A preview of my weekly read-around for Equitable Growth... 1) This is an excellent choice. This will be great fun. And we really do have a shot at changing the world, albeit in a small way: Equitable Growth: Names Economist Michelle Holder as New CEO: ‘Holder will take the reins at 7-year-old nonprofit research and grantmaking organization dedicated to accelerating research on how economic inequality affects economic growth and stability…
It seems possible maybe even probable to me, that SARS-2 may have in patient zero, been less contagious than the one that spread in late December 2019. That would account for the reported early cases, which if real present difficulties. I've read the virus was already well adapted to humans. Perhaps the first crossover had a very modest R value close to one, and it simmered along until it hit upon a lucky mutation?
Re: source of Covid outbreak. I agree with your reasoning. The NYTimes article is just speculative BS. This doesn't mean it couldn't have been a lab leak, but the probability is low. China has been its own worst enemy here, by covering up its early actions, suppressing evidence and warnings early on, possibly influencing the WHO, and certainly restricting full access to records. As they say, the coverup is often worse than the crime.
But this all deflects from the current problem - getting the globe vaccinated to reduce the petri dish of variant production, overcoming vaccine resistance (even from Covidiot Texam H/C workers), and preventing the next surges. The UK is facing a new surge in its zeal to open up the economy. The US is following suit.
My fear is that these responses will play into the hands of those who say we should have kept the economy open and let a fraction of the population die off and to hell with overwhelmed hospitals and untreated patients who are left to die at home or in the streets. We were truly lucky to get extremely effective vaccines so quickly (they certainly surprised me by how quickly they arrived). Had mask-wearing been initiated universally in each country from the start, even without social distancing the mortality rate would likely have been lower. Add in quick vaccine development and we would have been a lot better off. What we need now is adequate stockpiles of basic equipment, masks, PPE for health workers and essential workers, and a protocol with enforcement to ensure best practices are used and enforced. Whether we do this properly or play the same game as gun control is a political toss-up.
Re: Covid variants. I am starting to lose empathy for those who resist getting vaccinated without a good medical reason. Maybe there will be Darwinian [un]natural selection to remove these people from the gene pool. Shame they have already reached reproductive age so that they can infect others with their cultural memes. It seems to me that there are going to be populations who have no access to Medicaid and are going to be severely financially compromised if they end up in an ITU. "Red state" populations, especially may be affected. One would think their legislators might be worried about this unless they hope this will affect their Democratic Party voters more severely. [Or they cynically hope it will reduce the welfare rolls in toto.]
Come winter and flu season, I am hoping California ensure an adequate supply of vaccines for regular flu and the new, more transmissible, Covid variants. I am certainly going to wear a mask in places where I am in contact with others. This is one behavior change courtesy of S. E. Asia that I am thankful for.
Given the preponderance of data-mining results to write positive journal papers, and the known failure to replicate papers where serious money might be involved is it any wonder this goes on. Journals expand in number to reap monies from those needing to publish, even garbage, polluting the knowledge-space. Are academic institutions also to blame if they have publish or perish rules? Sadly, much research is rarely more useful than a blog post. The trash that ends up in arxiv and bioarxiv as placeholders is becoming a problem. I recall the immense amount of time needed to track references in university journal stacks in the 1970s. The internet with search seemed to be a savior. But now the sheer number of papers that need to just have their abstracts read is getting ever longer.
There also needs to be some way to publish negative results. That might increase honesty and reduce P=hacking. It would mainly solve the problem of repeatedly doing the same experiment as there was no way to tell it was done before and failed.
I an of the opinion that p-values have limited use, especially in a world where data is easy to come by. Tiny effects can give you very low p-values with enough data. Better to look at the effect size coupled with a p-value. The other issue with p-values is the selection of the wrong statistical test. I recall a study that showed that at least 30% of examined journal papers used the wrong test. Software often lets you use a selection of tests which allows the user to select the test that provides the best p-value, even if the test is inappropriate. I gather most reviewers rarely comment on statistical tests, possibly because of a lack of knowledge (especially true in biosciences). Maybe peer review would be better if there were fewer papers to review, or some of the publisher's profits paid for a review of the data and statistical tests used?
