In this episode of K9 Conservationists, Kayla speaks with Dr. Charles Van Rees about data. They discuss the ins and outs of working with data and how to turn data into information scientists can use.Episode suggestion: Take the time to see, smell, and notice things outside!
What is/are data?
- Information that we can use for inference and for learning things
- Data doesn’t necessarily “teach you things” but it helps you learn an answer to something you’re looking for
Examples of data for conservation science
- Occurrence data: evidence of the occurrence of a species in a specific area
- Abundance data: identifying and counting every individual of every species in a given sector
- Time series data: a series of data points indexed in time order
- Monitoring: the process of watching the changes and patterns in the species
Challenges in data collection
- Biases in both humans vs dogs
- Dogs have less of a visual bias than humans as they rely on olfactory vs sight to find data
- Important to consider if the questions you are asking from the data are going to be worth it to have dogs involved.
- Sample size: you have to have enough data in order to be able to work with the data
- Trying to collect data in a cost effective way
- You can’t control the data you collect
What happens to the life-cycle of data?
- FAIR principles
- Meta-analyses and reviews
- Or if not, advocacy! Action! Management!
What are ways to collect data?
- Conservation detection dogs
- Radio collars
- Camera traps
- Audio surveys
Links Mentioned in the Episode:
- Mirror test (Kayla got this slightly wrong – dolphins spent extra time head circling, which may indicate recognition)
- Dogs self-recognition with scent
- Wolverine Way book
- What the Dog Knows
- Inside of a Dog
- Are We Smart Enough to Know How Smart Animals Are?
- Chicken Wire Gauge preference study
- Nathan Hall
Where to find Charles: Website | Twitter | Instagram | Nature GuysYou can support the K9 Conservationists Podcast by joining our Patreon at patreon.com/k9conservationists.K9 Conservationists Website | Merch | Support Our Work | Facebook | Instagram | TikTok
Transcript thanks to volunteer Meg du Bray
015 – Data Nerdfest! Part One
Sat, 2/12 5:46PM • 44:12
Data, data collection, longitudinal data, conservation biology, ecology
Hello, and welcome to the K9 Conservationists podcast, where we’re positively obsessed with conservation detection dogs. Join us every week to discuss ecology, odor dynamics, dog behavior and everything in between. I’m your host, Kayla Fratt, and I run K9Conservationists, where I train dogs to detect data.
Today, I have the absolute joy of talking to Dr. Charles van Rees about the lifecycle of data, how to collect data, what even is data, the types of data, it’s just a big old data nerd fest. Aside from being one of my very dearest friends and a board member at K9Conservationists, Charles is a conservation scientist and naturalist who combines research and nature interpretation to change how we manage, protect, and relate to the natural world.
He is a published scientist with so much incredible research under his belt. He’s currently a postdoc at the University of Georgia. He was a Fulbright scholar at the Estacion Biologica de Donana in Seville, Spain. He has worked at the Flathead Lake Biological Station in Montana, which is where we originally met. He’s an incredible person, and I think you’re really going to enjoy this podcast. It does get pretty nerdy, and we do go into the weeds, but we try to bring the things back to dogs a lot. I’m really excited to get into this interview; again, Charles is one of my best friends and you’re going to be hearing more of him on this podcast, because we’ve got a lot of things left on the table to talk about.
Before we get to it, we’ve got to talk about our weekly suggestion. This one comes from Charles. He suggests that we make sure we take the time to see, smell, and notice things outside. You know, one of my favorite things about being outside with Charles is we both find so much delight in a little bug on a flower or a beautiful yellow warbler that flies across a river. Even as you’re walking down a street, on your way to the grocery store, you might notice an almost perfect little ladybug on a leaf. Just notice those things and take a deep breath! I can’t agree with that suggestion more.
