Pages

Thursday, March 6, 2014

My Take on ABC News Article "State Fat Index Crowns New Heavyweight Champ"

Let me start with a little disclaimer. I put this together during my free time at work. So there are probably some spelling and/or grammar mistakes. Possibly even a calculation error! I'm also a very technical oriented person. So some things in here may get a bit technical and dig into the details a bit. With that all being said, here we go...




My wife send me this article on Facebook titled State Fat Index Crowns New Heavyweight Champ. She sent it to me to identify states that we are not allowed to move to and promoting that New York is a less obese state than most of the other states. So I read it as I do all of my wife's posts (I do, I really do) and have a few comments on it.

Gallup has compiled this list every year since 2008 by asking a random sample of 178,072 adults about their height and weight...
  • Random sample? From my experiences with data, random sampling does not produce great results.
  • Asking? Problem with this is what people lie and sometimes people just don't know. The results, bad data.
  • 178,072 adults? What was the selection criteria for these adults? Is this enough data?
  • I guess getting data from doctors is difficult and would be cost prohibitive.


Anyone with a BMI of over 30 is classified as obese.
I'm sure that I am not the only person that questions the use of BMI as a valid method of determining obesity.


...the national obesity rate edged up to 27.1 percent from 26.2 percent.
More than 25% of Americans are obese. This is just terrible. Simply terrible.


Eleven States With Highest Obesity Rates:
I took a screenshot and counted, up to 10 not eleven. Just in case anyone thought I was nuts.

If you count the number of states listed, there are 10. I guess the editors at ABC News missed this. Maybe I should be an editor for ABC News. Hmmm...

But regardless of all these comments at this point in the article, I continued. But at the end of the article I felt like there was more to this story. I didn't feel that this was good enough. Being in the data / IT industry professionally, I naturally wanted more details. So I looked and found the source article; Mississippians Most Obese, Montanans Least Obese from Gallup. This article lists out the obesity rate details for each US state. Exciting! More input! (Yes, a Short Circuit reference.) I also gathered United States population details from the US Census from July 1, 2013.


So I decided to combine the data and look into it a bit myself, well, because I'm nuts.


The first thing that I noticed about the ABC News Article, is that they got 2 of the states obesity rates incorrect! See screenshot below for details! Sure the differences are minor and will not significantly affect any of the calculations, but it should be a direct copy from the source article! (There is a slight possibility that the source article changed, but I doubt it.)




The Census data from 2013 does not distinguish between adults, children, infants, etc. But the Gallup obesity rate data is exclusively for adults. A bit of a problem. So I have analyzed the data in detail by removing the births from 2010-2013 to see if that change would significantly affect my analysis steps. In comparing these individual results, there was an insignificant difference between the results, so I am going to use the full population numbers (not removing births) from my analysis. I am going to be looking at ratios/percentages, so this may work itself out. Math! In addition, there was no data from the Gallup article detailing the obesity rate (% obese) for Washington DC. So I have excluded this state (not a state, as someone has kindly pointed out) 
Washington DC from my analysis.

Here are the fields that I put into the final spreadsheet with some details as to what each of the fields represent.


  • Article Rank - the rank that I interpreted from the articles.
  • Dense  Rank / Tie - ranking based on ties for matching %'s.
  • Actual Rank - the rank based on my compilation and sorting, by % then population.
  • % - sourced from the Gallup article.
  • State
  • Population - 2013 population from the Census website.
  • # of Obese - calculation of the total number of people considered obese; population * %.
  • # Times Greater - for that state, the number of times greater it is than the smallest state's obesity number; # of Obese / minimum # of Obese.
  • % of Total National Obese - the percentage of the total national obese; # of Obese / Sum of Total # of Obese.
  • % Obese of Total Population - the percentage obese people of the total US population; # of Obese / Total US Population.

I am not trying to downplay the obesity rate per state. It is important and scary. But I think that there are other conclusions that can be drawn from this data as well. Here is a summary of my conclusions based on the data.
  • The Actual State Rank differs from the assumed Rank listed in the Gallup and ABC News articles. Now when I read this article, I was happy to see that New York was on the lower side. But if you look at the title of the section it reads 10 States With Lowest Obesity Rates, not the 10 States With the Lowest Obesity Rates. What a mis-representation of data! A bait and switch! Take a look at the following picture! New York is ranked the 20th most obese state out of 50 in terms of percentage! Oh my! Not as good as I originally thought! But I understand why they did this; New York is, well, New York. How could New York not be represented in the top or bottom 10?
    • Now how about Utah? Utah has an obesity rate of 24.2%, over 4% better than New York? Come on now!

  • You can also look at the following bits of criteria against New York; 28.1%. Both show that New York is worse than the average / midpoint.
    • Percentage span, 35.4% - 19.6%; 27.5%.
    • Median / Average; 27.1%.
  • In addition, New York then ranks 3rd based on the following pieces of calculated criteria.
    • # of Obese - sure New York also has a high population, so I do not hold this bit of criteria too problematic.
    • # Times Greater (than the lowest state, Montana); 35.
    • % of Total National Obese; 6.5%.
    • % Obese of Total Population; 1.8%.

Embedded below is the Google Spreadsheet that I made. I have also shared this document via Google Drive Spreadsheets which will allow you sort the data.




I personally think that this is a bad article from ABC News and Gallup. It provides a misrepresentation of the results and an incomplete summary of the data. Hopefully this sheds some light on to the data behind this 

As always, I look forward to feedback on what has been written here! Enjoy!

Have I made a mistake? Did I interpret something incorrectly? Let me know!

As a side note, this comment is the ABC News article is amazing.
If you argue against math, you will lose.

No comments:

Post a Comment