The Signal and the Noise: Why So Many Predictions Fail – but Some Don’t  

by Nate Silver

For those of you not familiar with the author, Nate Silver is the founder and editor in chief of www.fivethirtyeight.com, a website which features statistically-oriented articles on sports, politics, economics, and whatever else might catch the interest of the contributors. Silver became the ire of partisan political pundits in 2008 when he correctly picked the winner of 49 of 50 states months before to 2008 presidential election. FiveThirtyEight manages to present statistical-based discussion of topics in an interesting way.

The Signal and the Noise follows this style as Silver explains how predictions are central to things as diverse as meteorology and baseball. As the title implies, the book revolves around a central issue facing predictions: separating signals (what actually is associated with resulting phenomena) and noise (what might be confused for meaningful information, but really isn’t associated with resulting phenomena). An example he returns to is that who wins the Super Bowl is indicative of how the stock market will fare in the following year.

In his book, Silver advocates for a Bayesian approach to statistics. Simply put, Bayesian statistics are one tool which can be used to determine the probability of two phenomena being related, and adjust that probability as new information is acquired. Many of his examples throughout the book explain Bayesian approaches to predicting earthquakes, terrorist attacks, or winning a hand in poker.

The Signal and the Noise emphasizes that through statistics, we can understand much of the world around us, but that we must do so cautiously. I found it very interesting and entertaining, and found its approach very accessible for non-experts in statistics.

My summer TA assignment is with Field Zoology at Stone Lab, Ohio State’s island research station on Lake Erie. The course takes a hands-on approach to learning about the animal diversity on and around the Lake Erie Islands. In the first week, the class has visited South Bass Island, Kelleys Island, as well as the shores of Gibraltar Island to sample and study.

I’ll share more about the class and our experiences in coming weeks, but I thought I’d share a few cool pictures and descriptions of some of the things we’ve seen.

Freshwater Drum (also called Sheepshead, Aplodinotus grunniens) is a very common fish in Lake Erie, video here. In fact, they’re common in a lot of places, and their range covers most of the Midwest, north into Canada, and south to Central America. One of their closest relatives is the marine Red Drum (Redfish for Floridians, Sciaenops ocellatus). In contrast to their marine cousins, Freshwater Drum suffers from a bad reputation amongst anglers and are considered an unpopular fish to eat, although there is a movement to increase their popularity. If I get a chance to eat one up here I’ll report.

You’re probably quite familiar with aquatic crayfish in lakes and streams, but a number of species are also terrestrial and live in burrows several feet deep when they’re not venturing out at night to find food. These species build chimneys out of their burrows, some of which we came across on Kelleys Island. There is even a species of frog that will take over these burrows and live in them. Shockingly enough, that frog is known as the Crawfish Frog.

This weekend some friends and I had an opportunity to look at some of the fish species in a local creek. It was a great time canoeing, seining, and relaxing; and we luckily saw plenty of fish representing many species.

You can get a lot of information about the ecology of animals based on their body type. For fish this is especially true, particularly in the orientation of their mouth. Take for example a fish like this Banded Darter (Etheostoma zonale):

Notice the mouth is on the lower half of the head. This is a pretty good indication that this fish lives on the bottom. Darters do in fact live on the bottom of creeks and streams where they move between and under rocks looking for insect larvae to eat.

Its not the greatest picture, but this next fish is known as a Blackstripe Topminnow (Fundulus notatus):

Notice the upturned mouth. As its name implies, these fish live along the surface where they feed on other types of insect larvae.

An upturned mouth wouldn’t be very useful for finding food on the bottom, just as a downturned mouth wouldn’t be useful for finding food on the surface. In response to the specific foods fish eat and the habitats they live in, fish have evolved a great diversity of mouth types and orientations. This specialization can tell you a great deal about the ecology and behavior of particular fish.

My name is Adam Reddon, and I study the behavioural biology of social living. I have a particular fascination with both aggression and sociality and I endeavor to better understand these behaviours through the integration of functional, developmental and mechanistic perspectives. My Ph.D. work focused on social decision making in the contexts of grouping behaviour and resource contests in a highly social cichlid fish. I am finishing my PhD at McMaster University this month and will soon be moving to Montreal to take up a postdoctoral fellowship at McGill University in Simon Reader’s research group. During my postdoctoral work I plan to examine the effects of maternal experience on the social behaviour and physiology of her offspring in guppies.

 

Email: reddonar@mcmaster.ca

Website: http://www.science.mcmaster.ca/psychology/people/262-adam-r-reddon.html

Google Scholar: http://scholar.google.ca/citations?user=MM1KKXcAAAAJ&hl=en

Twitter: @adamreddon

Fight Club

Among those that study how animals interact with one and other, there are two basic groups: those who are mostly interested in what the boys have to say to the girls (or vice versa), and those who are mostly interested in how animals settle disputes among members of their own sex. The former group is certainly larger than the later, as evidenced by the fact that this Animal Behaviour Society Meeting had around 60 or 70 talks on sexual selection broadly defined, including a couple of plenaries, while there was but one session on agonism, or fighting as it is better known. Furthermore, the lone agonism section happened to fall into the last slot of the last day, a time when many conference attendees are often running short on attention span or have wandered off to explore whatever lovely locale in which the conference is set (Boulder, CO has this in spades). So, you can imagine my trepidation when I discovered that my talk was in the last slot of this last session, on the last day, in a topic that warranted only the lone session out of around 80 or so. I envisioned speaking in front of the other speakers from the session, 100 empty chairs and perhaps a janitor catching a nap in the back row. Much to my delight, the agonism session was very dynamic and well attended! This, I think can only be credited to the exceptional talks that joined my own in rounding out this session. All five of the other talks very interesting and engaging. I learned a great deal and very much enjoyed this session. Below are some brief descriptions of each of the 6 talks:

Sara Decker, University of Wyoming – Sara’s talk was about dominance hierarchies in a very interesting species of bird, the long-tailed mannequin. Long-tailed mannequin males form groups and work together to attract females. Only the dominant most male gets to breed though, and he has a second in command that is next in line for breeder status. Sara found that if both of these two top ranked males disappeared at the same time, the group descended into chaos and in-fighting, which apparently long-tailed mannequin females do not like, as they stop coming around until the chaos dies down and the hierarchy becomes stable again.

Bruce Lyon, University of California Santa Cruz – Bruce talked about dominance interactions in gold-crowned sparrows. Gold-crowned sparrows, like many birds, use badges of status to settle disputes without needing to escalate to costly violence. Males of this sparrow have a bright yellow patch of feathers atop their head. Birds with a bigger patch are dominant over those with a smaller patch. In a staged interaction between two birds, you can change who wins by using paint to give one bird an artificially larger patch. Interestingly, Bruce presented data showing that this manipulation doesn’t work if the birds are from the same flock and already know each other. Presumably, they remember the other bird, know what that bird is really like and hence are not fooled by the new dye-job.

Chad Johnson, Arizona State University – The rapid expansion of Phoenix, Arizona from small desert settlement to one the largest, most bustling metropolises in North America has posed many challenges for the local flora and fauna. One animal that seems to be doing ok with the new neighbors is the western black widow spider. This poisonous character is doing so well it has become a major urban pest. Interestingly, Chad presents data showing that although western black widows are super abundant in the city, the city-dwellers are actually less healthy than their desert counterparts that have to eke out a living under much tougher circumstances. Chad found that this unexpected difference seems to owe to the fact that the city spiders eat mostly a particular cricket species, which is low on phosphorus, a vital nutrient for western black widows. Basically, city spiders are gorging themselves on junk food, which lets them reproduce quickly, but spiders from more natural areas are getting a more varied and complete diet, resulting in better overall health.

Russell Ligon, Arizona State University – Russell studies aggressive interactions in veiled chameleons, which can change their body colour rapidly, and may do so to communicate with one another. Russell found that the brightness of a chameleon’s body predicted its likelihood of escalating to physical fighting, while the chameleon with the darker head was signaling submissiveness and was unlikely to win the fight.

Vikram Iyengar, Villanova University – Vikram presented data on sexually selected trait divergence in a pair of related damselflies. These species may live together in the same areas, or in different areas. When they co-occur, one species tends to have smaller spots on its wings than it does when the other species isn’t around. Previous research has shown that these wing spots are attractive to females. Vikram’s study showed that the wing spots also draw the ire of males of the other species, causing them to attack more often. Vikram’s data suggests that males in areas of overlap have been selected for smaller wing spots because although it makes them a bit less sexy, it also saves them some costly beatings.

Adam Reddon, McMaster University – Last up to bat was my talk. I told the audience about a recent study comparing the decisions during fights between two closely related species of cichlid fish, one of them highly social (frequent star of this blog, Neolamprologus pulcher) and the other (Telmatochromis temporalis) much less so. I predicted that because the social species may have more information about their opponents before a fight because they have interacted with them before, and because the social species may have a greater overlap in interests with their opponents, the social species would have less costly fights. I also predicted that the social species may find a way to end a conflict that allowed them to stay in the same area and continue to interact with their opponent after the fight, whereas the less social species may just run away and go somewhere else. My results confirmed both of these predictions, suggesting that changes in the way animals fight may be tightly linked to their social system, how they live with and get along with other members of their species. The morning of my talk, I found out that the paper this data was drawn from was accepted for publication in the journal Behaviour. I considered this a good omen, and sure enough, my talk went great. I was happy with my performance and got lots of positive feedback afterwards. Best of all, lots of smart people hung in there to the end and listened to me speak. It was a great way to end a great meeting.

