Fake master Yi Fuxian's paste statistics: "planned death" finally became a reality under his pen

Around the beginning of February 2018, I went to the University of Wisconsin group on reddit to reveal that Yi Fuxian's prediction that there will be tens of millions of families who lost their only child in China by 2035 was a data fraud. Unsurprisingly, I got a bunch of "shit", "garbage" and more than one person calling me "50-cent dogs" or "fifty cents." (Writing here, let's smile first ^_^).

Amid the piles of hostile comments, though, I also made a major discovery.

A netizen named "Gibborim" and majored in "Electrical and Computer Engineering" pointed out to me that "for every 10,000 people born, x people die before the age of y". "It's not the "0-y-year-old mortality rate" as I understand it, and also explained to me very carefully how Yi Fuxian might have calculated his evidence. Of course, in the end he wanted me to believe in Yi Fuxian's "professional" approach.

Although his words were fierce, I finally realized that his first statement was correct after thinking for a few days with my dull brain that failed in elementary school arithmetic: I did misunderstand Yi Fuxian's statement . What Yi Fuxian said "For every 10,000 people born, there are x people who die before the age of y", which refers to the total number of deaths of these 10,000 people before they reach the age of y, that is, put them at y+1. The total number of deaths added during the year.

But does this mean that Yi Fuxian and Gibborim's calculation method is correct? Now I will make a non-professional analysis. If there is something wrong, please point it out.

1. How did Yi Fuxian calculate that "for every 10,000 people born, there are x people who die before the age of y"

Regarding Yi Fuxian's algorithm, Gibborim made a very detailed analysis, and even made a graph of his calculation results. Netizens who understand English can refer to his reply at the back of my post .

It's just that his algorithm is too complicated for my brain who failed arithmetic in elementary school, but I still understand the general idea of how he and Yi Fuxian calculated this data. For the sake of illustration, I will paste some screenshots of the table they used to calculate here (please click here if you need to view the original text of the entire table: http://www.stats.gov.cn/tjsj/ndsj/renkoupucha/ 2000pucha/html/t0604.htm ):

Simply put, to calculate Yi Fuxian's "every 10,000 people born, x people die before the age of y", you need to know the annual number or death rate of these people. For example, if we want to calculate how many people born in 2000 in total died before the age of 10, in addition to the number of their deaths at age 0 already provided in the table, we also need to know they died at age 1, 2, 3... …the number of deaths per year up to the age of 10 (11 data in total), then add up these numbers, divide by the number of people they were born with, and multiply by 1/10,000 to get the kind of data that Yi Fuxian said .

But this table only provides the number of deaths, the total number and the death rate for a single age group in the current year. The other 10 data needed to calculate how many people born in 2000 died before the age of 10 are not available in the above table. .

There is a saying in China that "a clever woman can't cook without rice", but such a huge lack of data is obviously not difficult for our forgery master Yi Fuxian. Because he directly used the mortality rate of people born in other years in the table to calculate the number of deaths in the relevant age group at the corresponding age .

For example, from this table, we know that people born in 2000 have a mortality rate of 26.9‰ at age 0, then they reach age 1, age 2, age 3, age 10, age 25, age 44 What is the death rate? This table is not provided. According to the calculation method of Yi Fuxian, they actually use the death rate of people born in 1999 at the age of 1 to calculate the number of deaths of people born in 2000 in 2001 (that is, when they were 1 year old), and use the death rate of people born in 1998 at the age of 1. Mortality at age 2 counts how many people born in 2000 died in 2002 (and when they were 2 years old)...and so on.

(To put it bluntly, the essence of his algorithm is that all the people of the dynasty before and after that, whether they were born in the era of great famine, the era of the Cultural Revolution, the era of reform and opening up, or the 21st century, must die according to the mortality rate in the above table. The "planned death" that the anti-birth control faction longed for finally became a reality under the "magical pen" of Yi Fuxian. )

For the sake of illustration, let's take a relatively short period of time as an example. If we use n to represent the number of live births in a certain year, and A, B, C... to represent the deaths at the age of 0-2 in the above table rate, then the number of deaths in the group at age 0 is n*A, the number of deaths at age 1 is (nn*A)*B, and the number of deaths at age 2 is [nn*A -(nn*A)*B ]*C

Therefore, according to the algorithms of Yi Fuxian, the total number of deaths in this group at the age of 0-2 is: n*A+(nn*A)*B+[nn*A -(nn*A)*B]*C= n*A+n*Bn*A*B+n*Cn*A*Cn*B*C+n*A*B*C=n*A+n*B+n*Cn*A*Bn*A* Cn*B*C+n*A*B*C

