The Statistical Shell Game or How To Spin A Chart

There are three kinds of lies—-lies, damned lies and statisics. (Mark Twain)

I am not a big fan of charts or statistics because they can manipulate data and distort the truth. I’d prefer to see the raw data. Yet, some insist on parading them to support their point of view. Here’s an example of statistical manipulation of the following raw data: 14, 15, 20, 22, 1,000, 2,000 3,500. The median is 22, the average is 935. Depending how I want to spin the story, I’ll present the median or the average. Another example: An increase of 1 can be a 100% gain (from 1 to 2) or a 1% gain (from 100 to 101). Statistics and charts can be an expert’s shell game.

Here’s the latest illustration.

This chart shows that the S&P 500 follows the National Association of Home Builders (NAHB) Index. This chart would say the S&P 500 is headed for a crash.

But wait. Before you sell your stock, examine th chart below. It shows that the NAHB Index has no effect on the S&P or a slight negative correlation, i.e. that they move in opposite directions. Depending on which chart an expert shows you, you will get a completely opposite point of view. Damned lies!

Source: Business Week online

sellsius°

Great saying John. I am filing it in my "sayings to remember" folder.

John Occhi

Great Post and so true. When I first got into business, back in the 1970's, I had a 80 year old mentor who still worked everyday. At the risk of being politically incorrect, this guy was the sterotypical old jewish business man (not a bad thing) - his name was Seymour.

In any event, Seymour tought me that "Figures don't lie...Liars figure"

Michael

Exactly!!! Just like RealtyTrac is selling subscriptions by Paul Rever-ing "Foreclosures are coming, Foreclosures are coming!!!" Yup, in one of our neighboring towns, they actually doubled this year (went up 100%), from 1 in 2005 to 2 in 2006!!! Watch out world (and put your credit card number in the spaces provided by RealtyTrac - sucker!). md

Larry Cragun

My contention is that the bubble lovers take a stat and make their point, even if it skewed or not relevant. Lar

sellsius°

Thanks anonymouse (that's funny)

anonymouse

link to the paper:

http://urbanpolicy.berkeley.edu/pdf/CQSAdvMacro...

sellsius°

Thank you for the comment Pat. I am interesting in reading the Case paper. Do have a link?

Pat Kitano

Great find Sellsius... I had just read Peter Coy's gloomy 9/19 "Can Wall Street Withstand Weak Housing?" Bus Week article, and hadn't seen his retort above on 9/26... granted, the two charts are tracking two different decades and had different correlations, but Merrill, who published the scary chart, had the duty Not to sensationalize by showing two decades worth of data.

There's now a third Bus Week article dated 10/2 - Stocks Can Handle The Housing Chill - that rationalizes the new view that the scary Merrill chart might be just a facade:

"Bears worry that a housing downturn will hurt because it accounted for such a big share of economic growth over the past five years. The real estate boom created tons of jobs in construction, sales, lending, and related industries. Plus, it gave consumers billions in extra spending power through cash-out refinancing, home-equity loans, and the like.

But the employment hit in homebuilding is being offset by growth in nonresidential construction. As for the hit to consumption, there's no sign of it yet, and there may never be much of one. Americans may just continue to spend even though their houses are worth less. A 2005 research paper by economists Karl E. Case of Wellesley College, John M. Quigley of the University of California at Berkeley, and Robert J. Shiller of Yale University concludes that while increases in housing wealth spur consumer spending, surprisingly, "declines in housing market wealth have no effect at all upon consumption."

--pk

Sellsius real estate marketing blog

The Statistical Shell Game or How To Spin A Chart