OT Covid-19 models
Locked
 
|
I received this link from a colleague who is a senior official at the
American Statistical Association https://www.jhsph.edu/covid-19/articles/10-tips-for-making-sense-of-covid-19-models-for-decision-making.html. ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
|
Thanks.
If you erase the particular article-name, you get an index with
links to dozens of other articles.
--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of Art Kendall <[hidden email]>
Sent: Thursday, April 30, 2020 11:31 AM To: [hidden email] <[hidden email]> Subject: OT Covid-19 models I received this link from a colleague who is a senior official at the
American Statistical Association https://www.jhsph.edu/covid-19/articles/10-tips-for-making-sense-of-covid-19-models-for-decision-making.html. ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
In reply to this post by Art Kendall
In 1976, George Box said "All models are wrong". Few years later, Box said "All models are wrong, but some are useful". If already available, where are now the useful COVID-19 models? Box, G. E. P. (1979), "Robustness in the strategy of scientific model building", in Launer, R. L.; Wilkinson, G. N. (eds.), Robustness in Statistics, Academic Press, pp. 201–236, doi:10.1016/B978-0-12-438150-6.50018-2 Johnny Amora On Thu, Apr 30, 2020 at 11:32 PM Art Kendall <[hidden email]> wrote: I received this link from a colleague who is a senior official at the |
|
|
Art and Johnny, thanks for those pointers at a time when public figures everywhere are showing uncritical
worship of one class of model and having little regard for the individual variables included and omitted Warm regards/gary From:
SPSSX(r) Discussion <[hidden email]> On Behalf Of
John Amora In 1976, George Box said "All models are wrong". Few years later, Box said
"All models are wrong, but some are useful". If already available, where are now the useful COVID-19 models? Box, G. E. P. (1976), "Science and statistics" , Journal of the American Statistical Association, 71 (356): 791–799, doi:10.1080/01621459.1976.10480949. Johnny Amora On Thu, Apr 30, 2020 at 11:32 PM Art Kendall <[hidden email]> wrote:
===================== To manage your subscription to
SPSSX-L, send a message to [hidden email] (not to
SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF
SPSSX-L For a list of commands to manage subscriptions, send the command INFO
REFCARD |
Administrator
|
In reply to this post by John Amora-2
I'm not responding to the question about which COVID-19 models are useful.
I'm just adding that Box made that same observation (in various forms) several other times. This Wiki page catalogs them nicely: https://en.wikipedia.org/wiki/All_models_are_wrong I quite like this version from a 1997 book: "It has been said that "all models are wrong but some models are useful." In other words, any model is at best a useful fiction—there never was, or ever will be, an exactly normal distribution or an exact linear relationship. Nevertheless, enormous progress has been made by entertaining such fictions and using them as approximations." John Amora-2 wrote > In 1976, George Box said *"All models are wrong"*. Few years later, Box > said *"All models are wrong, but some are useful".* If already available, > where are now the useful COVID-19 models? > > Box, G. E. P. (1976), "Science and statistics" , Journal of the American > Statistical Association, 71 (356): 791–799, > doi:10.1080/01621459.1976.10480949. > Box, G. E. P. (1979), "Robustness in the strategy of scientific model > building", in Launer, R. L.; Wilkinson, G. N. (eds.), Robustness in > Statistics, Academic Press, pp. 201–236, > doi:10.1016/B978-0-12-438150-6.50018-2 > > Johnny Amora > > > On Thu, Apr 30, 2020 at 11:32 PM Art Kendall < > Art@ > > wrote: > >> I received this link from a colleague who is a senior official at the >> American Statistical Association >> >> >> https://www.jhsph.edu/covid-19/articles/10-tips-for-making-sense-of-covid-19-models-for-decision-making.html. >> >> >> >> >> ----- >> Art Kendall >> Social Research Consultants >> -- >> Sent from: http://spssx-discussion.1045642.n5.nabble.com/ >> >> ===================== >> To manage your subscription to SPSSX-L, send a message to >> > LISTSERV@.UGA > (not to SPSSX-L), with no body text except the >> command. To leave the list, send the command >> SIGNOFF SPSSX-L >> For a list of commands to manage subscriptions, send the command >> INFO REFCARD >> > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD ----- -- Bruce Weaver [hidden email] http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." NOTE: My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
|
Reasoning by analogy (e.g., models) is, of course, something that needs to be
done with care. Models are certainly weakened by failure to get good input. But think of the hurricane models like the one that was amended with a sharpie. The plume is the useful information, e.g., the hurricane track will most likely be in the NW quadrant from where it is now, start westerly, and at some time bend North. That gives a rough idea of what will happen. Other infectious diseases have commonly behaved in these kinds of ways. This is an infectious disease. So far it has done such-and-so. This is (similar to) (analogous to) (modeled after) behavior of other infectious diseases A good guess is "It may behave in similar ways". ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
|
|
In reply to this post by John Amora-2
The first, basic insight for Covid-19 models is that uncontrolled
spread of the epidemic is /exponential/. Unfortunately, it takes
time for people to wrap their minds around the implications of
that, if they can do it at all. (I have found "relative humidity" to
be tough that way, too).
Second, for urban areas around the globe, the doubling-time
(with no social distancing, etc.) is similar, at 3 days or less. That
means that the log of cases or deaths can be plotted against time,
giving a straight line. Straight lines are nice for comparisons.
