statistical misuse

Upload: cecilia-thomas

Post on 02-Mar-2016

11 views

Category:

Documents


0 download

DESCRIPTION

Presentation about the misuse of statistics in the world today

TRANSCRIPT

Three Types of LiesAt the age of about 9 years old, I was doing some light browsing on the internet. I was going through several sites with mundane articles or jokes, until eventually I stumbled upon an article called The One Thing Could Put You At An Increased Risk For Polio. Id like to point out that every word in that title was capitalized, because apparently that makes articles more appealing (and in my infinite wisdom or lack thereof it sure made the article much more appealing to me). So I clicked on the article, a snack of organic carrots at my side. I thought I would be able to prevent all polio worldwide just by looking at this one article. It actually looked somewhat legitimate, with real facts and figures, but I mean, I was 9. I was looking for the answer to my question about how to prevent polio. So I scrolled right to the bottom, and find an article that shows a direct correlation between polio and much to my dismay organic foods. So I stare at my snack of carrots with this look of what on earth did you DO? as if the carrots knew anything and they were supposed to stop being polio-inducing just from my stare. I didnt look at organic foods for a few years after that, until eventually and Im sorry if this is anticlimactic I just forgot about the whole polio thing and ate what I normally do. But then I got to thinking about it again recently; how did they derive the idea that carrots and polio were even tangentially related? After doing some research, I found that all they did was show that from the period 2005-2010, polio and organic food consumption rose at a similar rate. That was literally the whole thing. I could switch out polio with the U.S deficit, post that online, and some kid would probably find it and swear off anything organic in hopes of saving the U.S economy. Then as the debt level slowly rose even though he was keeping away from organic foods, the kid would grow up with a general sadness inside, like did I not do enough? But thats another story.Famous author Mark Twain once said there are three kinds of lies: Lies, damned lies, and statistics. I fell victim to the third of those lies, and was duped into legitimate fear by some misused numbers on the internet. But its not just me, and its not just people my age. In tons of places in the media or in everyday life, someone is either misusing statistics or just making them up to gain some sort of advantage. Obviously, some things are done just as a joke, like if I told you that 7% of dolphins are known to practice Judaism. But a lot of the fake or misused numbers out there arent made for fun. Theyre made to intentionally mislead people to gain an advantage. Today, well explore three different topics regarding the misuse of statistics. First, well discuss the nature of the misuse of statistics. Second, well examine what kinds of impacts come with the misuse of statistics. Finally, well explore how and why the problem should be controlled, and realize that both the fabricator and the audience have to work towards mitigating the problem.

