Claims about networking and social distancing

Our models demonstrate that while social distancing measures clearly do flatten the curve, strategic reduction of contact can strongly increase their efficiency, introducing the possibility of allowing some social contact while keeping risks low. Limiting interaction to a few repeated contacts emerges as the most effective strategy. Maintaining similarity across contacts and the strengthening of communities via triadic strategies are also highly effective. This approach provides empirical evidence which adds nuanced policy advice for effective social distancing that can mitigate adverse consequences of social isolation.

That is from a new paper by Per Block,  I do not consider this a confirmed result, but it is consistent with how my intuitions have been developing, and the success in containing Covid-19 on various smallish islands.


Well, sure. Your network is your network plus fan-out. If you and your contacts bias toward repeated interaction you reduce both your primary group, and the fan-out.

Of course, supermarket checkers might find themselves dropped from such groups. I've heard it suggested that all markets should do online orders and curb pickup for this reason.

I'd support that.

Also exclude diehard runners and bicyclists. They are nuts to be doing it, as I saw yesterday, in Los Angeles on April 21st, 2020.

Or, monoamory is in, polyamory is out.

Re: Also exclude diehard runners and bicyclists.

Why? Assuming people are doing these things solo, or just with other household members, they are not at any unusual risk.

Maybe that was the theory, but I see them congregating. Perhaps by force of habit they run the same park-loop they always did. Given no "one way" rule maybe 10 close oncoming interactions per minute?

Clockwise people and counterclockwise people.

Exclude men you know are irate about the masks thing. That attitude, i suspect, correlates highly with men who don't wash their hands. Gallup did a poll a few years back. In some countries half the men wash their hands. In others, italy, france, spain, netherlands, nearly none do.

While it's nice that this model includes network structure and moves beyond the binary choice of zero or full restrictions, it would be even better if we developed models with network structures that directly mirror the real world. I elaborate on this here:

I agree with you, but I wanted to bring attention to this quote from your blog post: "I’d like to imagine that like the Manhattan Project, there is a sequestered band of scientists somewhere doing the difficult work of slogging through demographic, transport, and city planning data to construct realistic pandemic models, and that they’re so occupied with this real work that they aren’t wasting time Tweeting. I am skeptical they exist, though. They should, however — it would a good resource for the next pandemic."

That band of scientists almost certainly does not exist. You have thought about this problem a lot. Even if you aren't a scientist, why don't you launch that project yourself? The problem is that everyone assumes that someone else is doing it, when no one is.

Thank you! Your question, "why don't you launch that project yourself?" is a very good one. I don't have a good answer, though. Here are three partial answers:
1. There are many people more skilled than I am at data analysis, modeling, and related topics who are also not working on pandemic modeling; with my magic wand I'd grab them, not me.
2. I already have a lot of work to do, running my research lab (currently focused on data analysis and writing and other non-lab things), teaching, etc. (I'm a professor, by the way.) Even sparing the time to write that post was probably not a great idea from the perspective of my actual job. One could argue that my normal work is less important, but...
3. ... this statement, and the means to draft and organize people to do important work, should be coming from NIH, NSF, etc. It's not, which is another puzzle. It's not unique to Covid-19, though. Even before all this, I've complained ( that these organizations are curiously uninterested in actually changing the structure of the scientific enterprise. Oversimplifying: we act as if we don't really care about tackling, or answering, big questions.

These answers are probably not too satisfying, even to me!

Thank you for responding. You are still uniquely positioned to do something, if you have the will. It doesn’t take nearly as much time as building the model ——> use your professional network to ask around and find someone who can spearhead such a project, and offer a promise of some kind of support, leveraging your university, other people in your network, companies that can provide funding, etc....

This is not new. Which is why if you interact with some folks who have few or no in or outlinks, and are in the same community as you are and your community has few cases , the probability of getting something is low, unless the community you are in is hot and they have many in and outlinks. Low density. Low degree.

Also, its not social distance, it's physical distance. Big difference.

Compare that to a prison, where prisoners eat communally, have many physical in and outlinks to each other, and to the guards who come in from outside and move back into the community.

The more interesting network issue is connections between firms who are dependent on each other for inputs and sales, and each firm is in a different state with a different closure program. Bound to have bottlenecks in the chain, particularly if some states go hot and have to close at different times.

So, if two months from now Georgia is a hot spot, and I depend on it for an input or product, and they are shut down entirely, even though my workforce is healthy and there is low transmission in my community, my assembly or distribution network will come down.

