Lies, Damn Lies and Statistics

The phrase used above was first popularized by Mark Twain. It questions the use of statistics by questioning it's validity. Especially when it supports an argument that seems weak. 

It is damning because people put a degree of faith in statistics. The reason for this is because statistics are factual in nature. The way the figures are presented appears to be cold facts as opposed to opinions. At best, we see statistics as representing and reporting on data that have been processed by more knowledgeable people, therefore more trustworthy. But if examine closely this relationship, there lies the flaw. When we trust statistics, we are really trusting the people who came up with them. But doesn't that mean when someone is quoting statistics, its really hearsay? Ouch.

Basically yes. Don't assume hearsay is bad although the word is usually used negatively in courts. There are ways to figure out if statistics quoted to us really mean what they say. It involves asking hard questions like:

1. How did the researchers define the population? Who did they ask?  How did they get the information? The source of data is called the population. Take a look at the raw data or find out more about the sample taken. The simplest form of data manipulation is to omit sources of data that you know will give the result you don't want. 
2. How was the data collected?  Find out how the data was processed and the statistics came about. Pay attention to when data is discarded or summarized multiple times. For opinion polls, the way questions are asked and in what order can affect the answer given. 
3. Who published these statistics? Find out more about the organisation that published the statistics. With Google search, this is a simple matter. There are unfortunately organisations that have acedmic-sounding names and structure but are committed to forwarding a narrow agenda or have purely commercial interest in mind. It's an inside joke in the academic community that you can find a professor somewhere that agrees with you.

One rule to always remember is "Just because is sounds true, it doesn't mean it's true".

Continued on Lies, Damn Lies and Statistics 2