeking, where they're going to spend precious time trying to gather the information that they need to make that decision easier for themselves.
00:49:33.840 --> 00:49:52.159
And I think that's the really strong takeaway from this is not just the extension that we need to make for theory in um kind of the research space, but it also has, I think, strong implications for how we think about uh prioritizing information in the messages that we put out.
00:49:52.719 --> 00:49:55.119
I want to be clear that it's one experiment.
00:49:55.440 --> 00:49:59.119
Um, you're never gonna be 100% right in every experiment.
00:49:59.360 --> 00:50:13.840
But I think what we found here is really indicative of something that makes a lot of sense uh in the context of other research that we've done, research that other colleagues have done, and kind of the logic of um orienting messages towards helping people make themselves safe.
00:50:14.000 --> 00:50:30.559
And when all of those kind of come together and you get some evidence like this, it's it's I think prudent to start to put that into practice while we gather more information about how these dynamics um kind of change under different circumstances or with different populations in the future.
00:50:31.039 --> 00:50:33.440
And I know there's only been one experiment, right?
00:50:33.599 --> 00:50:34.639
But I think that's the point.
00:50:34.800 --> 00:50:38.480
There's only been one so far about this that's identified this.
00:50:38.559 --> 00:50:43.360
And I can't find any information that's contrary to what this is showing.
00:50:43.679 --> 00:50:56.880
And and this is one of those, like, if I could reach through this, whatever medium you're if you're listening or if you're watching this on YouTube and just shake you and say, why do we struggle so much with this concept?
00:50:57.199 --> 00:51:03.679
Because we've been out there in training and we're we're saying, hey, because and I'm gonna focus not on on naming the hazard.
00:51:03.920 --> 00:51:05.199
Okay, we're good, right?
00:51:05.360 --> 00:51:11.280
And I I say that, I'm gonna use the air quotes now, we're good, because some of us aren't, but overall we're getting better.
00:51:11.519 --> 00:51:30.000
But Jeanette, we've we've seen resistance to putting in impact statements or sometimes specific guidance because people don't want to um cause trauma or they they don't want to cause you know pseudo effects for for someone, and we'll take a hazardous materials, right?
00:51:30.159 --> 00:51:32.000
One that that we've done in training.
00:51:32.239 --> 00:51:40.320
They're like, well, I can't say any of those things that could happen to somebody because then they're gonna start calling 911 and they're gonna say, now I'm feeling all those things.
00:51:40.800 --> 00:51:42.159
It's interesting.
00:51:42.320 --> 00:51:47.760
I I don't necessarily want to call it pushback, but um maybe growing pains.
00:51:48.320 --> 00:51:50.960
But now you have the research, right?
00:51:51.199 --> 00:51:56.079
That definitively in this in this study shows the importance of it.
00:51:56.480 --> 00:52:03.119
So, how can we help practitioners take that next step and just become okay with it?
00:52:03.920 --> 00:52:11.360
It's so interesting because it it kind of reminds me a little bit about um like the disaster myths that we talk about sometimes.
00:52:11.599 --> 00:52:15.119
Like that people are gonna panic if they get an alert.
00:52:15.280 --> 00:52:17.280
It's not this is not accurate.
00:52:17.440 --> 00:52:23.360
Um, that people are gonna act irrationally when they experience an environmental cue of a building falling down.
00:52:23.440 --> 00:52:25.599
It's not irrational to run away.
00:52:26.000 --> 00:52:30.000
Um people are gonna be um, you know, they're not gonna self-organize.
00:52:30.079 --> 00:52:33.920
They're just gonna be, you know, uh malicious towards their neighbors.
00:52:34.000 --> 00:52:35.039
That is not true either.
00:52:35.199 --> 00:52:37.840
People always converge to help one another.
00:52:38.159 --> 00:52:50.719
Um and so maybe this is another part of one of those things that we think about human beings that it's a concern that we have, but maybe it's not as valid as we think it might be.
