8 good reasons to ignore any statistics

Not all statistics are equally poleznaStatistika may cause distrust for at least two reasons. Reason number one: the human brain is very weak in the intuitive understanding of statistical data, we are not able to "estimate" (and hence check) Statistics on the eye. Reason number two: easy to manipulate statistics (and it happens constantly), it gives a wide field for speculation and juggling data, despite a strict, mathematical its character as a science.

So, before you look at the statistics and see its details, answer a few simple questions ...

1. Who shares this statistic? The first way is to check the statistics is to ensure to check out who the data shared.

The phrase "statistical survey" can often mean «a couple of people decided to publish a bit of nonsense» h4> It seems unscientific, as should respect the data, and not feel prejudice against those of their provided, as well as to those to whom they are already affected. However, data should be considered under all possible angles, not only from the point of view of those who have supported.

Properly prepared statistical information, in theory, should be unbiased, but in full, this condition could not be implemented because at any given moment, people have read this information, there are certain prejudices and they begin to defend a particular point of view , thereby preventing an unbiased perception of statistics.

That is why such people should be avoided. Even if their arguments sounded interesting, they should be ignored, and then familiarize yourself with the original statistics.

2. Who is the source? It often happens that there are no "committees statistical research" simply does not exist. And the phrase "statistical survey" can often mean "a couple of people decided to publish a bit of nonsense».

Check the source of statistical information - it is entirely your responsibility and only after checking the source, you can take into consideration what the statistics show.

Any web page statistics, usually filled with beautiful graphics, there is usually only one sentence, consisting of numbers, and the source of the statistics indicated the fine print at the bottom right, blue letters on a blue background. And you will not see who is the source, as long as all the graphs view, and scroll down the page.

And it's strange. After all, in real life, we usually first look at who we face, and only then decide if he believe or not. So the first thing to do, looking at the page with the statistics - a link to its source. If there is no such link, statistics are not credible.

3. What is the sample? < If the size of the statistical sample is not listed next to the results of the statistics, the data - a potential debris. If the size of the statistical sample is not specified at all, statistics - clear the garbage, the main purpose is not to inform you about something, and to lull your vigilance and turn off critical thinking.

There are many examples of amazing statistical studies lasting for decades, during which gather invaluable sociological data.

But still there are many examples of senseless waste of paper when some bored students are plotting their own "statistical survey" does not hold water.

4. What is the margin of error? How wide the limits of permissible error on the graph?

If they are not very wide, then The given statistics can be useful.

The physics of elementary particles is possible only a certain percentage of the measurement accuracy, and that many do not like, as the person does not like uncertainty, and loves to specific numbers h4> If the limits are wide, the data may less accurate, but also useful. And if outside the margin of error reduced to zero, the data just garbage.

Exact figures look more "scientific", but on the sociological level, this "precision" and "scientific" a performance full of illiteracy.

For example, in the physics of elementary particles it is possible only a certain percentage of the measurement accuracy, and that many do not like, because the man does not like uncertainty, and loves to specific numbers.

For example, you can take the statistics about the depth of the holes, the laser burns in different materials. With all the power of modern technology to measure exactly the depth of the microscopic holes it is not possible, so that reliability of the results of this statistic necessarily indicates the maximum permissible error. Pointing it, you can still get useful statistics.

And you can do anything else: just ignore this error misled the whole world, but giving people their favorite specific figures.

5. "The Last" does not mean "best" The easiest way to create a loud headline - declare that there is some "recent" study refutes all previous knowledge about a particular subject. If you already have a million studies claiming one thing, the sudden the opposite is bound to attract everyone's attention.

Most often this is due to the food.

The tabloids initially argued that the eggs are useful for you, then start to say that eggs are harmful, then the eggs are again useful, and no end in sight.

Opponents of Science are trying to use this mechanism to discredit science. They use these controversial "research" in order to undermine the credibility of the scientific progress.

The only way to cancel the results of the "last" of research - to develop his own theory, based on the results of the "old" research.

Then plan the experiment in which the suggestion put forward in the "final" study will do one thing, and your theory will do more, and finally run the experiment. If the assumption of the "last" Studies have confirmed then there will be an occasion to review the results of all previous studies. And if not, there will be a serious reason to doubt the competence of the "researchers».

6. Questionable headlines "The excess of sleep can increase the risk of stroke up to 50% inclusive, suggests a study».

Title study the above should only suggests a possible link between the change in sleep and stroke, and this possibility requires further serious consideration. A headline writer has made this assumption in almost claim that "excess sleep will kill you».

The goal of any analysis is to obtain answers, free of emotional coloring. These responses are not possible to get when we see in the headlines "from ... to ... inclusive». If you follow this logic, then any of your daily activities increase the risk of sudden heart attack to 100% inclusive. And now you're terrified already begin to imagine what it will be your sudden heart attack, completely forgetting about the fact that it is "up to 100% inclusive" includes other numbers, much less hundreds. For example, zero.

7. "secondary" matematikaA sometimes numbers are scary enough. But writer headers have a calculator. Right on your desktop, just below the solitaire. So if in some study says that the average person a chance to turn into a zombie for about one in a million, the writer multiplies the chance of a $ 7 billion, and for thousands of dead people hobble through the streets.

If we could simply multiply the number of statistics on even greater numbers, we generally would not need statistics. Because we all would be dead from thousands of disappearing unlikely reasons, which would have killed us. < br />
At a time when the writing fraternity drew his attention to the statistics, we should act as if they are talking about filling his car accident found a small amount of nitroglycerin. They do not know how it works, they have no idea of ​​the conditions in which it operates.

So do not go where they call you. Because their discovery will soon blow up in their faces.

8. Factors erundyKazhdy time a person stands up from his chair and injured his chances are doubled. This can lead to serious injuries, disability and possibly even death.

But all these assumptions about how horrible it would be if all this is really happening, do not change the fact, as it is unlikely. So that a person can hardly refuse to sit in a chair.

statistics - a mathematical "Tommy». It can be a very useful tool in the hands of a clever scientist, but it can also be used for other purposes, and then cause people harm h4> similar picture often seen in the reports of Ministry of Health, which says very seriously, that some factors could suddenly triple your chances of getting ill, and they amount to as much as much as 0 percent 0000001.

Statistics can give amazing results. After all, it works on levels inaccessible to our understanding. Often, the human mind simply can not imagine what it is - 7 billion of anything.

However, the statistics - a mathematical "Tommy." It can be a very useful tool in the hands of a clever scientist, but it can also be used for other purposes, and then cause harm to people.

Follow the rules listed above, and you'll be protected from the huge amount of nonsense. Yes, for that you will have to ignore most of the things that you have shared on the network, but it is quite normal, as the only universal statistical law is Sturgeon's Law: 90 percent of the world - is nonsense. And the work boils down to in order to find the remaining 10 percent to be built on the basis of their own world.

via www.cracked.com/blog/8-obvious-signs-statistics-are-lying-to-you/


See also

New and interesting