So for every 10,000 people born in this group, how many will die before the age of 2? Its calculation formula is:

x=[(n*A+n*B+n*Cn*A*Bn*A*Cn*B*C+n*A*B*C)10000]/[n*10000]= (A+B+ CA*BA*CB*C+A*B*C)*10000/10000

That is, x= (A+B+C+…-A*BA*CB*C-…+A*B*C*…) *10000/10000

(I haven't done math problems for a long time, and my eyes are dizzy or dim when I write here *_*)

Since the mortality rate in the table is per thousand, for example, the 0-year-old mortality rate is 26.9‰ (ie 0.0269), the 1-year-old mortality rate is 2.49‰ (ie 0.00249), and the 2-year-old mortality rate is 1.60‰ (ie 0.0016) Therefore, A*B =0.000066981≈0.00007,

A*C=0.00004304 ≈0.00004

B*C=0.00003984 ≈0.000005

A*B*C=0.0000001072

The more mortality numbers you multiply, the smaller the resulting number, except for those 0.00007 and 0.00004, the rest are almost negligible. No wonder the numbers Yi Fuxian calculated were similar to adding up the death rates of all ages!

Gibborim didn't hesitate to go to the trouble of calculating the total number of people who died before the age of 25 for each age in the original table. Judging from this formula, he/she was completely superfluous.

2. Is Yi Fuxian's algorithm feasible?

Based on the data in the above table, can Yi Fuxian's method be used to calculate "how many people die before the age of y for every 10,000 people born", and then calculate the number of families who have lost their only child?

The author believes that his method is not completely infeasible, but it must meet 3 very demanding conditions:

1. The country must have been in a state of stagnation for decades;

2. The country has not had any large-scale natural and man-made disasters affecting the mortality rate of the population for decades;

3. The country must be in a state of absolute egalitarianism.

"Stagnant development" in Condition 1 means neither development nor regression, because technological progress and economic development both reduce population mortality, especially infant mortality (i.e., 0-year-old mortality), and vice versa may increase these mortality rates. mortality rate.

The large-scale natural and man-made disasters mentioned in condition 2, such as wars and epidemics, will also increase the mortality rate of the population. For example, the Black Death in the Middle Ages in Europe and the two world wars (and the current global pandemic of the Wuhan pneumonia plague) have caused a large number of The death of the population, in turn, pushed up the death rate of the relevant countries and regions in those eras.

As for the state of absolute egalitarianism mentioned in Condition 3, it is to eliminate the difference in mortality rates between different social classes due to different economic conditions . The death toll of only children would be very unreliable.

I scratched my head and thought about it for a long time, and felt that we simply could not find a country on earth that fully met the three requirements of the appeal, and even a country that was close to these three conditions was very difficult to find, and Xi Dynasty was no exception.

If you don't believe me, please go to Wikipedia's " World Infant Mortality List ": In the past 50 or 60 years, the data of all countries in the world have greatly improved, and Singapore ranks first, from 1950 The 60.69‰ in the early 1990s dropped to 1.92‰ in the early 21st century; while Afghanistan, which was at the bottom, also dropped from 275.03‰ to 135.95‰. This is evidence that technological progress and economic development can reduce the mortality rate of the population.

There are so many blablabla, I actually want to explain one thing: in the real society, Yi Fuxian's absurd calculation method is almost infeasible.

3. Where did Yi Fuxian go wrong?

If you have a little understanding of statistics, then it is not difficult to see that Yi Fuxian made two very obvious mistakes when using the data in the above table for analysis and calculation:

Error 1, using the data of group A in a whole to extrapolate the data of other groups . For example, this is like calculating the total GDP of Gansu in that year or other years based on the per capita GDP of Shanghai in a certain year (that is, multiplying the per capita GDP of Shanghai by the total population of Gansu in the corresponding year). This is wrong.

The second mistake is to use the overall data to estimate the data of one of the groups . For example, this is like using the national per capita GDP in a certain year to calculate the total GDP of Gansu or Shanghai (that is, multiplying the national per capita GDP in that year by the total population of Gansu or Shanghai in the same year). This is also wrong.

Specifically, when Yi Fuxian put the years 2000, 1999, 1998, 1997... 1975... 1956, people born in 1956 were at the age of 0, 1, 2, 3... 25 years old... …the mortality rate at age 44, when taking the mortality rate for one or all of the age groups at the corresponding age, he made the first mistake. In fact, by calculating the number of deaths in any of the age groups within a certain number of years, the data he calculated can be obtained, because we can already see from the previous formula that his so-called "every 10,000 people are born, how many people are there death at the age of 2” is about adding up the mortality rates for the corresponding age groups.