When I found out that an innocent snack that was supposed to be keeping me healthy was actually increasing my risk for one of the deadliest diseases in the world, a few things happened. First, I started to gain this increasing distrust for everything my mother had ever told me to do. But more importantly, I started to pay closer attention to how I and those around me were being fooled by faulty uses of statistics. And this faulty misuse is actually a much more prevalent problem than we realize. According to the Wharton School of Business, there are several different types of flaws that so-called condensed data that is, people mining through countless numbers from different sources to produce nicer, more clear numbers can have after being released to the public. One of the biggest flaws that were found is that companies often try to look for evidence in their sales data for savings trends. For example, when Geico says it can save you 15% or more on car insurance, that just means that someone found a trend in which a large number of Geico recipients saved 15% or more on Geico products. But if said examiner of data looked at 30 different regional samples of Geico savings and found savings of 15% or more in 8-10 of them, the company can still claim that it COULD save you 15% on car insurance because it COULD. Another problem that was brought up in the study was the idea of detached statistics. When Geico says it can save you 15% on car insurance, what is it being compared to? One random insurance company? Every other insurance company? If that were really true, how could any other insurance company be successful There has to be some sort of knowledge gap between the producer and the consumer for this relationship to work. Finally, there is the problem of totally factual data that doesnt quite get translated correctly to visual representation. Fox News, the worlds preferred source for in-depth reporting, once showed a graph comparing Romneys poll ratings to president Obamas poll ratings. In this particular poll, Obama had a polling of 44%, whereas Romneys was 47%. This seems like accurate data, but on the bar graph representing the data, Obamas bar was tiny almost nonexistent and Romneys bar was huge. When looking more closely at the X axis of the graph, however, you will find that the graph started at 43% and went up in huge increments for every additional percent. The representation of the data wasnt wrong if you looked closely enough, but for people who understand visual representations of data more easily for example, all the people who watch Fox News the wrong message could definitely be sent. Finding out that organic food consumption led to polio may have given me a scare, and I may give a scare to some nave kid out there if I ever decide to replace polio with the US deficit. But eventually, like I did, the kid will grow out of it and discover that the scare wasnt really all that scary. What is scary, though, is how much bigger the effects of misconstrued data can be on other, more mature people who fall for more elaborate, more believable, and more impactful traps. Sometimes, broad statistics that describe large portions of the population can be misused, and that means, according to journalist Alex Perry of Time Magazine, that we get the world wrong. Take the popular statistic, for example, that only one in ten Americans have a passport. For a person living in America, this may not seem like a huge deal. But people outside the United States use this data to suggest that Americans are isolated and care too much about their own country to care about any other country. But it turns out that the statistic is derived not from how many Americans possess passports, but rather, how many of those passports had been used in the last five years. When we look at a different statistic of the same situation the fact that the US is the third most involved country in international travel, behind Britain and Germany, the situation looks different all of a sudden. Thats something that indirectly affects everyone who lives in the U.S, but there is also some statistical manipulation that hurts people on a smaller scale. According to the American Statistical Association in August 2012, one statistic has led many to hold back on one of the biggest decisions possible: the decision to have a child. According to the study, a CNBC article finds that the cost of raising a child from birth to age 17 has surged 25 percent over the last 10 years. 25% more money for something that is already so expensive could be very consequential for many Americans, especially the ones going through hard financial times. Except for one thing: the problem isnt nearly as bad as it looks. The Association goes on to explain that this rate is equivalent to about a 2.25% yearly increase which is almost the same as the 2.26% annual overall inflation in the United States over the same period. The fact that the original statistic was stated in such a drastic way, with an extreme word like surged, communicates the idea that this is an intense problem that needs attention from a reader. And no matter what adverse impact that information might have on the reader, the use of extreme language is necessary to attract people to the site. And a Department of Agriculture study with the same information didnt use that language, because it didnt need to attract people to read its study; anything that in-depth is meant for someone who is doing some serious studying on the topic. But for popular appeal? That kind of over-dramatization is not only necessary, but effective as well.Finally, we need to think about how we can solve the problem, so that any type of person, ranging from the little kid who develops an irrational fear of carrots, to the woman who decides she just cant take the burden of having a family, doesnt have to go through agony just because of some misused numbers. The first thing we need to do, as a society, is to think a bit more about the information we see on a daily basis. For example, every single thing I said in this speech could have been totally made up. The names, the dates, the statistics especially the statistics but you wouldnt try to verify my statements. Why? Primarily, because you wouldnt feel the need to, because, at least for the time being, I havent given you any direct reason to doubt me. (Well, I did just now, but lets disregard that for a minute). But lets think back to that CNBC article. It seems, for a second, that the writers of this article have not given us any reason to doubt them either. But this CNBC source is the only source in which the statistic is written in this way, and even though similar information is discussed in other articles and studies, it is never phrased in such a way that makes the issue look more extreme than it really is, in order to increase readership. As the public, the people who stand the most to lose collectively from being fooled by articles with fancy titles and extreme wording, we need to stat keeping an eye out for those sources of information which clearly seek to espouse an agenda if any kind or push a certain action, without giving us all the details. Moreover, for every statistic we see, we need to think about its counter-statistic. The Darton College Center for Math notes that certain statistics like this yogurt is 75% fat free can also be understood to mean that they contain 25% fat, which, for someone on a diet can still be a bad thing. Darton also points out that looking for detached statistics such as This yogurt has 33% more nutrients is important as well, because the 33% isnt being compared to anything outright. Sometimes, there will be a small asterisk by the 33% statistic, and at the bottom of the container, it will say versus the leading yogurt brand, another piece of information which is also very uninformative. We shouldnt necessarily stay away from products that have such information on their labels, because more nutrients probably wont be a bad thing. But if we want to know exactly what we are dealing with, its important to be on the lookout for the full picture of the information that comes at us every single day, and statistics, in all their imperfections, can never provide that to us if we merely glance at them without any extra thought.Misrepresentation of data is an often neglected problem, because we see so much seemingly harmless data out there that it never occurs to us that a lot of it may not be what it seems. But by implementing the aforementioned solutions, we can see not only when a certain problem is a lot worse than it looks, but also when its not quite as bad as it seems. We can enjoy any organic foods without having to face deadly disease, and our guard will be kept up for things that we actually need to pay attention to.