Not seeing a Scroll to Top Button? Go to our FAQ page for more info.

Lies, Damn Lies and Statistics



Lies, damn lies and statistics is a term that was popularised by Mark Twain, who attributed it to the British Prime Minister Benjamin Disraeli.  It was my university statistics course professor’s favourite saying. And it is a very fair way to summarise how the Australian press and pundits have presented the results of the 2016 census.  

A bit of background on the census for those of you who don’t live in Australia.  Australia does a census every five years, with the first conducted in 1911.  It is mandatory by law to complete your census on time, and the government can impose a fine if you do not follow the rules.  The census is online, and you need to fill it in with the data that is valid on a specific date, in 2016 this was the 9th of August. The date is critical as you need to complete the survey as of the facts on that day.  For example, if you have ten people staying with you on the day of the 9th of August, all ten people need to be included in your census report you lodge with the Australian Bureau of Statistics (ABS).  An incentive to not have too many guests on the day of the census.

At the end of June 2017, the results were available for all to review, comment and draw conclusions.  While the actual data collection process had many IT issues, which led to the government extending the deadline that individuals had to complete the form, the summary process appears to be very efficient.  In less than nine months over 24m people worth of data was checked, indexed and made ready for all to see on the ABS website. The results are an incredible set of data for anyone to use.  Data to contribute to making decisions regarding products, services, markets and consumers.  On the ABS website, there is a wealth of information to underpin decisions in commercial businesses, NFP’s and government organisations.

For example, the ABS kindly does “averages” for anyone who wants to read them.  Here is their summary of the typical Australian (if there is such a thing).  And for my local area, if I wanted to open a restaurant, I could easily find target market data.  For the area I live, there are 21,500 people and I can analysis this population by their average income, do they rent or own (with or without the bank), how many are men or women and their average age, just to name some of the stats available to all.  A gold mine of information for anyone wanting to set up or expand a business.

And of course the press did their only analysis of the data and in some cases produced what would fit well with Mark Twain’s quote.  For example, one of the questions asked is “what religion you align yourself?”  There are many choices, as you would suspect, but the headlines from one of the popular press read “Muslim religion in Australia up 77%”.  While the up 77% is technically accurate, it takes Islam up to 2.6% of the population.  As everyone knows who has done anything with numbers if you are starting with a small base, it does not take much to get a double digit increase. 

The above is a good example of how it pays, whether you are looking at census data or any numbers in your business, to understand the real relevance and materiality of the statistics quoted.  And to be sure you know how did they calculate the result, what was the starting point and were they consistent year to year in the methodology?  These and other questions should be asked when reporting a finding or approving a financially based decision.

The Guardian did a summary of what they saw as the key findings, taking away some of the misrepresentation that can occur from the “maths”.  Their review is worth reading, and it has several good examples of how numbers can be manipulated to tell the story you want to tell.  This was my statistics professor’s favourite point when lecturing on the use of stats.  A good statistician can make numbers prove almost any point he or she wishes to prove or disprove. 

The Australian census is an excellent case study in collecting and summarising data and then making the results available for everyone to use as an “open source”.  And it is an example which can be translated into a business context of how statistics can cause misleading headlines and potentially making the wrong decisions. 

As HR Professionals we really need to be aware of what story we are telling with the people related data and statics we have in our organisations.  We don't want to be the newspaper head lines that raise alarm bells on percentage increases in something if that is not material or relevant to the business.  All good things to think about next time you review a set of figures or are asked to report on some data regarding people in your business.

Article by Mary Sue Rogers

Picture Credit - First Dog on the Moon - The Guardian

 

Posted On : 11-07-17

Commented by:  Gary Kendrick

So True... summarised by a statement " we have increased our turnover by 100% year on year since inception" = last year we set up the business and our turnover was $1,000 this year it is $2,000 = not as impressive. However maybe with a little context like the business was a lemonade stall outside the front of my house run by my three children and the net profit was 90%, looks pretty entrepreneurial.

Posted On : 30-07-17

Commented by:  David Ellis

An interesting read - thank you, Mary Sue. I am not a statistician but I worked on the Census in 2006 and 2016. As you might expect, many more people took the online option in 2016 than in 2006! It was interesting to note that once it was publicised that the census had IT problems (some people had tried to upload their response on 9th August but been unable to), people's level of support and cooperation declined markedly. Perhaps they felt that once they had tried, their obligation was partly or wholly satisfied and they should be off the hook! Or maybe it's that everyone loves a winner and no time for losers!

Posted On : 27-07-17

Leave a Comment