Articles & Blogs - US Commentary

User Rating: / 0
PoorBest 

“Lies, Damn Lies and Statistics” September 14, 2007

By Jeff Nielson

Don’t believe everything you read.

We have all heard the expression "statistics can be used to say anything." In fact, this is not true. A statistic is a numerical representation of an event, trend, or simply some data aggregate. As such, it can never be anything other than what it specifically represents - nothing more, nothing less.

We are a society of statistics. Statistics allow us to 'see the big picture'. They guide governments and corporations in their policies and decision-making, as well as the decision-making of each of us, as individuals.

Where statistics acquire their bad reputation is through the (faulty) analysis - or extrapolation - attached to statistics; or, through the use of statistics which are based on flawed data. In the case of the former, the examples of faulty analysis inevitably outnumber the instances of disciplined, accurate analysis by a factor of (at least) ten to one.

Our ineptitude in employing statistics – despite their prevalence and importance – is rooted in a number of causes. There are technical issues, which generally involve the complex, mathematical aspects of statistics, which are neither known nor understood by the vast majority (including many of the people creating and using statistics!). Primarily, however the misuse of statistics is a failure of logic.

Logic is an analytical tool as important as statistics, themselves. However; it is not taught in schools. The principles of logic are even less-understood than those of statistics. Consequently; we take statistics which are 'pure' numerical representations, and 'pollute' them with logical interpretations and extrapolations which are nothing short of laughable.

Mistakes employing statistics are inevitable. What is far more regrettable (and dangerous) is the deliberate misuse of statistics – 'polluting' a statistic in order to make it a tool of propaganda. The beauty of lying with statistics (in the eyes of the liar) is that such deceptions almost always provide plausible deniability – if confronted, the liar insists he is guilty of nothing worse than misinterpretation, rather than deliberate deception.

It is up to each of us, therefore, to be on our guard. First, we must learn to think for ourselves. When a 'number' is thrown at us, often with multiple conclusions attached; we must examine the raw data and carefully consider whether the 'number' actually validates the analysis based upon it.

An example of faulty analysis is in the U.S. housing market. There are many to choose from here, but let's pick the "construction of new homes" statistic, since it is an easy one. While this reading has been generally declining throughout the last year, there have been occasional upward ”blips.”

In one recent blip, a plethora of “experts” pointed to the increase in this statistic as an indication of stabilization in the collapsing U.S. housing sector. While an extremely superficial analysis might conclude this to be a reasonable view, it becomes immediately obvious how foolish such an extrapolation is when viewed in context. In a sector with declining sales and prices, the inventories of unsold homes were already at or near all-time record highs (depending on which statistic you use). What could be worse news than a new statistic that showed a significant addition to already-bloated inventories?

As the preceding example illustrates; correct statistical analysis often requires more than simply sound logic. General knowledge concerning the subject in question is also necessary, along with mathematical expertise in the many instances where statistics are the product of detailed and complex calculations (such as calculating the Gross Domestic Product of an economy).

When we lack the necessary general knowledge or statistical training, we are at the mercy of the person who produces the statistic – with no way of ascertaining whether we are being fed with information or deceived and manipulated with propaganda.

In my next discussion of the misuse of statistics, I will delve into what could be the greatest single deception of our modern society – accomplished almost entirely through the subtle distortion of one number.

{rokintensedebate} 

TrackBack URI for this entry

Comments (0)

Subscribe to this comment's feed