small typo, tweaks

[martinamorris]

typo: Maybe this will create much less connectivity than?

tweaks:

we won’t just set an average of one non-essential tie per household; we’ll make it the maximum—every household has exactly one.

we won’t just set an average of one non-essential tie per household; we’ll make every household have exactly one.

largest component -> largest connected component (but see below)

bullet point these:

• They may be multiple degrees of separation away from you.
• They may be your friend’s mother’s patient’s father.
• They may be your sister’s friend’s grandfather’s doctor.
• They may be the daughter of the grocery store worker where your friend’s parents shop.

Another word for component?

Might be worth using reachability -- how many people are able to reach you, how many people you are able to reach.

[sgoodreau]

Thanks - all but the "component" points processed in a straightforward way.

For component , I agree that's a weird, jargony word, and I've added in the concept of reachability explicitly to explain it. But I've still left it in because there isn't another good noun to describe the actual structure that I can think of. I also put it in quotes the first time I use it to emphasize that it's a new term. And I added "connected", even though technically I believe it's redundant (that's the definition of a component) because it emphasizes the most important feature in familiar language..

[martinamorris]

For component , I agree that's a weird, jargony word, and I've added in the concept of reachability explicitly to explain it. But I've still left it in because there isn't another good noun to describe the actual structure that I can think of. I also put it in quotes the first time I use it to emphasize that it's a new term. And I added "connected", even though technically I believe it's redundant (that's the definition of a component) because it emphasizes the most important feature in familiar language..

the tension between technically accurate and effective communication to a lay public is inherent here.

i'd suggest removing "component" entirely, replace it with "cluster" and focus on the words/concepts:

"reachable" -- Now your household can reach 60% of the households in your community, and they can reach you! You may not even know the people in these households, but the virus can easily travel along the network of connections to get from them to you. It's a very connected, and effective network of transmission

"x steps of separation" -- In this central connected cluster, households are just x steps away from each other (on average).

Then with the tabs, you can get technical.

[sgoodreau]

"Cluster" as a vernacular network term has a clear sense to me of much denser connections within a small group of people. It also has a technical network meaning that doesn't match this. So that's still not right.

We do need a specific word or phrase given how much we refer to it, and specifically measure it. I'm happy for you to keep making suggestions, but I'll stick with component for now. I guess I don't think that having one technical term that is clearly defined is too worrisome.

I like the reachable language, and I'll add that.

Calculating average distance within the largest component is more than I have time for right now (and does sna even do that?) but the "x steps of separation" language is good, and I'll add that somewhere.

[sgoodreau]

OK I went to add degrees of separation, and then realized I needed to define it, and ended up with this:

You may have heard of the concept of "degrees of separation" - how many steps you have to go to get from one person to another. This is usually talked about in terms of the relationship of "knowing" or "being friends with" someone, but we can use that idea here too, in terms our relationship of interest---hanging out with. You can see that this idea is closely linked with these paths we are talking about. For example, households that have three degrees of separation are those that are connected by at least one 3-path, but not by any shorter paths. The smaller the degrees of separation between two households, the easier it is for the virus to get from one to the other.

In the end, I think this actually ends up just adding more complexity instead of making things more intuitive. Largely because the relationship between path lengths and degrees of separation is trickier than it seems.

In retrospect, it might have made sense to go with average distance as the metric considered instead of path counts, but my understanding is that those algorithms are much much slower, and I think in the end the visuals drive the message anyway - the rest is just filler. So I'll leave this bit as it was. If you think the paragraph above about degrees of separation is useful, I can put it in after all.

[martinamorris]

On Tue, 7 Apr 2020, Steven Goodreau wrote:

"Cluster" as a vernacular network term has a clear sense to me of much denser connections within a small group of people. It also has a technical network meaning that doesn't match this. So that's still not right.

i'm not sure that's true. you might ask some other folks, but to me the word "cluster" just means connected, not densely connected. in the sna literature, the word is clique/clan. in the physics network literature the term is used quite loosely.

i think, for the general public, it's an appropriate and effective term.

We do need a specific word or phrase given how much we refer to it, and specifically measure it. I'm happy for you to keep making suggestions, but I'll stick with component for now. I guess I don't think that havening one technical term that is clearly defined is that problematic.

i just know from teaching the "Know Your Network" stuff that the word component, even when defined, is a stumbling block

Calculating average distance within the largest component is more than I have time for right now (and does sna even do that?) but the "x steps of separation" language is good, and I'll add that somewhere.

yes, the geodesic distance distn

[martinamorris]

honestly, i don't even understand what the path counts mean.

again, i think you need to balance the public accessibility with our obsession with scientific accuracy. i honestly think everyone understands the idea of 6 degrees of separation. you don't need to define it here. i'm guessing the average geodesic distance in the cluster may be < 6. but that's an empirical question, and (i'm sure) depends on net size.

and use the tabs for more extended text/definitions. it'll take the pressure off.

You may have heard of the concept of "degrees of separation" - how many steps you have to go to get from one person to another. This is usually talked about in terms of the relationship of "knowing" or "being friends with" someone, but we can use that idea here too, in terms our relationship of interest---hanging out with. You can see that this idea is closely linked with these paths we are talking about. For example, households that have three degrees of separation are those that are connected by at least one 3-path, but not by any shorter paths. The smaller the degrees of separation between two households, the easier it is for the virus to get from one to the other.

In the end, I think this actually ends up just adding more complexity instead of making things more intuitive. Largely because the relationship between path lengths and degrees of separation is trickier than it seems.

In retrospect, it might have made sense to go with average distance as the metric considered instead of path counts, but my understanding is that those algorithms are much much slower, and I think in the end the visuals drive the message anyway - the rest is just filler. So I'll leave this bit as it was. If you think the paragraph above about degrees of separation is useful, I can put it in after all.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, orunsubscribe.[AB6QTYSLKQAAES3SEELC4VTRLNVLHA5CNFSM4MCYKGT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOERS C6XI.gif]

---

Professor Emerita of Sociology and Statistics Box 354322 University of Washington Seattle, WA 98195-4322

Office: (206) 685-3402 Dept Office: (206) 543-5882, 543-7237 Fax: (206) 685-7419

morrism@u.washington.edu http://faculty.washington.edu/morrism/