Please login to the form below

Statistics and bikinis – 2 men united

I want to write today about 2 men that between them have reminded me of a fine line that we have to tread every time we analyse a dataset...
The first is Professor Aaron Levenstein. In truth there’s not too much to say about him, he was associate professor emeritus at Baruch College. He taught there for a grand total of 20 years, so I imagine that there are a lot of alumni of the college who have fond memories of him, if you happen to be reading this piece then let me know I’d like to hear what he was like as a man. His most widely published book is the well respected title “Escape to Freedom: The Story of the International Rescue Committee”. It’s a chronicle of some very dark times in the world’s history and yet it focusses on a positive outcome that is the IRC and the work it has done since its inception during WWII. I’m not about to review the book, there are plenty of other people who have done that. What has struck me is how he is remembered for a simple catchphrase, probably meant as a throwaway line, although it does emphasise a lot of the challenges we face when writing reports and interpreting research data.

Then there is my second man, Sam Warburton. He is currently preparing to knock heads with a number of Australians on the rugby pitch in Cardiff. Meaning no offence to Mr Warburton, he is a fine man and I respect his position as the captain of the Wales rugby team, but he doesn’t strike me as the type of man to be reading the work of an American business professor! Yet here he is misquoting Professor Levenstein’s line, “Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital”. 

It’s a fun line and I can easily understand why it has spread so widely. It also holds a lot of truth. Every time I sit down to analyse the results of a piece of research I am faced with a decision about what to include and what to exclude. Where do I put the emphasis and what do I need to caveat so that the limitations of the data are understood. What is it important to reveal and what should be left to modesty to remain under wraps.  There are some instances when the data only goes so far. For example a survey that shows that a brand has grown its market share, but in many cases observational data doesn’t clearly explain the ‘why’ of this event. It’s not intentionally hiding this information, rather, the data needs to be investigated further, or another study might need to be run to explore the why’s more explicitly. Also there are instances where the analyst needs to take a specific line to answer a specific question. If the research objective is to show the overall satisfaction with a brand vs. its competitor there’s no point focussing on the level of dis-satisfaction rather the emphasis should be on the positive numbers, and the negative numbers can easily be inferred i.e. if 20% of doctors are reporting a satisfaction level of 5 – 7 then you can easily see without having to spell it out that 80% of doctors are less satisfied with the brand. To some what is revealed is the most interesting point to make, to others perhaps it is the glimpse of something to be revealed that is the nugget that will uncover the real gems!

Research Partnership is one of the largest independent healthcare market research and consulting agencies in the world. Trusted partner to the global pharmaceutical industry, we use our expertise and experience to deliver intelligent, tailor-made solutions. We provide strategic recommendations that go beyond research, helping our clients to answer their fundamental business challenges. Find out more here: http://bit.ly/1F7l2AY

2nd April 2015

Share

Company Details

Research Partnership

+44 (0)20 8069 5000

Contact Website

Address:
Chester House
Fulham Green
81-83 Fulham High Street
London
SW6 3JW
United Kingdom