The last thing before we get to this is that we finally have a new review! You all know how much these new reviews mean to me, how excited I get when we get reviews on the Apple podcast app. If you give us a review, you will just absolutely make my day. I know that there’s like, 900+ of you listening to this podcast, and yet we only have like 20 reviews, so step it up guys! This week’s review highlight says, “I found this podcast while browsing a while back. Binged all the episodes in about a week. It’s super great to learn about the field of conservation canines. As a dog trainer trying to find my training passions, this podcast has inspired me to chase conservation training for a while. Even when technically heavy, Kayla finds a way to make the information easy to understand.” I hope that we live up to this reviewers expectations in this episode, because again, it is pretty heavy. Again, these reviews just make my day, so if you can do one nice thing for me, because I do put so much time and love and effort into this podcast for free for y’all. Go ahead and leave us a review on Apple podcasts.
So now let’s get to it with Dr. Charles van Rees. All right, well, welcome to the podcast, Charles.
Charles Van Rees
Thanks so much, Kayla. It’s a pleasure to be here, I’m a real fan of what you guys do. And honestly, I just think the whole idea of conservation detection dogs is a fantastic overlap of things that I think are lovely.
Yeah, it’s an easy thing to get behind, in my experience. I mean, obviously, I’m biased.
So let’s start out with: what is data? What are data?
Charles Van Rees
Yeah, data, I guess is plural for datum. I think if you’re trying to be a stickler about it, you would say what are data and some people get a hang up about that. I sometimes do, but usually in scientific contexts. But what is data? I think from a practicing scientist’s point of view, from the point of view of someone like me, data are information that we can use for inference and use for learning things. It doesn’t necessarily mean that we have this point of data, and now we know something, but it’s something that we can then use through processes of statistical inference to understand the probabilities that certain things are true or not true. To test certain hypotheses to make syntheses and generalizations. All that to say, in a fancy way, data are individual bits of information that, in summary we can use for useful knowledge.
Okay, yeah. Can you give us an example of a dataset? Maybe something you’ve used for your work as PhD, postdoc, as an actual scientist? The reason we’re talking about this today is that our tagline at K9Conservationists is “Dogs Detecting Data.” But we have this question of, what is the data? And where does it go? What does it do? That’s the whole thing we’re talking about today. Why don’t we talk about an example of data that you’ve used. I think the listeners are probably familiar with some of the data that my dogs and I have already collected.
Charles Van Rees
Yeah, I like this idea. This reminds me of – maybe I’m dating myself here – of Schoolhouse Rock, like, the life of the bill. What are these data doing?
To get past my very hand wavy, academic, fancy-sounding definition of data, sometimes examples are a better way to define things, so I think you’re very right to ask that question. I think that the data that are most often used in conservation biology and conservation science, which is my field and discipline, one of the most basic ones are what we call “occurrence data,” which is having a place and a time where a species or an individual from a population was.
If we get enough of those across time, we get information about the “spatial distribution,” as we call it: where those animals are found in time and space. Does that change over the course of the year? Has that been changing over years? That’s a big deal in conservation, because we’re worried about what we call spatial-temporal trends. Do we see the range in which this animal is found shrinking? Or that distribution? Do we find fewer of them across time? Do we see them less often – is that a concern? That’s, in my opinion, some of the most basic, bread and butter data in ecology, and especially in conservation.
Collection is another word that will probably come up. When you collect those sorts of data across time intentionally to look for patterns, you’re doing what’s called monitoring, which is another major place where conservation detection dogs contribute. A lot of my work, especially during my PhD, focused on endangered water birds in Hawaii. A major thing with a lot of organisms is finding them in the first place so that you can get that occurrence data. You need to go out there and see them somewhere, and confirm that you absolutely saw them somewhere to have a useful data point. The birds that I was working with the beginning are the ‘Alae ‘ula, or Hawaiian common gallinule, really funny little animals. They are extremely shy, and very secretive, so detecting them and being like, “Oh, yeah, there’s one here,” was really difficult. And that really mattered, right?