I like the graphs using "days since 50th case". Success in slowing
the spread is shown by the "flattening of the curve."
Third, there is a terrible delay of 5 or more days between infection/
ability to spread the disease and overt symptoms, and two more
weeks until death. For nursing homes, prisons, and meat-packing
plants (as we have observed, so far), that means that disease is rife
by the time of the first diagnosis, if (as in the U.S.) there has been
no aggressive program of testing.
--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of John Amora <[hidden email]>
Sent: Friday, May 1, 2020 1:36 AM To: [hidden email] <[hidden email]> Subject: Re: OT Covid-19 models In 1976, George Box said "All models are wrong". Few years later, Box said
"All models are wrong, but some are useful". If already available, where are now the useful COVID-19 models?
Box, G. E. P. (1979), "Robustness in the strategy of scientific model building", in Launer, R. L.; Wilkinson, G. N. (eds.), Robustness in Statistics, Academic Press, pp. 201–236, doi:10.1016/B978-0-12-438150-6.50018-2 Johnny Amora
On Thu, Apr 30, 2020 at 11:32 PM Art Kendall <[hidden email]> wrote:
I received this link from a colleague who is a senior official at the |
|
|
Also, many people have trouble dealing with large and small numbers. One of
Goffman's books maybe /Health Effects Of Ionizing Radiation/, has a great appendix on large and small numbers that I used to recommend for House and Senate staff and others who were going to deal with public policy. Another little trick that belongs in a class on quantitative thinking is to write the numbers down. In a conversation, something like this came up. "The pillow guy made 50 thousand masks. This was a good thing but not sufficient to reach the 3.5 billion needed." So I suggested writing the numbers down and stacking them. 50,000 3,500,000,000 This an example of people's difficulty with large and small numbers. Another time a local politician who is a financial planning advisor repeatedly said our town had 17 million people. It has 17 thousand. ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
|
Administrator
|
Art's mask example might be one of those rare cases where a pie chart is
helpful. As the following code shows, the pie is all one colour. NEW FILE. DATASET CLOSE ALL. DATA LIST LIST / x(F1) n. BEGIN DATA 0 50000 1 3499950000 END DATA. FORMATS n(F12.0). LIST. VARIABLE LABELS x "Masks". VALUE LABELS x 0 "Masks produced" 1 "Masks still needed" . WEIGHT by n. * Legacy pie chart. GRAPH /PIE=COUNT BY x. * Chart Builder pie chart. GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=x COUNT()[name="COUNT"] MISSING=LISTWISE REPORTMISSING=NO /GRAPHSPEC SOURCE=INLINE. BEGIN GPL SOURCE: s=userSource(id("graphdataset")) DATA: x=col(source(s), name("x"), unit.category()) DATA: COUNT=col(source(s), name("COUNT")) COORD: polar.theta(startAngle(0)) GUIDE: axis(dim(1), null()) GUIDE: legend(aesthetic(aesthetic.color.interior), label("Masks")) SCALE: linear(dim(1), dataMinimum(), dataMaximum()) ELEMENT: interval.stack(position(summary.percent(COUNT))), color.interior(x)) END GPL. Art Kendall wrote > --- snip --- > Another little trick that belongs in a class on quantitative thinking is > to > write the numbers down. > In a conversation, something like this came up. > "The pillow guy made 50 thousand masks. This was a good thing but not > sufficient to reach the 3.5 billion needed." > > So I suggested writing the numbers down and stacking them. > 50,000 > 3,500,000,000 ----- -- Bruce Weaver [hidden email] http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." NOTE: My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
|
In reply to this post by Bruce Weaver
Great point, Bruce. I also like Norbert Wiener...The best material model of a cat is another, or preferably the same, cat. On a related note, for any of you who've read The Signal and the Noise by Nate Silver, he talks about how difficult it is to get a good model in a dynamic system, where an event in one area changes events in seemingly unrelated parts of the system, and the system is constantly changing. Among the chapters he has to illustrate this is one on epidemics. Brian Brian G. Dates, M.A. Consultant in Program Evaluation, Research, and Statistics 248-229-2865 email:[hidden email] On Fri, May 1, 2020 at 11:37 AM Bruce Weaver <[hidden email]> wrote: I'm not responding to the question about which COVID-19 models are useful. |
|
|
One Star and Two Stars for the military, and senior people from CDC, NIH and
staff from both parties and both chambers agreed on the need for planning for spontaneous or deliberate chemical, biological, nuclear, or radiological events. (CBNR). The DPA was a major part in responding to CNNR events. It is over 30 years since I really looked at the DPA. However, I recall some main take aways. (1) emergency needs could be met. (2) producers could be sure they would be paid even if they over-produce (3) price-gouging would be reduced. (4) producers would be lent the money for facilities, new staff, and materials. The need for data as input to modeling for monitoring and evaluation was well recognized. Over the years things like these were in the briefings. They should not be surprises. hand sanitizer disinfectants things needed for whatever testing was appropriate contact tracers places for people to self-isolate when they should not go home PPE for ICUs, EDs, ambulance drivers, nurses, physicians, hospital janitors, testers, food workers, grocery clerks, delivery people ... barbers, etc. PPE for anybody who leaves his/her house body bags cremation facilities ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |