0:00

Hello, this is Matthew Jackson from Stanford University, and this is our

Â first lecture in social and economic networks.

Â And I'm going to start with just a brief introduction of some of the material

Â we'll be covering. And, of course, the most important

Â question we can start with is, why study networks?

Â And I think from a social scientist's perspective, it's that many economic,

Â political, and social interactions are embedded in social settings, and the

Â structure of those relationships is very important in both determining how people

Â behave and, and determining what outcomes are.

Â So things like trades and good and services, most markets are actually not

Â centralized but, but occur between different parties and bilateral

Â relationships. Sharing of information, sharing a favor,

Â risk, transmission of viruses, opinion. How did you find out about a job?

Â often through somebody you knew, you know.

Â how do you choose who you vote for? How do you make decisions about products?

Â A lot of time you're talking to different individuals, what did they hear?

Â how did you hear about your information? Political alliances can be represented in

Â these networks, trade alliances, there's all kinds of, of different settings where

Â network structure is very important. And critical to this is the fact that the

Â networks actually influence the behavior. So if we look at crime, we look at

Â employment, we look at people's investment in human capital, education.

Â how they vote, whether they smoke all kinds of decisions they make are, are

Â embedded in these settings and are influenced by the social structure.

Â So in, in most important from our perspective, is that networks come in

Â different sizes, and understanding how they're shaped, what they look like is

Â going to very important in understanding what the outcomes are.

Â And so there's, a lot, to understand and to model.

Â So,[COUGH], the primary questions we're gonn be interested in, in this course,

Â will start with just some background on, on structure of social networks.

Â What we know about social networks. we'll must of the course we'll be looking

Â at how networks form, so do the right ones form?

Â If we could actually affect the network formation from different perspectives,

Â would we want to? How do networks influence behavior?

Â So how, how what's the relationship between how dense a network is and what

Â the outcomes are? And so forth.

Â now this is an area, obviously, which has been researched in, in many different

Â disciplines. And so when we look at the literature's,

Â there's sociology, economics, computer science, statistical physics,

Â mathematics, random graph theory, and so what we'll do her is try and synthesize

Â these and bring together a random view point.

Â And pulling models from different perspectives.

Â And trying to understand what we've learned and also what are important areas

Â for future research as we go along. Now, in terms of areas for future

Â research and current research. We'll be looking at, at theoretical

Â foundations, modeling basically of, of network formation, modeling of dynamics.

Â design of networks, how then understanding how networks influence

Â behavior. There will be some emphasis at the, the

Â end of the course on co-evolution, so that, what does that mean?

Â That means that, that who my friends are, influences my behavior, but my behavior

Â also influences who my friends are, so there's a co-determination.

Â It's not as if one is set in stone and affects the other, but both evolve in

Â together. We'll be, be looking at, at a lot of

Â empirical work and experimental work as we go along.

Â Observing networks, seeing what patterns we do see and, and also you know with an

Â emphasis on testing theory and understanding regularities and patterns

Â that are out there in the data. one other thing that we'll see as the

Â course goes along is methodology. So they'll be a whole series of

Â definitions about networks. for instance, you know, understanding

Â who's central in a, in a network could be measured many different ways.

Â And we'll try and say something about what are good and bad ways of, of doing

Â it for different applications. So there might not be a single way to

Â approach a problem, but understanding what are the different methods and, and

Â is there something we can say about the methods themselves.

Â 4:20

central focus in this course is really going to be on the models and the types

Â of, of techniques we'll be using are one, pulled from random graph theory, pulled

Â from mathematics. The other, we'll be using some strategic

Â and game theoretic techniques, and we'll also be using some hybrid models that

Â involve some, both choice and chance. And looking at some statistical models

Â for fitting and analyzing networks, dealing with data.

Â goals, I'm not going to presume, prior knowledge of network analysis.

Â I'm going to try to introduce you to variety of different approaches, so the

Â idea here is really breadth, more than depth, so its an idea of giving you, some

Â exposure, so you know what's out there. The types of different tools, which tools

Â might be appropriate in different settings.

Â There's a lot more that can be said about each of the subjects we're going to talk

Â about, but this will be more or less an introduction, to give you an idea of

Â exactly what the tools are that might be appropriate for different parts of

Â analysis. It'll also give you some sense of

Â different disciplines' techniques and, what the kinds of questions and

Â perspectives that they take. In terms of, one important aspect when I

Â start the course here is, is really emphasis.

Â Why do we care about modelling things to begin with?

Â And I think this is an important question that, that will shape the structure of,

Â of what kinds of models we work with and, and how they're formed.

Â And, you know, when we look at models, one thing they do for us is give us

Â perspective into why we see certain things.

Â So why do social networks have short average path lengths for instance.

Â Why is it that there's six degrees of seperation in the world.

Â Well, we'll see an answer to that, that will come out of random graph model.

Â So, just understanding the structure of how things arise at random, can help us

Â understand why we might see something like that.

Â So understanding a basic tree structure that underlies social networks, will help

Â us understand path length. models also about to compare the

Â statistics. So if we understand that models changfe

Â as we change different parameters, that can help us make predictions about how

Â the world might change. So, how, how does the component structure

Â change with density. If a, if a network has more and more

Â links, what does that do to the overall component structure of the network.

Â It will help us make predictions out of samples, so if you want us to come in

Â with a new policy for instance you are trying to, to stamp out a flu, epidemic.

Â how effective does the vaccine have to be in order to, to limit, the extent of a,

Â the epidemic. That's a question we can begin to answer

Â with network analysis. things will also the models will allow

Â for statistical estimation. So, if we wana understand, for instance,

Â is their significant clustering which means, you know, are my friends friends

Â with each other. does that happen, because of some social

Â force, or is it happening just at random? we can test models.

Â So we can take models and, and then ask does this appear that this happened at

Â random, or does it appear that it something else was going on.

Â So there'll be statistical tests that we can use once we have models for analyzing

Â that kind of question. in terms of a basic outline of the

Â course, it's going to break into three parts.

Â The first part's going to be background and fundamentals.

Â So, definitions. How do we analyze networks?

Â What are some basic, properties of networks, characteristics.

Â And along with this will be empirical background.

Â The second part of the course, and the central part of the course is going to be

Â network formation models. So we'll look at random graph models, and

Â then we'll also look at strategic formation models when people are actually

Â making choices. the third part of the course is networks

Â and behavior. That's going to then take networks and

Â understand how the shape of networks and the structure of networks,

Â Who do you know, how many people do you know, who do they know and so forth.

Â How does that influejnce what you're dedcisions are, your behavior and so

Â forth. So we'll look at things like diffusion

Â and contagions. We'll look at learning models.

Â And then finally, what's known as games on network or situations where what I do

Â depends on the choice of my friends. So if there's a new app out there, do I

Â want to get it? Well it might depend on how many of my

Â friends get it and that might depend on how many of their friends get it, and so

Â forth. And so how do we analyze that in a

Â network context. So more or less these three main parts

Â are going to be the core structure of the course.

Â And there's a text book which is completely optional, that I've written

Â where a lot of the material is going to be pulled for.

Â In terms of this outline, the numbers on the side here indicate the chapters, so

Â one two, three four five and so forth, these indicate the relative chapters out

Â of the book that, that correspond to the lecture structure of the course.

Â So we'll be moving along through the book with, with a, a couple of exceptions in

Â terms of which chapters are covered in which part.

Â So that's the basic outline. And so let's get started.

Â