You might want to read Prof Matthew Jackson's network books and look at this recent article:

Early on it was obvious events and places with international appeal—NYC and Mardi Gras—were spreaders and not events with no international appeal like a Trump rally in SC on March 1st.

Outsiders. That's been the topic of many Cowen blog posts. Don't let outsiders in, whether in the country, in the state, in the city, or in the working group. We are all immigrants, now!

Not a criticism, just an acknowledgment of reality, and the future. How about everyone wear a badge that reveals where she is from, where she has been, and where she is going. That will tell where she can and cannot go.

Besides a badge, wear a mask. Finally, a clear and concise explanation of their use:

1. I don’t understand how this is not obvious.

2. I don’t understand how this is not what is already occurring. I don’t know a single person who completely isolates themselves... everyone is currently “limiting interaction to a few repeated contacts.”

Maybe I’m dumb and someone can explain. However, what I’ve almost concluded by reading MR posts over the past several months is that the intuition of the lay-people typically is just what these epidemiological studies find. It’s just that the intuition of the lay people comes faster. This is not usually the case when it comes to economics, which is why I like this blog... but honestly I’m not sure I’m being made any smarter about the coronavirus by reading about it here (or anywhere).

Since this started it's often been clear that the optimal policies in a time of contagion resemble those of more conservative attitudes on social mixing (or of people from more traditional societies, or people with high social anxiety) - repeatedly socialize with a few people who you know well, be suspicious of people who spend too much time outside the group, and avoid going to new places. This is not an argument for social conservatism, but it shows how these attitudes make more sense under certain conditions than others.

Abstract Tree.
They want to partition society such that an infection in any one partition is equally surprising, make them independent events.

He gets a Planck's curve, a probability of emission per number of steps in a loop on the graph. But he assumes triadic intervention rates are not overloaded. Since infected people are no longer a problem, the infections are emissions from a black box that is cooling off.

Change the problem, let people have some immunity rats, and they go back to infected. There are no phases of matter given, they must be constructed. The graph is bandwidth limited. The group size anywhere will be uncertain, but in total must add to the original size. The graph becomes a three coloring problem, infected groups, immune groups, and triadic intervention. You have a beach ball, you must paste it with three colored buttons and make it look as white as possible. We know that solution because we have a general method for defining any N colored beach ball.

For any number of colors, the coloring algorithm will consumes qubits of color at each move, always making the surface more white than it was. If we know the algorithm then we know the arrival rates of the qubits. We are not flat, we pack sphere. But If I know the arrival rates for one surface then I can derive the arrival rates for the surface just interior. That surface must be equally white., with lower resolution.

My arrival rates must be primes, else one of my dimensions is just a multiple of another. At some point I do not have enough primes to compute the spots on the outer surface as the absolute count available is M factorial.. At some point in the interior my arrival rates go to the value one, zero is not allowed. We know how to determine the size of your beach ball, its Avagadro, in any dimension. The AVagadro is the total number of color unit that must be applied per dimension. We have a systematic way of reducing a factorial graph into a congested managed directed graph with the minimum number of deviations. Those deviations are the coloring algorithm, a very small finite set.

Well yeah, the behavior of people, in terms of both networking and physical distancing, will affect the spread of the virus.

And we could lift some parts of the lockdown more quickly if we (a) could identify the people who are less tied in to the contagious networks and (b) if we could control people's behavior and get them to stop doing the risky networking and do more of the non-risky networking.

But I'm not sure that we have the state capacity (including the population's willingness to acquiesce to authority) to control behavior like that, beyond gross controls such as "stay home".

IOW, people's networks and behavior can change, rapidly. Maybe we try to tell people: living through the lockdown is easier if you get outside and get some exercise while being careful to maintain physical distance. E.g. I had been going to some local parks and a local high school track to do some hiking or just walking.

But people's behavior at those locations has changed: I see more and more people at what had been largely empty spaces. So much so that I've had to change where I walk, because previously under-used parks are now getting more use than average.

People can and do change their behavior and networks, so attempts to limit the virus' spread by exploiting those networks may not have the intended effects. People might react to the closing of Portland's popular city parks by going in larger numbers than ever to the less-popular ones, rendering them more contagious than the old data would've predicted.

one tricky part is "nuance". The media doesn't handle that very well.

Comments for this post are closed