When Yi Fuxian used the death data, which was calculated in the wrong way, to calculate the number of families who lost their only child, he made a second mistake. Because, as I said in "One of the Biggest Lies of the Anti-Birth Control Faction: "Tens of Thousands of Families Losing Their Only One" ", the areas where China strictly implements the one-child policy are mainly cities. The gap between urban and rural areas in China has always been large, and accordingly, the gap in mortality between urban and rural populations has also been created.

So even if his first data were accurate, using this national average to calculate the number of deaths of only children as one of the groups would have a large margin of error. What's more, he still uses historical data to calculate the future situation, which is even more unreliable.

This is also the reason why many netizens shout that their income is "averaged" every time the Bureau of Statistics releases the national per capita income. Because we can't use that per capita income to calculate the income of a single individual (or part of a group). What we can do is to compare our actual income with the national per capita income to see roughly where our economic situation is in the country.

So, is there a more reliable way to estimate "for every 10,000 people born, how many people will die before their age" from limited data? The answer is: yes.

For example, if we can find reliable data elsewhere on the number of births in 1990, 1975, and 1956, then combined with the above table, it is easy to calculate that "for every 10,000 people born in 10 died before the age of 25 and 44."

However, before this sentence, please be sure to add the corresponding year. You can never generalize the data calculated from this and take it for granted that it applies to all eras like Yi Fuxian did ( this is also the fault of Yi Fuxian). The third mistake ), let alone use them to calculate the number of deaths of only children. At most, based on these data, we can only compare very, very, very roughly whether the death trend of the only child born in the first two years (the third year is not counted, because family planning was not implemented at that time) is high or low. Low.

From the analysis here, it is not difficult for us to come to the conclusion that Yi Fuxian made at least three mistakes when calculating the "tens of millions of families who lost their only child". Essentially, they are all anti-statistics and anti-science .

To put it a bit harsher, these ridiculous methods of his are just a bunch of nonsense, which is simply " slurry statistics ", which is also the methodological basis of his " data falsification ". However, by virtue of his status as a "senior scientist at the University of Wisconsin in the United States", his nondescript data can confuse the general audience, including many reporters, including reporters from some internationally renowned media.

4. How to interpret the data and conclusions in media reports?

The reason why Yi Fuxian's "tens of millions of families who lost their only child" has been circulating for many years and spread around the world is mainly caused by some groups deliberately hyping up the topic of anti-birth control and family planning, but it is also related to the collective lack of professionalism of the global media reporters.

Looking at various media reports on Yi Fuxian's data (and some of his other "research results"), we will find that from the Global Times to the New York Times, there are quite a few articles that almost only show Yi Fuxian's views and/or similar views to Yi Fuxian are rarely interviewed by journalists who have views that are different or even opposite to his. Therefore, naturally there will be no professionals to test Yi Fuxian's data and "research results".

That is to say, such articles are far from the "balanced reports" that some media have regarded as the standard. They are completely propaganda like the Communist Party media . A recent article by Radio Free Asia, " Yi Fuxian: Many Chinese Policies Are Based on Incorrect Demographic Data ," is such a biased propaganda article.

Therefore, when we read various news and news from various media and even from the media, we must be vigilant against those reports with a single point of view and position (this kind of report is the mainstream of "news"), because such reports even Not outright lies, or at least biased.

If there is a guy who claims to be an expert who claims to calculate the results based on what authoritative statistical data (especially those seemingly huge numbers), but is very secretive about his calculation process and even the source of the statistical data, and the article If you likewise only list his or her and his or her supporters' views, be more vigilant. Because this is what Yi Fuxian and his likes have always done.

Writing here, I also want to make fun of my experience inside and outside the wall: when I tried to tell the truth inside the wall, the netizens in Xichao (mainly the fifty cents navy) scolded me as a cents party; When I tell the truth, some western netizens call me a five-haired dog. It seems that people who tell the truth, whether inside or outside the walls, are not welcome, because the truth is often dazzling .

It feels like Yi Fuxian is now like the new coronavirus that is spreading around the world. The more his lies are exposed, the more media such as Radio Free Asia and the BBC pursue him, and the farther his lies spread. Some websites of the "Republican Party News Department", after I left a message criticizing Yi Fuxian's data as unreliable, changed Yi Fuxian's name to "expert" and continued to spread his lies.

Soon, the "New York Times" as the "Democratic Party" will also dispel the old quarrel with the "Republican Party" and join the spread of lies that regularly appear at the end of the year and the beginning of the year (it appears at the same time as the flu and the coronavirus). The virus has been around for the same time, that's it!), will the anti-birth control fake news at the New York Times this year be a bit more sophisticated than it used to be? Let's wait and see.

(This article was originally published on my blog, because the original layout is ugly, so I reposted it on Matters after a little modification and arrangement.)