"If we can hold Facebook’s algorithms to account, why can’t we do the same for our government?"
We know the reason for this. The wealthy buy the legislators to make the rules suit them. How many more books need to be written about this? It isn't just legislators but corporations. High-Frequency Trading only works because the NYSE sells access to front-run trades. Once total illegal, access is now sold. Corporations even {illegally} backdate options to be most favorable to the CEO. Banks continue to time charges on foreign transactions at the rate that maximizes their returns. Robocalls claiming debts owed made up of whole cloth continue, as well as the innumerable scams (like the current renew car warranty) and the carriers slow-walking blocking these calls. Liz Warren has plenty of posts on fixing the many predatory practices but nothing gets done. I wonder why?
re: New Report Lays Out the Path Forward for Faltering U.S. Unemployment Insurance System
That would be nice, but over my lifetime in the UK and US, the opposite has happened.
In the UK up to the Thatcher era, one just went to the unemployment office on the day stipulated by the first letter of your last name. They noted you were not working and a payment was made in some way. I recall seeing actor Patrick Troughton there too, no doubt in "resting" mode.
When Thatcher came into power, the focus was on supposed cheaters. It was harder to register and the state did surveillance to find cheaters.
Now with Universal Credit I gather that it is a disaster, resulting in people dropped through the cracks, the disabled cut off, and reported suicide. This seems to have happened with a similarly hostile home office making passport renewal for expats much harder.
My experience with California's unemployment system is better, but the people they use to sort out their own errors make Comcast customer service look good. I was lucky to get a state representative to fix a problem caused by the office after being lied to by the customer support people over my claim for reimbursement. I suspect the same mentality of assuming everyone cheats has infected the system.
The same has occurred with disability payments. It is now routine to deny disability unless several attempts are made to claim disability. You now need lawyers to plead your case and they take 30% of any back payments owed. It is scandalous.
So yes I would like to see a smoother system for those needing the systems, but I fear those making the rules and laws have no comprehension of the true situations and assume those needing help are just cheating the system. (Reagan's "Welfare Queen" meme just will not go away.)
It seems possible maybe even probable to me, that SARS-2 may have in patient zero, been less contagious than the one that spread in late December 2019. That would account for the reported early cases, which if real present difficulties. I've read the virus was already well adapted to humans. Perhaps the first crossover had a very modest R value close to one, and it simmered along until it hit upon a lucky mutation?
Re: source of Covid outbreak. I agree with your reasoning. The NYTimes article is just speculative BS. This doesn't mean it couldn't have been a lab leak, but the probability is low. China has been its own worst enemy here, by covering up its early actions, suppressing evidence and warnings early on, possibly influencing the WHO, and certainly restricting full access to records. As they say, the coverup is often worse than the crime.
But this all deflects from the current problem - getting the globe vaccinated to reduce the petri dish of variant production, overcoming vaccine resistance (even from Covidiot Texam H/C workers), and preventing the next surges. The UK is facing a new surge in its zeal to open up the economy. The US is following suit.
My fear is that these responses will play into the hands of those who say we should have kept the economy open and let a fraction of the population die off and to hell with overwhelmed hospitals and untreated patients who are left to die at home or in the streets. We were truly lucky to get extremely effective vaccines so quickly (they certainly surprised me by how quickly they arrived). Had mask-wearing been initiated universally in each country from the start, even without social distancing the mortality rate would likely have been lower. Add in quick vaccine development and we would have been a lot better off. What we need now is adequate stockpiles of basic equipment, masks, PPE for health workers and essential workers, and a protocol with enforcement to ensure best practices are used and enforced. Whether we do this properly or play the same game as gun control is a political toss-up.
Re: Covid variants. I am starting to lose empathy for those who resist getting vaccinated without a good medical reason. Maybe there will be Darwinian [un]natural selection to remove these people from the gene pool. Shame they have already reached reproductive age so that they can infect others with their cultural memes. It seems to me that there are going to be populations who have no access to Medicaid and are going to be severely financially compromised if they end up in an ITU. "Red state" populations, especially may be affected. One would think their legislators might be worried about this unless they hope this will affect their Democratic Party voters more severely. [Or they cynically hope it will reduce the welfare rolls in toto.]