I mentored a fantastic pair of undergraduates when I was at Tufts who did this great project of trying to understand what makes good habitat for the ‘Alae ‘ula. We want to know that so that we can make more of that habitat because they’re endangered. But if we don’t have an understanding of what aspects of the environment are associated with them being there, then how do you figure out what they like? You can’t go ask them! So what we need to do is just go to a bunch of different places. And of course, it gets fancier with the math and the statistics, but essentially, what we’re doing is visiting a whole bunch of places and saying, “Well, what’s similar among the ones that have them, and what’s different from the ones that don’t have them?” That’s what occurrence data could do for you. You could see, well, where are they found? Where are they not found? What does that tell us about the habitat needs of this animal? And then in terms of the lifetime of that data, you know, that then goes to the managers and they say, “You guys always found ‘Alae ‘ula in marshes that had more patches of open water.” Okay, we can do that, we can get in a backhoe, and open up some space in the marsh and do more for this endangered species.
That’s an example of kind of the applications and one really basic form of data. And you can scale up from there. And there’s all sorts of other things too. Abundance data is usually the next scale; you didn’t just see whether or not they were there, you counted how many were there. That’s even more difficult, because you have to count them. And then there’s, you know, did you miss one? Or did you over count accidentally, there’s all these issues around that. All these things that dogs are ludicrously good at.
As a conservation biologist, those are the types of data I work with on a regular basis, and you use them in so many different ways. You can use them to figure out habitat quality, historical trends, types of management, and whether or not they work, things like that. And then that goes into all sorts of larger decision making, because then people like me, who are sort of on the border between the science part and the doing part, and also talk to decision makers, we’ll be figuring out those patterns. And then the idea is you communicate those patterns with people who are doing the stuff on the ground, or the people who are making the big decisions, and they say, “Okay, now we’ve we learned something from that information, there’s always ways we can do that, let’s make some decisions about how we deal with environmental policy or protecting that species,” and so on. And of course, there are always problems with this. But ideally, every one of those data points is contributing in some way to some useful learning that then can be moved forward into a policy decision or something that affects us in the real world.
You’ve already made a bunch of really good points here. In terms of hitting on the things that make for a good dataset, you mentioned making sure that you’re not double counting a bird, which I would imagine from what I know about the ‘Alae ‘ula, they’re not necessarily all that individually identifiable unless you’ve got a banding system or something. So it’s really challenging, if you’ve got five, and then they kind of go around a bend in the marsh, and then they come back, and then there’s three and it’s like, “Oh, God is that eight, or is that five?” So it’s not just information, it’s not just counting, but also making sure we’re counting in this really reliable way.
Charles Van Rees
Yeah. Our jargon in the sciences for that is “rigorous.”
Yeah, exactly. One of the other things that I’m thinking of that you didn’t hit on, but is an interesting point when we’re talking about dogs versus humans for collecting data, is that it’s not dogs versus humans. Every time I’m out in the field, I’m also working as a field tech. So you’ve always got both: every time you’ve got a dog, he’s got a person. And we don’t want people to think that we’re poo-pooing their field techs or their people.
Detection dogs often have different biases from people. So one of the things that can come up is that it’s much easier for human searchers to find the sexually mature males, or the animals that are ranked higher in the dominance hierarchy, because those are the more visible animals. They move through the environment in a way that is easier for humans to see. Because dogs work in this olfactory world, it’s easier for them to tap into some of these subordinate animals, the immature animals, the transient animals. And that goes for plants as well. It’s really easy for a human searcher to find a plant that’s as tall as they are in this environment, or when it’s flowering or whatever. But as soon as you’re talking about a two-inch-tall rose, humans really don’t have a prayer in a lot of cases.
Charles Van Rees
I really like the point you’re making here, which makes me think of another sort of ecology science nerd thing we can touch upon briefly. There’s a field in ecology called sensory ecology, which is this idea of studying the ways that different organisms view and experience the world through, for example, what you were touching upon with this idea of dominant sensory modes. As humans, we’re the strange mammals. Most mammals are living a very sniffy world. Everything is about the sniffs and what they’re smelling and what they’re putting out. And everybody’s stinking a different way. And that’s how they’re working out their stuff. And primates happen to be unusual. And we are all these weird looking animals with flat faces and these eyes facing super forward. We don’t look like a lot of other mammals. And we’re super, super visual. And, I guess, I’m not super up on human olfaction. But the things that I’ve read have mostly said that we have a pretty darn sensitive nose for certain things, but –
I believe Dr. Nathan Hall has talked about this before, probably on the on the K9s Talking Scents podcast, humans can actually – we either have more receptors, or we can actually outperform dogs in finding vanillin.
Charles Van Rees
Yeah, that sounds familiar. There were certain things that we’re probably better at.
Yeah, Inside of a Dog, by Dr. Alexandra Horowitz, and What the Dog Knows by Cat Warren, both of these books talk explicitly about this. Both of those authors, I believe, are search and rescue handlers, so they’re talking a lot about being out searching for missing or deceased people, and talking about learning what the dogs perceive. And yeah, we’ve got these this binocular vision, we’re upright, we’re taller than a lot of animals that are much, much bigger than us, which makes it easier to see more stuff. You know, being taller sometimes helps for olfaction, but not really.
Charles Van Rees
Yeah, we’re so visually focused that we’re going to have those biases. And again, if you’re studying anything that is operating in a sensory space that is different than humans, probably a dog’s going to be a lot better way to find those things in a lot of cases. In fact, one of the few places where humans seem to have an easier time, and you can see this in both people’s hobbies and in the science, is ornithology. And birdwatching. Birds operate in a scent space, and there’s a lot of stuff that dogs can do with regards to birds, and we’ll get there, I hope. But the reason people like bird watching so much is because birds operate on much more of a similar sensory space to primates than primates do to the rest of mammals. Birds are all about sounds that are specifically mostly in our hearing range. And they’re about types of light and color that are mostly within our vision range. They can see some colors a little further on either side of the spectrum than we can, but that’s why it doesn’t require a lot of fancy equipment, or a lot of time to study birds. You can just go out there. You can hear what they’re saying and see what they’re doing because they’re speaking the same language, but I can’t go out and study wolves like that. They’re living this whole universe of sniffs that I can’t. I’m noseblind to that.
Exactly. And that’s an actual term that’s used quite often around humans. Noseblind is a term that I’ve heard quite frequently. I
can go out and I can tell like, “Oh, that male song sparrow is starting to set up a territory. He’s mad, because so and so’s moving in over here.” And I can hear all of that, and I can be aware of it. I can’t go out and say, “Ooh, that mink that passed through last night? She was coming into heat,” you know, like, I can’t do any of that with a mammal. And we might not know this, both Charles and Kayla, but also humans in general – I wonder if being more visual, it would make sense that that would track to some degree with being a little bit more arboreal and operating in this 3D space. You need some amount of visual acuity to be operating in that 3D space. I don’t know. Bats are kind of an exception, but they’ve got a different superpower. Who needs to who needs to see when you can use echolocation.
One of my favorite things to just think about is the concept of umwelt, is that something you’re familiar with? Oh my gosh, Frans de Waal talks about it a lot. He wrote an incredible book called, Are We Smart Enough to Know How Smart Animals Are? And one of the things that is really common in this field of animal cognition is this this idea of umwelt, which is basically the idea of understanding how your study subject, or how another animal interacts with the world. Kind of like the sensory ecology. They use it a lot in animal cognition studies. So if you’ve got a dolphin, and you’re trying to figure out, can this dolphin recognize itself in a mirror, and therefore it does it have a theory of mind, one of the things you would do with people, or with chimps is you might knock them out and then put something on their face to see if, when you show them the mirror, they try to rub that, that red dot off their face. And theory is then they recognize that that mirrors themselves and then they’re trying to get that thing off their face. And people did this at some point with dolphins, and then dolphins didn’t do anything.
Well, there’s two major things wrong with that. This dolphin, which lives a world of echolocation is inside of a tank, and the mirror is outside of the tank. And a dolphin doesn’t have hands, so how on earth would a dolphin try to – like, how would you know that they’re trying to remove a red dot from their face? I’m not sure in the study if they used red, but if they used red, that’s a horrible choice for an aquatic animal or a marine mammal! Similarly they’ve done this study with dogs, and dogs also didn’t really react. But again, these dogs are so olfactory! Then later, someone thinking about umwelt came along and did another study. I will try to dig up all these studies, because I’m sure I’m quoting my 2015 college course right now, so I will try to dig them all up for our show notes.
But for dogs, I think they did something really weird where they altered the scent of the dog’s urine somehow, and then re-exposed the dog to their own urine. And with that change, the dogs spent much longer investigating their own urine when they were like, “That doesn’t smell like what I ate yesterday!” Which, if dogs were designing a test to see whether or not humans could recognize ourselves, they would probably mess with our urine. And we’d do horribly! Like, why isn’t the toilet flushing? It comes down to that saying of, if you want to test a fish’s intelligence, don’t base it on how well it flies or something along those lines. I don’t remember what it is. But this is a fascinating crossover of sensory ecology and animal cognition.
So we’ve got this occurrence data, we’ve got this abundance data, then what about these longitudinal studies? I just finished reading Doug Chadwick’s book The Wolverine Way, and that was a fabulous book. Incredibly beautiful, and just really cool science journalism. In the book, they’re following these wolverines using both GPS collars and other implanted transmitters. So, that’s not really occurrence, or abundance necessarily; that’s following an individual longitudinally. Is there a word for that? What is that?
Hey, everyone, just popping into this episode with an update on our Patreon. We still have the $3 a month Doggie Detector level, which allows you to ask questions for me and the guests to answer each episode. But that also lets you join our monthly training video analysis calls. These calls are going to be recorded of course, and we’ll also publish the video afterwards, for everyone to view and ask questions about prior to the call to ensure that all timezones can participate fully. We’ll publish the video we’re going to analyze so that you can ask questions and view it and prepare it ahead of time, then we’ll have the call where we talk about it, we can have beverages, it’ll be a good time. And then all of that is going to be shared later, so you can participate before, during and after, again, just for three bucks a month.
Now, at the $10 a month Scentsational Scientist level, you get everything that we got before at the $3 level, plus you get to submit videos of your training sessions for those calls. This is perfect for the aspiring canine conservationist. And your target odor doesn’t really matter here as long as you communicate what it is so we can think intelligently about your goals. This is going to be great for nosework competitors and other canine handlers as well. We’re really striving to make these video calls super kind and supportive and helpful, so it’s going to be a nice safe place on the internet to get good feedback on your training sessions.
Then finally, the Canine Conservationist patrons get everything from those other two tiers, plus a private 30 minute training call with me to go over whatever you’re running into with your dog, and that tier is just 25 bucks a month. That’s cheaper than booking my time at JourneyDogTraining.com for behavior modification, and that’s just because I love you, and I love my patrons. You can join that Patreon over at patreon.com/k9conservationists or at the link at k9conservationist.org. It’s like a tiny link up in the top bar. And then we also drop that link into our show notes. So if you’re listening on your podcast app, you should be able to find it just right from there. Thank you guys so much, and let’s get back to the episode.
Charles Van Rees
Oh, jeez. In terms of the typology of data, I think typically, we’re referring to that as time series data. And there are certain mathematical assumptions that you have to account for when you’re analyzing that, but that’s not really something we dig into so much. But studying individual animals, or I guess, I might even talk about that as identity data. You’re not only just looking at how many of something, or whether it’s there or not, but who is it? And then people worry about anthropomorphizing or what have you, but like, really, who is this? Is this number 875? Or is this number 217B? Or is it Bob, or George, or Helen or whatever you choose to name? Yeah,
Yeah, Flo and Flint, and Fifi.
Charles Van Rees
I think my birds ended up with nicknames despite my best efforts.
I was actually naming chips from Jane Goodall’s study, because I’m the sort of person that knows all of their names.
Charles Van Rees
I mean, I’m not super surprised.
