Falsified figures do not make statistics!

| by Dr U. Pethiyagoda

( April 27, 2012, Colombo, Sri Lanka Guardian) Not a day passes without our being assailed with statistics. Be it GDP, per capita income, unemployment, income distribution, road accidents, child molestation, economic growth, crime , tourist arrivals, energy, garbage disposal - you name it, apparently we have it. There are some entertaining features. Everybody seems competent to use whatever numbers that they can get hold of, to support their cause. Rather as they say, a drunkard uses a lamp post - more for support than for illumination! Public officers to secure their tenures. Politicians to sharpen their lying ability. Almost none to draw useful conclusions. Another feature is that none disclose the methodology employed to reach the numbers that they quote. The sole purpose of a figure is to endorse, or even better, dismiss a popularly held notion. Never to pretend erudition or unjustified exactitude. The economists are a fascinating example. Faced with a discipline barely describable as an art and most emphatically not a science, they believe that numbers based on dubious techniques and believed to be supportive of some theory that is either totally obscure or painfully obvious, somehow lends respectability. They do not display evidence of appreciating the function of a decimal and its implication of accuracy of the raw figure! They are rightfully to be pitied! Economics is possibly not all tosh - after all a Nobel Prize is awarded annually to select practitioners. But after all, Barak Obama was the recipient of the Nobel Peace Prize even before he had taken his first breath as US President! The Economist is doubtless a respected regular journal devoted to the subject - but significantly, also a lot besides.

Purveyors of statistical jargon may profitably learn from the format of a good scientific paper. That is "Introduction" (answering the question " why?"), "Material and methods", "Results" (numbers arrived at), "Discussion" (a reasoned conclusion/s) "Summary", Acknowledgments" and "References". Otherwise statistics and economics become the substance of "beer table" talk and no more. Personally, I dismiss much of the statistics I encounter because of a steadfast belief that no amount of manipulation or obfuscation can yield good conclusions from bad data.I have for example, yet to succeed in getting anybody to explain how the seminal figure of GDP is compiled and from which so many other indices are derived. Almost without exception they evade the question of accuracy and veracity by mumbling moth-eaten definitions of "goods and services" and like unascertainable garbage! No help at all!

Another favourite gimmick is to quote some external agency - World Bank, ADB or UN bodies as authorities to endorse some "statistic", being careful to not disclose the fact that their point is derived from the very defective numbers given to them by Government itself!. One of the most potent tools of poor governance is the photocopier!

So next time you see numbers quoted with pretended precision, ponder on the possible original sources. Then reach for the WPB! An entertaining (and obscure) dialogue is currently proceeding in the media, between Dr Harsha de Silva and Mr R. M. B. Senanayake - two practitioners of the art, whose words are generally oracle. They are fencing on some issue relating to divergences of GDP figures. Characteristically, they sidestep the issue of the reliability of the raw data on which the figure is calculated, while engaging in their polemics! The uninitiated may be forgiven for believing that this factor is germane.