The law (and danger) of averages

February 5, 2014 | By | Add a Comment

“Just think about the average, what use have they for you?” (Rush, “2112”)weighted_average_symbol

I was at a presentation a while ago, the subject of which was economic data.

“Of course, the Central Coast has a higher than average aging population”, we were informed. Which made me smile, because not 10 minutes before, a colleague had commented to me (with actions), while we were discussing the danger of averages:

“I’ve got one hand in the fire, and one foot in the freezer, but on average my temperature is normal…”

Therein lies the problem. As I wrote in a previous post, the Sydney Morning Herald committed this sin when deducing that Central Coast workers were, in effect, lazy (or less hard-working) than workers in the Eastern Suburbs of Sydney, based on average weekly working hours for Potts Point compared to Central Cost suburbs such as South Kincumber and Patonga. The fact that the latter are predominantly retirement communities was completely missed by the journalist.

And similarly, with respect to the oft-repeated description of the Central Coast being “God’s waiting room”, a bit of analysis finds that, well, it depends

Drilling down on suburbs reveals dramatically different demographies, with some communities having a huge proportion of kids under 15 and a very low proportion of over 65s, for example.

So context is everything. While the media in particular use averages as a way of proving a point they know they want to make, a more pragmatic and objective approach is to recognise the potential for ‘judgemental bias’, particularly in the use of averages.

Filed in: General

Patrick Spedding

About the Author (Author Profile)

Patrick Spedding is Senior Director of BI R&D for Rocket Software, and IBM Champion for IBM Collaboration Solutions. He is also a Non-Executive Director on the Board of Eastside Radio in Sydney, Australia. Prior roles include Director of Product Management for IBM Cognos, Director of Field Marketing for Cognos, Founder of Tableau partner See-Change Solutions, and SAS Solution Manager for BI and Strategy Management. Patrick's qualifications include an MBA degree in Marketing (AIU), Diploma in Management (University of Michigan), BSc (Hons) in Mathematics (Loughborough University, UK), Fellow of the Australian Institute of Management (FAIM), and member of the Australian Institute of Company Directors (AICD). Find Patrick on Google+

Leave a Reply

Trackback URL | RSS Feed for This Entry