Come winter and flu season, I am hoping California ensure an adequate supply of vaccines for regular flu and the new, more transmissible, Covid variants. I am certainly going to wear a mask in places where I am in contact with others. This is one behavior change courtesy of S. E. Asia that I am thankful for.
Re: P-hacking
Given the preponderance of data-mining results to write positive journal papers, and the known failure to replicate papers where serious money might be involved is it any wonder this goes on. Journals expand in number to reap monies from those needing to publish, even garbage, polluting the knowledge-space. Are academic institutions also to blame if they have publish or perish rules? Sadly, much research is rarely more useful than a blog post. The trash that ends up in arxiv and bioarxiv as placeholders is becoming a problem. I recall the immense amount of time needed to track references in university journal stacks in the 1970s. The internet with search seemed to be a savior. But now the sheer number of papers that need to just have their abstracts read is getting ever longer.
There also needs to be some way to publish negative results. That might increase honesty and reduce P=hacking. It would mainly solve the problem of repeatedly doing the same experiment as there was no way to tell it was done before and failed.
I an of the opinion that p-values have limited use, especially in a world where data is easy to come by. Tiny effects can give you very low p-values with enough data. Better to look at the effect size coupled with a p-value. The other issue with p-values is the selection of the wrong statistical test. I recall a study that showed that at least 30% of examined journal papers used the wrong test. Software often lets you use a selection of tests which allows the user to select the test that provides the best p-value, even if the test is inappropriate. I gather most reviewers rarely comment on statistical tests, possibly because of a lack of knowledge (especially true in biosciences). Maybe peer review would be better if there were fewer papers to review, or some of the publisher's profits paid for a review of the data and statistical tests used?
re: Tim O'Reilly's pos
"If we can hold Facebook’s algorithms to account, why can’t we do the same for our government?"
We know the reason for this. The wealthy buy the legislators to make the rules suit them. How many more books need to be written about this? It isn't just legislators but corporations. High-Frequency Trading only works because the NYSE sells access to front-run trades. Once total illegal, access is now sold. Corporations even {illegally} backdate options to be most favorable to the CEO. Banks continue to time charges on foreign transactions at the rate that maximizes their returns. Robocalls claiming debts owed made up of whole cloth continue, as well as the innumerable scams (like the current renew car warranty) and the carriers slow-walking blocking these calls. Liz Warren has plenty of posts on fixing the many predatory practices but nothing gets done. I wonder why?
re: New Report Lays Out the Path Forward for Faltering U.S. Unemployment Insurance System
That would be nice, but over my lifetime in the UK and US, the opposite has happened.
In the UK up to the Thatcher era, one just went to the unemployment office on the day stipulated by the first letter of your last name. They noted you were not working and a payment was made in some way. I recall seeing actor Patrick Troughton there too, no doubt in "resting" mode.
When Thatcher came into power, the focus was on supposed cheaters. It was harder to register and the state did surveillance to find cheaters.
Now with Universal Credit I gather that it is a disaster, resulting in people dropped through the cracks, the disabled cut off, and reported suicide. This seems to have happened with a similarly hostile home office making passport renewal for expats much harder.
My experience with California's unemployment system is better, but the people they use to sort out their own errors make Comcast customer service look good. I was lucky to get a state representative to fix a problem caused by the office after being lied to by the customer support people over my claim for reimbursement. I suspect the same mentality of assuming everyone cheats has infected the system.
The same has occurred with disability payments. It is now routine to deny disability unless several attempts are made to claim disability. You now need lawyers to plead your case and they take 30% of any back payments owed. It is scandalous.
So yes I would like to see a smoother system for those needing the systems, but I fear those making the rules and laws have no comprehension of the true situations and assume those needing help are just cheating the system. (Reagan's "Welfare Queen" meme just will not go away.)