Yeah. But Flo was the major issue; she was the main female who was the most reproductively successful female in Gombe. And all of her babies had F names, so they’re easy to remember. She named them with matrilineal letter names. So there’s the G family and an F family.
Charles Van Rees
Nice to have a system, I like that. But this brings up a very important point. There are two things we’re touching on now: the extra dimensions of data. We have the longitudinal side; when you’re collecting data over a long period of time, you can learn all these special things about what changes at long timescales. Another issue we have in our sensory world as humans, is that we exist on a certain timescale. We live 80 years or 60 – we have a certain generational time, we operate on those scales. Some animals live way longer than us. Some of them live way shorter than us. But we have a perception of time, not just a perception of space and vision, and whatever else, that’s different. And different ecological phenomena and species are operating at these different temporal scales, so in order to understand them, we usually study things across time, especially when we’re dealing with things like what we call now global change, planetary change, which has to do with both global climate change, but also deforestation, water scarcity, all these changes we make in the planet. Those impacts operate at such big scales that we would not know that there are things happening, unless we had people studying them for long periods of time. And that’s hard, right? All it takes is for a few things to go wrong, and suddenly you miss five years of data, or you run out of money, you can’t go do that study anymore. There’s a whole branch of the National Science Foundation, that is based around funding long term ecological studies.
I’m just thinking of a bristlecone pine, which can easily live for 5,000 years. I don’t care how good the NSF long term funding plans are. It’s not gonna happen. And if we’ve got any geologists in the audience, like, I’m sorry, because at least most biotic things that we’re looking at, are on a timescale that we can kind of wrap our heads around. With the exception of some of these really long-lived trees. Even if you’re looking at a whale, or a tortoise, or something that does live on a scale similar to humans. And even studying humans! I’ve read and listened to some really interesting studies just looking at childhood effect or you know, childhood factors. And it’s really hard to do those studies with our own species. And we can follow our own species in theory, and like, keep up with their address changes or whatever. But you know, like, forget it when you’re looking at like a humpback.
Charles Van Rees
Right. That ties into the other aspect of data. We’re discussing this other possibility of identity, which is easier with us, because we can recognize faces and whatever. I think a lot of people imagine animals, for example, as being more homogenous than we are, because they can’t tell the difference between them. But again, that’s a sensory thing. We’re not usually staring at the faces of dozens and dozens of Magnolia warblers, so we can’t tell the difference between a bunch of male Magnolia warblers, whereas like, I guarantee you, if you were a male, Magnolia warbler, you’d be like, “Oh, that’s Frank. He’s a real jerk, and I hate that guy.” But when you know identity then you get to learn all sorts of interesting trends about what makes them more successful, or how many babies do they typically have in their life? How do their social dynamics work? Knowing the identity of an individual allows you to make these kind of narrative stories that can tell you all these other things, and look at correlates of behavior and interactions and things as another big dimension there.
And as we wander off into these tangents, I’m trying to tie it back to the dogs a bit. I think as humans, like we mentioned banding birds, like, that’s why we do it. Because, honestly, I had stared at so many ‘Alae ‘ula by the time I was done with my PhD that like, yeah, I could tell some of them apart just from something I was picking up on but not like that, you know, we still put these colored bands on them.
Well, even with our dogs! You know, I really like Border Collies like Barley, my border collie. He’s long haired, he’s black and white. He’s got floppy ears. He is the quintessential Border Collie. If you go to a pet food store and you go to buy a bag of Hill’s pet food, the dog on the front of that looks a lot like Barley. I can tell it’s not Barley, but I will tell you at least once at a park, I have had a different Border Collie run up to me and think it was my dog. And I live with him!