Cody Dey is an evolutionary ecologist who studies social behaviour in birds and fish. He is primarily interested in signalling, mating systems, whiskey and beards.

As you have probably heard in Isaac’s previous incarnations of “Weird Bio Sh*t”, animals do some crazy stuff. Some of the craziest traits and behaviours we observe in the natural world are those that are involved in reproduction. For example, males of many species of birds have extremely elaborate ornamentation and behaviours (check out the wild dances and songs in this video) that are used to attract mates. The evolution of traits used only for reproduction is extremely interesting to behavioural biologists because these traits are the product of strong selection (if you don’t mate your genes are quickly removed from the gene pool), however sexually selected traits often decrease the survival of their bearer (e.g. the peacock’s long tail makes it more prone to predation because it is bulky and conspicuous).

Sexual selection was first thought up by Charles Darwin himself. Typically, we think of sexual selection occurring when females choose a mate from a pool of different males. In order to stand out, males have evolved special attractive traits that are used to broadcast how sexy they are to females. Alternatively, sexual selection can also occur when males have to fight to gain access to females. In this case, males often evolve complex weapons that are used to duke it out (think moose antlers or fiddler crab craws).

This year at ABS, there was a great deal of discussion of how, why and when sexual selection occurs in animals and what consequences it has on the natural world. Robin Tinghitella told us how female choice for mates can be affected by mate availability and age. Apparently, female sticklebacks (a species of fish) are less choosy when males are harder to come by. Also, female sticklebacks get less choosy as they age. The ability of females to adjust how choosy they are allows female sticklebacks to balance the goal of gaining a high quality mate with the costs of not getting a mate at all.

We also heard how sexual selection can lead to speciation (the process of creating new species!). Biologists first thought that sexual selection might be involved in speciation when they looked at closely related animals and noticed that the primary difference between many species is in male ornaments. Since these ornaments are usually produced by sexual selection, it seems likely that sexual selection has some role in creating new species. Dr. Rebecca Safran showed that in different populations of barn swallows, females prefer either males with longer tails or males with darker breast plumage. Since females have different preferences in these different populations, it is possible that the different populations will diverge into different species (as males gain longer tails in one population, and darker breasts in the other population).

Finally, we heard how male white-crowned sparrows use the colour of their crown as a signal of dominance. Males with more white on their crown are dominant over those with less white on their crown. These ‘status signals’ determine which males get the best territories and therefore get to mate. However, status signals are thought to be advantageous to all males because they can be used to resolve conflict without resorting to a full on fight, which can be extremely costly because the combatants could get injured. Status signals are common in males of many species, but behavioural biologists are still struggling to explain how these signalling systems remain honest. It is still somewhat unclear why subordinate individuals don’t grow large status signals in order to appear more dominant, but there are some emerging ideas of why this kind of ‘cheating’ may not be beneficial.

Sexual selection is still a field of intense interest for biologists even though Darwin first formalized the idea in 1871. It is likely the study of sexual selection will continue indefinitely because it produces some of the most interesting behaviours we observe in the natural world. Additionally, sexual selection is a dynamic process that is being affected by our changing environment and the tools available for the study of sexual selection are evolving at an astronomical rate. Indeed, it is truly an exciting time to study the sex lives of animals.

Photo of a Social Spider nest – from http://www.texasento.net

There have been a number of really cool talks at this years ABS conference. I’ll briefly describe a  few that have been particularly interesting to me:

Duck penis morphology:

You may have heard in the news recently about Dr. Patricia Brennan and her research on duck penis morphology. While I could spend a whole post justifying why she should receive NSF funding for this work, she explained it extremely well here. Dr. Brennans talk at ABS focused on plasticity (changes in an organisms characteristics in response to environment) in duck penis morphology. Some species of ducks have high rates of forced extra-pair copulations (rape), and these species typically have larger and more complex penis structures to more efficiently inseminate females. Dr. Brennan predicted that social context would influence the growth of penises in these species. She found that in the species with more forced extra-pair copulations there was increased plasticity – males grew larger penises to deal with increased competition for females when multiple males were housed together with females.

Personality in Founding Individuals:

In light of HIREC (Human Induced Rapid Environmental Change), some species are entering new territories. Dr. Jonathan Pruitt suggests that the personalities of these founding individuals is important for understanding the pattern and success of these invasions. Personalities, or behavioral syndromes, are repeatable patterns of individual behavior across context. Variation in personality has been found in an immense number of species. In social spiders, Dr. Pruitt has found variation in how bold and aggressive these individuals are. When an individual forms a new colony, they deal with other spider species that parasitize their immense nests. Aggressive individuals kill these parasites off, while more docile spiders tolerate the other spiders. While colonies founded by docile spiders are initially more successful, in the long run they are far more likely to collapse than those founded by aggressive spiders.

