Using the right set of data, I quite often argue statistics could prove’ black is really white. The number of times I hear people quip “statistics prove” such a statement is as meaningless as the point they are trying to prove using such a baseless justification. This is not to say that statistics are useless. They have a purpose and those who choose to ignore them do so at their own peril however, without fully understanding the meaning of data collected, the information presented or conclusions drawn are just as meaningless.
A recent poll in my local area regarding recycled water misused statistics in the worse possible manner simply by the way in which the question was framed. The campaign against the issue played an emotional and fact-starved game of ignorance through suggestions of recycled water being poo’ water a crass and fundamentally flawed method of disinformation which ignored a number of facts supporting the other side. The inevitable outcome was a sound defeat of the proposal to recycle water and this came about simply by skewing the results through phrasing a number of questions designed to extract the desired outcome. So the polls suggested that statistically, the majority did not wish to proceed with a water recycling program yet the true result may have been different had the facts been more effectively presented, and the questions phrased without the emotive and crass exaggerations by the opposition.
On the other hand, a simple average can often be used as a means of drawing accurate conclusions, for example by placing two brands of soft drink on the shelf and recording sales performance, it is relatively easy to determine which one is preferred – simple average. From this information, the retailer (if he or she is sensible) would ensure the preferred brand is available in higher numbers.
The more information gathered over a longer time will produce even more accurate result using simple average as your method. What I mean is a number of variables will often skew the results and provide misleading information when calculating a simple average. For example, a convoy of tourist busses may come past and stop to shop during the period over which the data is collected. The people from a different area might prefer brand X, while the locals may have a distinct preference for brand Y, so a relatively short period used to collect data will clearly not reflect what is happening. Based on such a short period of time, the data used to determine which brand should be ordered at a greater quantity came collected when the tour busses happened to stop would not be a reliable guide for the retailer ordering supplies. If however the data were collected objectively over a two year period (or greater), even more useful information might be extracted from using the simple average method such as: at a certain time of the year, brand X sells higher than brand Y. The smart retailer will ensure stock is available accordingly. This is a fairly simplistic example of how statistical information can be used effectively.
When presented with statistical facts’, you should ask yourself many questions regarding the source of the information. Consider the following:
– Where did the information come from?
– Under what circumstances and for how long was it collected?
– What (if any) are the interests of parties presenting the information?
– How was/were the question(s) framed?
– Who was asked, and when?
– Who can you contact to verify the results independantly?
– What information was disregarded, and for what reason?
– What was the purpose of gathering the information in the first place?
These are only a few questions that I hope might provoke some thought and comment. So, next time someone comes up to you and argues with statistics which prove’ their point of view, start thinking about exactly what they are trying to tell you before blindly going along with what you are told.
Have a good day folks.
Statistics prove that most of you who are reading this will in fact have an excellent day – if not today, some other day perhaps!