00:52:51.280 --> 00:53:04.320
But I think that the bigger concern that people probably are facing when they're trying to think about including impacts is okay, I have 360 characters, how am I gonna include all that information?
00:53:04.639 --> 00:53:10.480
And now you're asking me to come up with that information, and I have not prepared for it.
00:53:10.960 --> 00:53:13.760
I have not prepared to include that kind of information.
00:53:13.840 --> 00:53:15.440
I haven't even thought about the impacts.
00:53:15.519 --> 00:53:17.519
All I've thought about is how do I name the hazard?
00:53:17.599 --> 00:53:22.880
And sometimes it's gonna be like a local incident, which is not clear at all.
00:53:23.440 --> 00:53:27.920
But this is the thing, you asked uh Hugh, like, how do we get this information?
00:53:28.159 --> 00:53:31.440
And we have the publications, we can make sure that people can get them.
00:53:31.599 --> 00:53:36.000
I always make a free version through the archives at the university so that people can have that.
00:53:36.239 --> 00:53:45.760
It's not typeset, so it's not pretty, like the the one that you would get from the journal, but it's the exact same contents, it's just not in the the beautiful version.
00:53:46.000 --> 00:53:54.800
But the other thing that people can get that gives them the impact statement for every hazard that we've been studying is the warning lexicon.
00:53:55.599 --> 00:54:06.400
We have already done all of the work to identify the most likely impacts from the individual hazards that people are likely to alert for.
00:54:06.880 --> 00:54:14.159
This is a document that is the foundation for the message design dashboard and is freely available.
00:54:14.320 --> 00:54:24.000
We paid for it to be publicly available through the Natural Hazards Review, which is a journal, a highly ranking journal, engineering social science.
00:54:24.159 --> 00:54:39.599
Um, but you could just look, you can go to my the Warn Room um website, you could go to my personal academic website, um, you could look it up through Google for um the warning lexicon, it click uh download, and it it's free.
00:54:39.679 --> 00:54:40.719
It will be downloaded.
00:54:40.800 --> 00:54:41.840
There's no paywall.
00:54:42.320 --> 00:54:53.679
The piece that you want is the supplemental material that provides to you 48 different hazards, and every hazard has an impact statement associated with it.
00:54:53.840 --> 00:54:57.360
And we also give you all of the guidance that goes along with the hazards.
00:54:57.440 --> 00:55:02.559
So it's basically giving you the six different pieces of information that you need for your message.
00:55:02.719 --> 00:55:15.760
So if you walk through the warning lexicon process to build a message, it's on you to determine that you're gonna leave something out because it's making you work through the process every single time.
00:55:16.000 --> 00:55:28.320
So you're gonna have a complete message because you're using the resources that have been researched, built, paid for, published, and freely available to everyone.
00:55:28.559 --> 00:55:32.000
And that is the one resource that I advocate for every time.
00:55:32.079 --> 00:55:34.960
And it's already got the impact statement built into it.
00:55:35.199 --> 00:55:40.800
So please, if you if you can't find it, send me a message and I will help you to get it.
00:55:40.960 --> 00:55:46.320
It has been downloaded almost 20,000 times around the world in the last three years.
00:55:46.800 --> 00:55:48.000
Fantastic resource.
00:55:48.159 --> 00:55:51.360
Please you know, use it, get it, make use of it.
00:55:53.440 --> 00:55:53.920
I don't know.
00:55:54.079 --> 00:55:55.599
Any other closing thoughts?
00:55:55.840 --> 00:56:03.199
Because I I think this has been honestly, again, from the practitioner spot, Hugh, you didn't nerd out too much.
00:56:03.360 --> 00:56:06.480
I mean, there's the word that started with the D, I wrote it down.
00:56:06.639 --> 00:56:08.239
Dela, is it deleterious?
00:56:08.639 --> 00:56:09.360
That's the one.
