Coronavirus - accurate information? what you believe ?

[mouse]

5 hours ago, DavyJonesLocker said:

Same website is showing around 220 today

 

The website (windy.com) isn't a record of historic emissions, it predicts emissions. Even if Wuhan burned a billion people tomorrow this website wouldn't show it, as it doesn't collect data in that way.

[889]

Alex Lam, a journalist with Apple Daily, reports that China has changed its definition of "confirmed case" in the past few days: those who test positive but have no symptoms are no longer considered a "confirmed case."

 

http://twitter.com/lwcalex/status/1226840055869632512

 

A reminder that when dealing with statistics, definitions count as much as abstract "accuracy."

[imron]

On 2/11/2020 at 10:38 AM, roddy said:

Under-reporting? Absolutely. But it's not like there are two sets of books here. 

This seems relevant: https://www.barrons.com/articles/chinas-economic-data-have-always-raised-questions-its-coronavirus-numbers-do-too-51581622840

[realmayo]

I'm obviously no expert, but perhaps be sceptical of some scepticism! 

 

Statistician quoted in below piece says "I have never in my years seen an r-squared of 0.99". If we fit exponential curve to Ebola cumulative case data from 2014, we get r-squared=0.992. We should be cautious about interpreting #COVID19 data in real-time, but for right reasons

 

https://twitter.com/AdamJKucharski/status/1230139831377944579

[imron]

1 hour ago, realmayo said:

but perhaps be sceptical of some scepticism! 

I'm skeptical of everything

[realmayo]

1 hour ago, imron said:

I'm skeptical of everything

 

I doubt it!!

 

[Dawei3]

A long standing aspect of Chinese culture is that accuracy in reporting numbers isn't valued as highly as in the West.  This shows up in many ways, including government reports.  (And to be clear, this is like any aspect of a culture:  it doesn't apply to everyone.  Some in China are trying to change this, as the below explains). 

 

The book, Mao's Great Famine, notes that one of the causes of the famine was that Mao kept pushing for higher rice production targets.  To meet the targets, lower government officials added sand and water to rice to make it heavier.  This usually destroyed the rice.  All that mattered to the local officials was "make the numbers."    

 

For those who have access to the journal Science, Liu & Yang (2009) discussed a more recent attempt to "fight statistical corruption."  They note the phenomenon of "numbers make the leaders" is a motivation for "leaders make the numbers". 

 

They give the pinyin phrases for these statements.  Can anyone offer the characters for these? 

Shu zi chu guan (numbers make the leaders)

guan chu shu zi  (leaders make the numbers)

https://science.sciencemag.org/content/325/5941/675

 

They note a similar law from 1983 was too vague to enforce (vagueness is another characteristic of Chinese law).  (i.e., this is nothing new)  

 

A lack of concern on the accuracy of numbers is a constant concern in drug development and laboratory studies in China.  In ~2016, the Chinese FDA (now called NMPA) issued a rule that if a company was found to submit false data in new drug application (NDA), the company would be barred from submitting NDAs for 3 years.  It was an ingenious approach because CFDA didn't need to prove falsification.  Notably, over 80% (!!!) of the submitted NDAs were withdrawn.  And >90% of the withdrawals were by Chinese companies.  This suggested that prior to this, CFDA had been approving drugs whose data on safety, efficacy, & quality were usually fabricated.   

 

This was so embarrassing that later CFDA retracted the story.  However, I had the slides from a Chinese regulatory expert who had the specific numbers of the NDAs that were withdrawn & when (i.e., the original story was true).  To be clear, any country can have unethical companies.  The importance of the above is that CFDA showed that virtually all of the Chinese drug companies were unconcerned about the veracity of their numbers.  (n.b., there are some good Chinese pharma companies).  

 

 

 

 

 

[carlo]

数字出官 官出数字。I'm also skeptical of most references to "cultural differences" -- there are many good statisticians and accountants in China. It might have something to do with power structures though.

[Dawei3]

10 hours ago, carlo said:

here are many good statisticians and accountants in China

Excellent point.  

 

In large pharma in the US, most of the statisticians are Chinese because they do excellent work.  Fraudulent data happens rarely.    

 

In Chinese pharmaceutical companies in China, the statisticians are likely  equally competent.  Yet,  the Chinese FDA found that the  Chinese companies were submitting a huge amount of fraudulent data.  Why?  Liu & Yang would likely say:   Leaders make the numbers (not the statisticians)

 

Accounting is also an excellent example.  The book, The Chinese Tao of Business, mentions that a Moody's found that the books of China's Big 4 banks were meaningless.  Odds are these 4 banks had excellent accountants.  Why were their books meaningless?  Leaders make the numbers (not the accountants).  

 

If just a few companies did these kinds of things, I would agree it's not culture.  Unfortunately, the above isn't rare or isolated.  There is definitely an effort to change this in China, but it is pushing against long-standing cultural practices.  

 

(I could give other examples, but this could sound like polemic against China.  It's not.  Most of the sources of these examples are Chinese authors who want to change things for the better).  

 

 

 

 

 

[suMMit]

So now two prisons (that we know of) have become infected. I was thinking about this weeks ago, as I know someone in prison in Hebei. There's been an outbreak in a Beijing hospital, as of yesterday 36 confirmed and hundreds being checked. There Daegu in South Korea now under self imposed lockdown. I do not see what is to stop all kinds of outbreaks starting in places all across china - 3rd and 4th tier cities, villages, hospitals and yes office buildings. Very worrying imo.

[somethingfunny]

17 hours ago, Dawei3 said:

A long standing aspect of Chinese culture is that accuracy in reporting numbers isn't valued as highly as in the West.

 

I’m not sure this is fair. Maybe “A long standing aspect of communist states...” would be more accurate.