Statistics are B.S. and I love Pizza

Back in 2007 I was sent to Australia on business. Throughout the year I was there, I was constantly bombarded with comments from Australians, how they love America, but they hate President Bush (no surprise there). Well, after recently getting married this past March, my new wife and I went on our honeymoon to Australia. I was happy to see the "I hate Bush" comments were replaced by much more positive comments about our current President Obama. Believe me I realize this is completely anecdotal, but interesting nonetheless. I recently came across an article entitled "Obama's popularity abroad on the wane, survey finds", and due to my recent experience in Australia I quickly clicked in. In paragraph 4, I read the statement:

"Approval ratings for the original pool of 116 countries, meanwhile, declined from 47 percent to 43 percent between 2010 and 2011."

This sounded like a tiny margin, so I was curious to see where the data came form. I then googled "Gallup Presidential Ratings", and after a few clicks I came across the source of the article's data at the following link: "U.S. Leadership Losing Some Status in Key Countries". I wanted to see their polling methodology, and as I scrolled down to the section entitled "Survey Methods" I read the following:

For results based on the total samples, one can say with 95% confidence that the maximum margin of sampling error ranges from ±1.7 percentage points to ±5.7 percentage points.

First of all, WTF does that even mean, with 95% confidence they still give a range? So 47% approval rating in 2010 minus 43% approval rating in 2011 equals 4%. So based on Gallup's methodology I can say with 95% certainty, that this falls within the uncertainty range of the survey, therefore the results are statistically insignificant, and yet you still have such a strong headline about Obama's popularity waning. As a side note, I am not even an Obama fan, yet this still bothers me tremendously. Even more disturbing are the anti-Obama comments on the article, and how angry they are that his popularity is waning abroad and "he needs to go!". All this over data that is clearly statistically insignificant.

For those of you who know me, I eat a ton of pizza. It's my favorite food. A close friend of mine, who is a total health nut sent me an article with the following headline: "Study Identifies Link Between Depression and Certain Foods", pizza being one of the foods. I laughed. Being that this person was a very healthy person, who works out, and eats a healthy diet, high in protein, and low in carbs, I decided to google, "high protein diet depression". Low and behold I came across the following article: "Does a High-Protein Diet Cause Depression?"and quickly sent it back. Her response was "Hahaha. Touché!".

So what's the point I am getting at? If you haven't figured it out already, I have come to the following conclusions. You can make statistics do and say whatever you want. It's all about changing the lenses you use to look at the data. Secondly, when reading an article with such a strong headline, before taking it literally, do your homework! Find out the sources of the data, look at the levels of uncertainty, look at the credibility of the source. Mainstream media is always gonna inundate us with loads of stats., but it's our job to be smart and figure out for ourselves what's legit, and what's BS.