Part of the reason I was asking about longitudinal individual data is because one of the things I find really interesting about it is it’s one of the areas where I think I see less application for dogs. I can see dogs supplementing that work, but not necessarily replacing other methods, particularly when we’re talking about these big studies with collared animals. I’m obviously picking on these wolverines because that’s the book I’ve recently read. But in theory, you could train a dog to track an individual wolverine; there are some dogs in Africa that have been trained to track individual rhinos for this longitudinal data. But at what point is it no longer practical, especially when you’re looking at an animal like a wolverine that treats the shortest distance between any two points as the right way to go? Regardless of the terrain in between. At least as a human handler, I disagree. Barley might not mind following wolverines around, but I would.
It’s just interesting to think about, like dogs are helpful for this occurrence data in particular, because they’re so good at finding stuff that’s hard to find. They’re pretty good. They’re really, really helpful for abundance data. But not quite as much for this kind of the individual level, like, how they’re moving through space, what their life story is like. Some of that you kind of need to see it, like just knowing where an animal is. Which dogs can do; is it here, was it here? When was it here? That’s useful, but what was it doing there? That’s where a camera trap or actually having eyes on an animal might be most necessary for your question, which kind of ties into one of the upcoming episodes I’m going to be recording. It’ll come out a while after this, but it’s going to be with a couple other conservation dog handlers and thinking about what you need to know before you hire a conservation dog team. And I think that’s one of the things we’re kind of circling here. If you’re a biologist, if you’re a conservation scientist, or a decision maker in any way, you know, and you’re listening to this podcast, and you’re like, “Oh my god, Kayla and her dog sounds so cool. I’d love to hire her – “
Charles Van Rees
Yeah, we’re not debating that point.
But are the questions you’re trying to ask questions that are best served by having dogs involved? And I’m planning on doing at least one more episode fully on that question.
I think we’ve answered this question a little bit already, but what are some of the challenges in data collection? Again, we’ve already hit on some, but are there any that we haven’t brought up yet that you wanted to make sure we mentioned?
Charles Van Rees
Well, let me think. Are we thinking specifically about challenges that are addressed by conservation dogs?
No! You haven’t worked with conservation dogs, so you don’t have to answer that question. What are some of the things that are hard about being a PhD in conservation biology and Behavioral Ecology?
Charles Van Rees
Well, speaking to data challenges, I think the big thing that we’re always thinking about as scientists when we think about having data sets, or multiple data points, is sample size. The issue is, we have to do something with these data, we can’t just have them. We can’t just sit there like, “Okay, we have the information, we have to do some form of inference, we have to try to pull something out of that distill that into useful information, which requires testing hypotheses, or comparing expectations to reality, things like that, right?”
We talked about confronting a model with data. This is a very, very interesting scientific term. But the idea being, we are coming up with models, which in this case, is our idea about how we think the world works, or how something works. We go and collect a bunch of data and say, “Okay, well, does this fit? Does this make sense based on what we’re thinking about?” And in order to be able to trust what you’re seeing, you want to have a lot of evidence. It’s not always as brute force as this, but there’s kind of a force of evidence approach where you want to have enough information to say, “Yeah, okay, I believe that.” If I told you that I had a perfectly fair coin, that every time I flipped it, there was a 50/50 chance of heads or tails. And I promised you that it was 50/50. And I said, “Okay, I don’t know.” We’re rolling something high stakes, right? You have to give me your van, or Barley, but if I only showed you one flip, would you believe me that it was a fair coin?
Kayla Fratt (KF)
Yeah, of course not.
Charles Van Rees
I mean, I trust you. If you’re a random guy at a bar, no.
Charles Van Rees
This might be getting to be a too loosey goosey example, but the point is, if someone’s like, “Well, it’s totally works,” and they flip the coin once, you’re not going like “Oh, yeah, that’s proof that it works.” You want to see a couple hundred flips or something, right? And see if the data actually looks like it’s close to 50/50.
If we’re trying to infer these more complicated patterns about the natural world, or about this endangered bird in Hawaii, or about this invasive vine that’s spreading through the river systems of Mississippi or something, we can’t just go and collect a couple things and say, “It’s okay, let’s go back and think about it.” We need large amounts of data. And that makes it challenging to have enough money to collect the data and things like that. So, ways of collecting information and data that are cost efficient, and quick, and things like that? That’s a major challenge, I think that we face as scientists: getting the data in the right way. That’s a big one.