Agricultural Amoebas:

Humans aren’t the only species that tries to privatize resources. Dr. Joan Strassmann described a species of bacteria-eating amoebas in which some colonies farm several species of bacterium. Some are kept as a food source; while another, which they cannot eat is farmed for its chemical product which they use as a weapon against non-farming competitor colonies.

Look forward to a post in the next few days from Adam Reddon on the Agonism (fighting) session which occurred this afternoon!

Dr. Constance O’Connor is a postdoctoral fellow at McMaster University in Hamilton, Ontario, Canada. She studies the mechanisms underlying social behaviour in cichlid fishes, and is trying to understand how social behaviour evolved, and how environmental stressors affect social behaviour.

———————————–

When most people think of a behavioural biologist, I suspect that the image that comes to mind might be something along the lines of an earnest, scatter-brained David Attenborough with a clipboard. Clad in a Tilley hat and Wellington boots, clutching a clipboard and binoculars, your behavioural biologist can give an enthusiastic discourse on the mating habits of the sugar crab, but needs help finding his spectacles. In other words, I suspect that when most people think of a behavioural biologist, they don’t think of someone who is particularly practical, and the research is not usually considered particularly useful.

A growing group of behavioural biologists is trying to change this, and a special symposium at the 50^th Animal Behaviour Society meeting was dedicated to discussions of how behavioural research can be applied in practical, useful ways to guide wildlife management. Animal behaviour is a rapidly changing field, and behavioural biologists are now using new tools and new approaches to understand why some animal populations are declining, and how to reverse these alarming trends.

Traditionally, behavioural biologists have examined how evolution has shaped animal behaviour. Why do animals do what they do? In todays rapidly changing world, however, these behaviours that have been shaped by thousands or millions of years of natural selection are often no longer be adaptive. Newly hatched sea turtles head towards city lights instead of the ocean. Naïve island birds don’t recognize invasive cats as potential predators. Fish migrating upstream are stymied by a dam, and spend the rest of their days swimming into the concrete. Male beetles are so enchanted by beer bottles – the biggest, shiniest beetles they have ever seen! – that they no longer mate with female beetles. In todays changing world, behavioural biologists therefore need to research not just how evolution has shaped animal behaviour, but also understand and predict how animals will behave in a world with habitat alterations, pollution, invasive species, and a changing climate.

Through the symposium, there were positive examples of how behavioural biologists have already helped wildlife managers. For example, behavioural biologists have helped reduce the number of animals hit by vehicles, reduce the bycatch of dolphins and sea turtles in fisheries, and helped discourage bears from harassing visitors in National Parks. More importantly, there were discussions throughout the symposium about how animal behaviour research can be better used in the future to understand and solve conservation problems. The take-home message of the symposium was that behavioural biologists are already poised to contribute practical solutions to the current biodiversity crisis. The only thing left to do is for us, the behavioural biologists, to take up the challenge! We need to put on our Tilley hats and Wellington boots, and march off into our rapidly changing world to research both why animals do what they do, and how to make sure that animals can keep doing it for generations to come.

I’m a little embarrassed I haven’t posted anything since returning from Zambia (I’ll post a recap sometime soon), but Jenn, Adam, Connie and I are currently at the Animal Behavior Society meeting in Boulder, CO. This is one of the major annual meetings for our field, and consistently has a large number of very interesting talks.

In years past while at this conference, I’ve sent a daily e-mail to friends recapping some of the most interesting and unusual talks of the day. Scanning through the schedule, I can already tell there is going to be some really cool stuff. This year, these e-mails will appear on this blog, and I’ve also invited a few friends to contribute to the daily updates. 

Sometime tonight or tomorrow look forward to a post on today’s talks from Dr. Constance O’Connor who is currently doing a post-doc at McMaster University in Hamilton, Ontario.

 

Its been a long few months. 90 dives. 4.85 days underwater. A ton of experiments complete. A bad ass beard.

We completed our final dive this morning, and spent the day packing gear and dropping off equipment staying in Zambia at fisheries. The visibility has gotten better overall, but it wasn’t as great this morning as I’d hoped for a perfect final dive. Plus my observations didn’t go as well as I would have hoped. But its done.

Tomorrow we take the 18 hour bus ride back to Lusaka, after which we need to run some errands around town before departing for a bit of a holiday to cap off our trip. Rumor has it there are other interesting animals to see in Africa besides cichlids, so we are going on a safari at South Luangwa National Park. We’ll be staying at Flat Dogs (www.flatdogscamp.com), which is supposed to be one of the best in Zambia.

After the safari, we’ll be on our way back to North America and surely be perplexed by the lack of random children shouting “how are you?,” looking north to see weather that will arrive a day later, and stinking amazing sunsets over the ridge.