00:56:09.679 --> 00:56:10.480
Deleterious.
00:56:10.719 --> 00:56:13.360
I will look it up and I will try to use it at some point.
00:56:13.679 --> 00:56:24.880
Um but but in in all seriousness, I hope everybody was able to capture not only as he talked about risk perception, right?
00:56:25.039 --> 00:56:33.440
Which which I think is is valid, which I think plays into everything that happened within the study, but the now the why.
00:56:33.599 --> 00:56:42.880
And now you've heard about how they got the information, you've heard why it's important and what can be some of the negative effects if you do not put all that information in.
00:56:43.039 --> 00:56:48.239
And we say negative, or just will be less, less impactful potentially, or less effective.
00:56:48.480 --> 00:56:53.679
Um, I any effort I appreciate, and you I if you're putting messages out there, that's great.
00:56:53.840 --> 00:56:57.119
But I'm also at a point where I'm expecting you to do more now.
00:56:57.280 --> 00:57:01.599
I'm becoming less forgiving because the information is out there.
00:57:01.679 --> 00:57:03.440
And just as Jeanette said, it's free.
00:57:04.000 --> 00:57:05.360
It is absolutely free.
00:57:05.519 --> 00:57:16.239
There, there isn't anything stopping you other than you know, it could be your policy, your you know, government decisions, whatever it may be, but the information is available.
00:57:16.480 --> 00:57:21.599
Um, and again, thank you, Hugh, for breaking it down in a way for us to understand.
00:57:22.000 --> 00:57:33.119
Um and I honestly think if you if you've heard this one, if you heard the the part one as well, it's really at a point where you can probably explain it to someone else as well and the importance.
00:57:33.199 --> 00:57:46.079
And if you're uncomfortable with some of the impact statements, like the one that I know catches people all the time is when we talk about an active assailant or an active shooter, where we list in there, it's like you could die, right?
00:57:46.239 --> 00:57:48.400
By not following these these things.
00:57:48.559 --> 00:57:51.360
And there's this, whoa, we can't tell anybody that.
00:57:57.840 --> 00:57:59.840
Like that this isn't a surprise.
00:58:00.079 --> 00:58:07.039
It's trying to prevent other people from getting into that danger area where they could put themselves in that greater risk.
00:58:07.280 --> 00:58:11.039
So if it is new to you, I say welcome, right?
00:58:11.199 --> 00:58:13.039
If you're already doing it, great.
00:58:13.199 --> 00:58:14.880
Um, and just keep spreading the word.
00:58:15.039 --> 00:58:16.079
But those are my final thoughts.
00:58:16.239 --> 00:58:16.719
Let's go around.
00:58:16.880 --> 00:58:19.920
Hugh, you want to go next, and then Jeanette can close us out.
00:58:20.559 --> 00:58:33.440
Yeah, I'm I'm just really grateful to be here, being able to talk about the work that we've done and uh kind of go back to school myself and and get some some good time in in thinking about how to break these basic concepts down uh for myself.
00:58:33.519 --> 00:58:42.559
Because we spend the more time we spend with the basics, the the better we're gonna be able to make decisions about how we do all these things in the future, which I think is true for everything, not just for research.
00:58:42.639 --> 00:58:43.920
But it's been a great opportunity.
00:58:44.079 --> 00:58:45.199
Really appreciate it.
00:58:46.400 --> 00:58:49.199
This has been a fun two-part series.
00:58:49.280 --> 00:58:55.039
And I hope that the audience that listens to the alerting authority has gotten as much out of it as I have.
00:58:55.280 --> 00:59:10.000
Um, and I've just really enjoyed talking through these theories, the methods, and sharing with everyone about our latest research and how important it is for shaping the way we think about alerts and warnings and moving forward in the future.
00:59:10.239 --> 00:59:24.239
So this is the Alerting Authority, where every second has a story, and we are so happy that we got to help share Hughes and our research story and yours today.