… and if it doesn’t move? Measure it until it does!

Statistics are all around us and I got to thinking, how much do we depend on analysis and are there times where they are just not worth paying attention to?

Before I continue, and at risk of a contract being taken out on my head by an axe wielding numberjack, I do think that some of the data analytics that are performed by our own Business Intelligence department (who have the coolest department name in our company) not only add value to our service delivery but they also help our clients understand their business better than they ever have before thus giving them the ability to perform better than they have ever dreamt of. From my point of view this type of reporting and analysis is very valuable and I rely on good data to make relevant decisions in the workplace all the time.

In the workplace we can use statistics & analysis to help us understand specific trends, patterns, cycles, behaviours etc. Where statistics start to lose credibility for me is when they are presented with a specific bias in mind. I recall the itv news at ten showing, with the same snappy graphics as they use for their ‘Happiness Index’, the shift in voter faith in the coalition government and other related questions. The part that struck me wasn’t the message that they were driving home; it was the percentage of people who had answered “I don’t know”. On some questions this category made up almost half of the responses but clearly it didn’t hit home a hard message that indifference rules, instead they opted to focus on something that would either create a skewed positive or negative message to make it news worthy.

In the recent Oldham East & Saddleworth by-election each political party relied heavily on trends and past statistics to get their excuses in early. On the lead up to the election each party had already peddled their statistic driven reasons for potential failure either by referring to tactical voting or previous outcomes where the sitting party ‘never wins’.

This also happens in televised sport, in the pre-match build up a barrage of statistics will be shown of past form in previous encounters. Often these mean nothing but at best it gives the pundit something other than idle banter (or sexist comments allegedly from Andy Gray & Richard Keys) in the low points of the game.

I have already had a pop at the media in another blog posting so, in an attempt to provide other examples, I will move to advertising. If you haven’t done so already, the next time you see an advert for a new shampoo or cosmetics line coming on the television or in a magazine or newspaper, take a look at the small print which will tell you the actual percentage of the people that agreed with the fact that the shampoo is the best around. More importantly it will also refer to the sample size. In some cases I would be better asking my extended family for their opinions…at least I could trust them to tell me the truth and I would have more respondents. Granted, a positive response from the general public to a survey can offer backup to the claims being made but when the advert on the TV for shampoo sees the model with hair extensions, really how much faith do they  have in their product?

As we move more and more towards social media as a driving force for public opinion I often wonder whether the rating that something has received has been a push-up by the creator/owner/business and often I am left still making up my own mind…

I would really like to hear your opinions on this topic. Do you trust and rely on statistics in your professional and personal life in the same way? Or do you think that a statistician is just a person who draws a mathematically precise line from an unwarranted assumption to a foregone conclusion?

Answers on a correlation diagram please.