By Ed McManus, ZBN Contributing Editor

Book:  “The Numbers Game” by Michael Blastland and Andrew Dilnot
(2009, Gotham Books, hardcover)

We’ve heard the old saws: “Figures don’t lie, but liars figure.”  And – my favorite: “There are three degrees of lies: Lies, Damned Lies, and Statistics.”

Now, courtesy of Messrs. Blastland (a Cambridge English graduate and BBC personality), and Dilnot (economist and British TV host), we have a book to help steer our way through the myriad numbers that are thrown at us each day.  How can we tell the real from the slanted and the false?  The authors suggest we know more than we think we do, and we just have to ask a few basic questions and think it through:

1. As children, we count with beads and fingers.  As adults we count with “strawberry jam,” and it goes all over the place.  In 1970, for example, the U.S. Census Bureau reported there were 106,000 centenarians in the country.  Very impressive, but they later determined that number to be less than 5,000.  Why the big difference?  The raw data was unverified and the sources spread out all over the government in Social Security offices, military and voting archives, Medicare reports, and old birth records.  Then, the seniors said how they felt, or maybe they wanted a little attention, or they were genuinely confused.
2. Governments, industry and the media hurl impressive sounding numbers at us all the time.  But, are they really big numbers?  In the UK, for example, the British Labour government said it would spend L300 million (pounds or about $600 million when the book was written) over five years to create a million new childcare places.  That sounds wonderful.  But, if you divide that L300 million by the million places to be created, it comes out to L300 or L50 each year for the five years.  Run that across 52 weeks and you get L1.15 ($2.30) per child per week.  Can you get quality childcare for $2.30 a week?  We think not.  In this case, L300 million is not a big number. 
3. In Britain, a group of concerned parents tore down a radio tower because there were nine cases of cancer nearby. Surely the cancer was caused by the tower’s microwave signal.  There were no other reasons to suspect the tower.  Statistically, this could be a Cluster Effect.  If you stand in the center of a room and throw handfuls of rice over your head, you will see interesting patterns develop.  Some areas of your rug will be evenly covered; other patches will be bare, and in a space or two there will be a heavy concentration of rice.  This is just the way random distribution works.  The authors show a convenient and less messy way to verify this effect by flipping coins or turning shuffled cards to see the color distribution.  Is this too a Cluster Effect?  Maybe there’s no tiger, just stripes - and chance.  We have to look at all other possible reasons for an event before we decide on true cause and effect.

The authors also discuss the placebo effect on medical trials, including the fact that we rely on trend studies to explain why numbers go up and down.  Often, they just do; it’s in the nature of numbers.  As for averages, the average color of the rainbow is white.  Does that help you understand rainbows?

Part of our problem, is that before we can accept numbers we have to insure we have all of the numbers for all dimensions of the issue.  We must have the nitty gritty about who collected them, the procedure used (shortcuts included), the number takers interest in the outcome, and what other alternatives exist.  The authors cite a British Police Federation report of a young man who went door-to-door collecting sponsorship money – and then stole it.  They nabbed him, and then had to decide if he had committed one crime or dozens.  They decided to go with dozens.  The police records now looked excellent from a crime clearing standpoint, but some other group, unaware of the decision, might be looking at a startling increase in the crime rate.

The authors use these and many other amusing and instructive examples of how we can understand and identify the accuracy, if not the veracity, of the data barrage to which we are subjected.  The book is rich and informative reading.

The authors leave us with several useful thoughts:  “Numbers are tidy – life isn’t,” “Up and down happens,” “Stripes aren’t tigers,” “You can’t see the wholes through keyholes,” and this classic, “Numbers are pure and true.  Counting never is.”