Make "communicating with data" more credible!
"A Oxfam survey by the philanthropic development organization shows that the combined wealth of the world's 85 richest people is equal to half the wealth of the world's poorest people combined."
This was the guardian headline in January 2014. For a time, the public was in an uproar, and many media followed suit and reported one after another, and everyone was very shocked by such a data comparison.
Three years later, Oxfam made significant changes to the survey report, changing the title from "85 billionaires" to "8 billionaires".
Is the inequality between rich and poor really 10 times worse? Did the billionaires' wealth grow 10-fold, or did the wealth of the poor shrink by 90% for no reason?
So, how many people are in the hands of the world's wealth?
Oxfam's publicity has caused inequality between the rich and the poor to arouse the indignation of the people at once, so that the people can no longer calmly distinguish and reflect on the authenticity of the data.
In the three years since the absence of a major economic crisis, Oxfam's survey data has changed so much that one cannot question whether the original report looked at inequality between the rich and the poor.
In fact, Oxfam's purpose is to attract the public's attention, and as for the truth, it is a secondary issue for them.
The rush to reprint this data has pushed this data to the wrong interpretation, in which the Independent newspaper said that "the wealth of the world's richest 85 people is the same as the sum of the wealth of the rest of the world's population." This statement directly confuses "half of the wealth of the poor" with "the wealth of non-billionaires", which is reflected in the difference between less than $2 trillion and more than $200 trillion, and the conclusion without repeated verification will lead to such a serious fallacy of a hundred times the difference.
This ridiculous mistake reminds us that when people are confronted with some data, they can easily be provoked and lose their minds. It's true that there are inequities in the world, but when we look at these data and just stare at them, it means that we start to be emotional.
In fact, such a mistake is already ridiculous, and it should be easy to find. But when we are dizzy about something instead of using our minds, there will be such a situation where we cannot distinguish between right and wrong, and emotion rather than reason prevails.
In the chaotic information age, people seem to have lost a cognitive and discriminating ability, passively accepting information "attacks" from various data. While statistics are easy to lie about, it's easier to lie without statistics.
So how do you mine valuable data from a world of disinformation, bad research, and bad motives? We have extracted five rules from the book "Patchwork the Truth" to teach everyone to use statistics correctly and make "communicating with data" more credible.
Title: "Piecing Together the Truth: Recognizing the Ten Laws of Big Data in a Complicated World"
By Tim Harford
Rule 1: Don't be confused in the heart, don't be trapped in emotions
The first step in learning can start with learning to control emotions. Learn to observe and reflect on your emotions. When you see the data results, pay attention to your own reactions. If you see those statistics, whether you feel angry or happy, or in disbelief, you have to pause and reflect. You don't have to be a person without feelings, but since you can feel with your heart, you can definitely think with your brain.
When we come across statistics on a world problem and think about whether to like and retweet on social media, or refute it vehemently, pause and ask ourselves the question, "Why am I so emotional?" "We're doing this not just for ourselves, there's a social responsibility in it. We've seen how much social pressure affects our perceptions and the way we think about problems.
We need to settle down slowly, learn to control our emotions and set aside our positions, and focus only on the facts themselves, so that we can not only think more soberly, but also provide others with the right mode of thinking about the problem, that is, we think and reason about the problem not as a member of a political faction, but as an individual with an unbiased attitude. I want to form such a habit. I hope that's your wish too.
Emotions can influence people's thinking. So when it comes to reading statistics, expertise and technology are important, but if we don't put the reins on the wild horse of emotion and let it carry us to believe and doubt, we will eventually lose our front hoof.
Rule two: A combination of a worm perspective and a bird's-eye view
Try to learn to see things from two perspectives: the worm perspective and the bird's-eye view.
Two perspectives will show you something different, which may also become your dilemma: is the two scenes true or false? Such questions will set you on a journey of inquiry. We will find out later that sometimes statistics can mislead us. Sometimes, our own eyes deceive us; Sometimes, once we understand the cause of the matter, the discrepancy between the number and what the eye sees can also be understood.
It is easy for people to understand what they see from their own perspective as the whole picture, which psychologists call "naïve realism," that is, thinking that what they see is the truth without any deviation. This kind of naïve realism can seriously mislead us.
Naïve realism can lead to a misunderstanding of many things. For example, the Morrie polling agency surveyed nearly 30,000 people in 38 countries on a range of social issues. It turns out that these people—who can represent most of us—have a serious mismatch with reliable statistics, and here's a list:
Our understanding of the murder rate is wrong. We thought homicides had been on the rise since 2000, but in most of the countries surveyed, the proportion has been declining.
We think that the number of people who have died of terrorism in the past 15 years is higher than it was 15 years ago, but in fact the number has declined.
We believe that 28 percent of prisoners are immigrants. The Morrie survey estimated that the true proportion of all countries surveyed should be 15%.
We thought that 20 per cent of adolescent girls would give birth every year. This number doesn't really have much credibility from a biological point of view. Counting from the age of 12 with fertility, an 18-year-old girl has a 20% probability of 6 births, so most 18-year-old girls should have at least one child. Let's look at the situation around us, is this true? According to Mori's survey, the correct figure is that only 2% of adolescent girls give birth each year.