And then there’s also things like replication and control. If you’re trying to compare different things, we’re not in the laboratory where you can control everything and have only two things differ and make these raw inferences. Ecologists and conservation biologists are dealing with a complicated reality. There’s all sorts of stuff going on. Maybe we’re trying to compare these two tracts of forest and how many of these beetles they have in it or something. And then one day, like a bunch of college students throw massive party in one of your study sites, and leave Solo cups everywhere, and that brings in spiders that murder all the beetles or something. The point being, complicated things just happen. I mean, if you only look at two sites, you’ve learned nothing. You also have this variety and these ways of controlling for all those dynamic factors.
I think those are some of the big, logistical, conceptual challenges we face as scientists: using the information we get, we need to find ways to make it useful. And if it’s not controlled and rigorous in those ways, you can collect data that are less valuable than other types of data.
Yeah, of course. This reminds me of something I noticed on the wind farm this summer with Niffler. Towards the end of migration season, we went from getting really high numbers of bats every single day to really, really low numbers quite abruptly. The first couple days that happened, I was just walking behind him being like, “There is no reason, training-wise, that he would go from finding all of the bats to finding none of the bats. So I have to assume that this is a migration change.” But it’s always this question of like, “Is the dog not finding something, am I not finding anything because it’s not there? Or because of something that’s going on with our training? Or something that’s going on with conditions?”
We had one day where we were searching, and it started pouring on us. And it’s like, well, that’s probably damping down the scent a little bit. Dr. Nathan Hall, who’s at Texas Tech, does a lot of really cool research looking at environmental conditions for dogs and their ability to scent. But figuring out how to extrapolate that out to the real world, taking lab data into the real world, and looking at the messiness of the real world is a challenge. Actually, when you were talking about the ‘Alae ‘ula and their habitat preference, you said, “Well, we can’t ask them.” And yeah, I suppose we can’t. But there’s a somewhat famous study with chickens; there was a law that was being passed, I believe in the EU about the gauge of wire that needed to be at the bottom of laying hens contained in cages. And they were going to change it to this thicker gauge wire, because they thought it would be more comfortable for the birds. And then some scientists, at some point were like, “Oh, I wonder if they actually prefer that?” And they set up a bunch of birds in these cages that were identical. They had food and water on both sides. But the birds spent more time on the side with the original cage floor, and I think they ended up reversing, or not passing that legislation, I can’t remember whether it had gone through or not. But yeah, we can’t do that. Even if we had a bunch of captive ‘Alae ‘ula, and started asking them like, “Hey, do you prefer more water or less water,” figuring out how to extrapolate that to the reality of Hawaii, which doesn’t have all that much spare land that we can just mess around with.
Charles Van Rees
There’s very little, in fact.
Yeah, so even when we can get some of these answers in the lab world, it’s hard to extrapolate it out sometimes.
Charles Van Rees
That’s a good point, yeah.
Alright, we are going to cut the conversation with Dr. Van Rees off here, because it did go a little bit long and we are going to turn this into a two parter. Tune back next week to hear a little bit more from our data extravaganza with Dr. Charles van Rees; I think you guys are going to continue enjoying this conversation. As always, thank you so much for listening. I hope you learned a lot and are feeling inspired to get outside and be a canine conservationist in whatever way suits your passions and your skill set. You can find shownotes, donate to K9Conservationists and join our Patreon over at K9conservationists.org. You can find Charles at gulothoughts on Twitter and guloshots on Instagram. You can also find him at VanReesconservation.com, and I’m sure he will be thrilled to hear from any of you. Thank you so much for listening, and we will be back in your earbuds with the continuation of the conversation with Dr. Van Rees.
-Data points on their own don’t necessarily have meaning; collecting a lot of data and using statistics allows us to make meaningful inferences
-Some of what happens in the lab doesn’t have the same relevance in the real world, particularly in the world of conservation biology and ecology