The abuse of statistics

%%%%%%%%%%%%%%%%%%%%

When you are watching an advert/commercial on TV and the company claims that 89% of woman saw an improvement in their skin within two weeks, it sounds great right? But in the subtext of some commercials, you can see that they product was tested on less than 100 people. Is that a reliable sample size? In addition, what counts for improvement? When products are being sold to us, companies constantly abuse their use of statistics, twisting them to sound better than they are in reality. I don't want to buy a mascara that only 20 woman have tried and 15 liked. The portion of sample is just too small.

Unfortunately this statistical abuse is not solely limited to product sales. I have also observed it (in growing amounts) within politics. Skimming the news the other day, I saw that it was claimed in America (by a reputable source) that terrorism was up 80%. Maybe that is true in one sense. But the way it was phrased implied that there was an 80% increase in terrorism towards America. The statistic could easily breed fear, hatred and obviously support all of the American anti-terrorism policies like monitoring, sending troop etc. I am not implying that the statistic isn't true. It could well be that globally terrorism is up 80%, this shouldn't come as a surprise because any attack, bombing, shooting by ISIS will count as a terrorist attack. So the statistic was somewhat twisted.

And again today, I saw the same thing. A global survey of 40 countries, it was claimed that 69% of people saw America in a positive light, including the American effort against ISIS. Having lived and travelled extensively abroad, I find this statistic very hard to believe. Maybe the people surveyed were all highly educated, working in government, and able to understand all the actions taken by the American government. But for the most part, I believe that average person tends to have a poor outlook on America. I know from experience that most people view the U.S as a country that is glutinous, violent, and has a very unequal treatment of people. In a lot of ways I agree. This is why I struggle to trust that 69% of people actually like America. Especially given that many find American war actions as over the top and violent, I doubt 69% agree with the US moves against ISIS. But perhaps I am wrong.

The point of this post was to highlight that when you read the news, or watch a commercial, don't trust every number. Statistics are abused, inflated, manipulated to support a point. In academics it is a common thing. If you find a statistic from one study that supports your own in some way, you use it. Does it convey a different point than the original statistic? If yes, than it is manipulating the numbers. I don't necessarily have strong feelings against it, I just think that more people should be aware that the numbers you hear can't always be trusted. Statistics sound lovely and nice, but often should be regarded with a hint of skepticism.