Food for Thought "If you have the right people, with good, basic values and good work ethic, you can have a tremendous journey."

Buy Now
Food for Thought 2 "The follow-up in my Food for Thought series, with more focus on my experiences with Six Sigma and Kaizen."

Buy Now: Paper Back
Buy Now: E-Book

February 26, 2016

The Assumption of Accuracy : Statistics and Social Media

Filed under: Food for thought for friends — Alec @ 4:33 pm


Of course you don’t and you should not. I don’t believe yours either; I don’t believe anybody has a high degree of accuracy.

Having 45 credits of Calculus, I can prove 1 equals zero. Having a masters in Statistics and also in teaching Statistics for the last 45 years, I know that ALL data is just the opinion of the person that collected the data. And maybe not even an accurate opinion.

You ask me then “Why do I bother to post data?” Because everyone else is posting data with the assumption that it is accurate. Neither are accurate. If my data contradicts your data, you may begin to question all data and sources.

Example, unemployment is a totally misleading number. Last time I checked it was the number of people between 18 and 55 that are on unemployment. In other words, that number is divided by the number of people collecting full time pay checks. So, if you have run out of unemployment and stopped looking for a job, you are not counted. If you are getting paid cash, you are not counted. If you are a small business owner and not getting a regular check, you are not counted. Not working full time? Not fully counted. If over 55 or under 18? Not counted. Illegal? Not counted. A lot of categories remain uncounted.   

Political polls are just as inaccurate. They are based on a series of assumptions:

  1. People will give you an honest answer about who they support
  2. That the person conduction the poll asks unbiased questions
  3. The person polled is a registered, potential voter
  4. That the poll is truly a random poll — it is very difficult to really randomly poll the population!

The best any poll or chart might show with a certain degree of confidence (i.e. 90%, based on how the data was collected) is that this is our trend at this specific time. And if the poll is not random? The data is totally inaccurate. And tomorrow or yesterday? The data collected could be totally different. 

So…I don’t believe your data, and I trust, you do not believe mine.

Please Note: I do not watch Fox or any TV news channel except CNBC Squawk Box. The only newspaper I read is the WSJ.  

No Comments

No comments yet.

RSS feed for comments on this post. TrackBack URL

Sorry, the comment form is closed at this time.

Powered by WordPress