We thought 34% of people had diabetes, and the real number was 8%.
We think 75% of people use Facebook. At the time of the 2017 survey, the figure was 46 percent.
The events reported in the news are also data to some extent, and although they are not representative data, they really affect our view of the world. In Kahneman's words, they are "fast numbers"—numbers that people can draw conclusions at once.
The numbers provided by the bird's eye view are boring and rigorous, but comprehensive and profound, and the data seen by the worm perspective is vivid but relatively one-sided, and it is not easy to balance the two perspectives. We need to always remind ourselves that while we understand these things, we may also overlook other things. Statistics, like other disciplines, rigorous logic and personal experience should complement each other and correct each other, and only by organically combining the two is the most ideal method.
Rule Three: See the definition of the data
When we want to understand any statistical result, we must first think about what the actual meaning of the result is. The COVID-19 outbreak raises similar questions.
On 9 April 2020, media reported that 887 people had died of COVID-19 in the UK mainland in the past 24 hours, but I happened to know that this figure was wrong. Scottish statistician Sheila Bird did a thorough investigation and told me that the true number was likely to be around 1500 people. 5Why are the numbers so different? This is partly because some people die at home, while the authorities only count those who die in hospitals, but mainly because hospitals that have expanded due to COVID-19 are too late to update the death toll reports, often laging for a few days.
Today, Thursday, the announced death data could be the number of deaths on Sunday or Monday. With the death toll soaring in recent days, it's easy to underestimate the seriousness of the current situation, telling us that the data from three days ago is.
Many of the problems are because people are going in the wrong direction in the first place. They are obsessed with statistical technical problems, such as asking about sampling errors and margins of error, debating whether numbers are rising or falling, believing, doubting, analyzing, and dissecting various numbers, and they have not taken the time to understand the first and most important question to ask: What is the statistical object? What standards are used?
We first have to figure out what the statistical object of the data is, and then we have to calculate the mathematics.
Rule 4: Learn to look at data in the big picture
Pulling away from the problem can make you feel macro. Every time you see a statistic, you can think, is that a big number?
Let's take former US President Trump building a wall on the US-Mexico border as an example to talk about the macro feelings. It will cost $25 billion to build the wall. Is that number big? It does sound a bit big, but to really understand the number, you need something as a reference.
For example, the U.S. annual defense budget is nearly $700 billion, or $2 billion a day. Therefore, the cost of building the wall is equivalent to the two weeks of military expenditure of the US military.
Or, there are about 325 million people in the United States, and $25 billion divided by 325 million people costs about $75 per person. Whether this number is big or small, you can judge for yourself, but I guess with these comparisons, your judgment will be more reasonable. Pulling away from the problem can make you feel macro.
If you can remember some numbers with ruler meaning, they can bring you a lot of convenience. You can compare them with numbers (a 10,000-word report may seem long, but an average novel is 10 times longer) or you can count averages (the U.S. defense budget is more than $2,000 per person per year).
These ruler numbers, whether you have memorized them in your head or you look them up, can be used to do the math, or to do the same with the calculator. It's a simple thing to do, but it's very inspiring.
Rule 5: Know if the statistical sample is comprehensive
In fact, we must often ask: "Who is missing from the data?" What's missing?
When it comes to data, scale doesn't mean everything. There are two problems to deal with: sample error and sampling.
The deviation sample error reflects a situation in which sometimes it is purely accidental that the people randomly sampled do not reflect the true opinion of the public. "Margin of error" refers to this risk, but the larger the sample, the smaller the error. Random interviews of 1,000 people are a large sample of polls for any purpose.
There is a bigger pitfall to be avoided in polling, namely sampling bias. Sampling error means that randomly selected samples do not happen to reflect the real situation, while sampling bias is that the sampling does not cover the full sample type.
How many blatant racists and sexists there are in society, you can see by looking around. But in general, what we count, or who we miss, is caused by carelessness in our own choices, less serious prejudices, and some inadvertent transgressions.
Unless we collect data ourselves, there is limited capacity for statistical leakage. But at the very least, when someone else gives us data, we can and should remember to ask who or what content might be missing.
Big data may seem comprehensive, and it may be very versatile, but "one is not less" is an easily deluded illusion: everything is in your own hands.
In fact, we must often ask: "Who is missing from the data?" What's missing? "This is just one of the reasons why we should be cautious about big data. Big data represents a huge change in the way data is collected and counted, and the impact of this change remains to be examined.
Titanium Time New Book Recommendation - "Patching Together the Truth"
Original price: ¥65 Titanium Empty Cabin Special Price: ¥42.2
Statistics can help us piece together the truth of the world, the book lists ten rules to help readers understand statistics, and teach everyone to use statistics correctly, to help individuals, organizations or government departments make better decisions, hurry up and buy!
Go to the Titanium Cabin homepage to collect coupons and place more discounts
Long press the QR code below and buy now
Beijing shipping Yuantong Express
▼
https://shop16192226.m.youzan.com/wscgoods/detail/36copvwawok3eef?scan=1&activity=none&from=kdt&qr=directgoods_2035332285&shopAutoEnter=1
---------------- gorgeous dividing lines-----------------