How we are deceived by statistics
How we are deceived by statistics
Anonim

On July 5, the Levada Center published a study that 91% of Russians have a negative attitude towards people walking in swimsuits. Among the opponents of bikinis and swimming trunks, the majority are respondents aged 40 to 54. Despite this, some media presented information from a different angle, saying that all Russians have a negative attitude to walking in negligee. We decided to figure out what tricks in statistics can be used to make the information seem more attractive.

How we are deceived by statistics
How we are deceived by statistics

Why ask only people between the ages of 40 and 54 about topless sunbathing? Perhaps we should go further and ask the age group over 80 about whether we need the Internet? By presenting the same information in different ways, you can radically change the way others see it. Here are some examples of how statistics are used to cheat.

Using metrics that are only good at first glance

Example: 90% of all vehicles sold over the past 20 years are still on the road.

It seems like a very good brand since the machines are so durable. But think better. Perhaps this car brand was released only 10 years ago? Then she no longer seems so attractive.

A more correct and less yellow title should have sounded like this: "90% of all cars over 20 years old are still on the road."

Performance claim without comparison to alternatives

Example: this pain reliever will relieve headaches as effectively as possible.

It makes no sense to talk about the effectiveness of a product without comparing it with others. "Most effective", "better than others", "highest quality" - these words should make you think about whether to buy this product. If you want to prove that your pain reliever is the best, you need to compare it to other brands. Otherwise, these are useless words.

Playing with graphs and charts

Example:

Apple presentation
Apple presentation

At this conference, Steve Jobs talked about the share of the iPhone among all smartphones in the United States. Despite the fact that iPhone is used by 19, 5% of residents, its share on the diagram looks larger than the share of "Others" (21, 2%). Visually, this can be achieved by giving the diagram a 3D effect.

Submission of information without confirmation

Example: after the legalization of marijuana, the number of smokers in the Netherlands increased.

Such "facts" are worthless without confirmation. Perhaps the site on which you read this just forgot to link to the study, but in any case, there is no point in believing this information.

The reference point on the chart is not zero

Example:

Obamacare Support Schedule
Obamacare Support Schedule

The photo shows that the number of participants in the Obamacare program has increased by 1,066,000. That is, the difference is about 17%. On the diagram, the difference between the columns is almost threefold. This is due to the fact that the reference point is not zero.

Statistics provided by the interested party

Example: we tested our new shampoo and came to the conclusion that it is more effective than all analogues on the market.

And finally, a rather obvious fact. If the research is conducted by an interested party, then you need to trust its results with extreme caution.

Recommended: