Professor Ian Angell

Department of Management, LSE, London WC2A 2AE    |    +44 (0)20 7955 7655    |    i.angell@lse.ac.uk

"A single death is a tragedy, a million deaths is a statistic"

These words of monster Joseph Stalin show his clear understanding of the monstrosity that is statistics. He knew that statistics deliver the clear conscience of a numerical justification that avoids responsibility for "man's inhumanity to man". His secret is out. Even our second-rate control freaks at Westminster realise they can camouflage bureaucratic abuse by deluging a gullible public with the 'smoke and mirrors' of statistical targets.

They tell us the 'big lie' that "the figures don't lie", as they treat us as categories and not as individuals. However, category is not truth, but merely an act of choice driven by hidden agendas and prejudged priorities. Categories are intrinsically ambiguous. They can be distorted so that most of the data is 'on message', while awkward numbers are ignored. "Domination is transfigured into administration."

Reduced to statistics, both human problems and problem humans are made anonymous. Brutally insensitive decisions and actions are reduced to bookkeeping. As a paper exercise, otherwise unpleasant acts are purified. Treated as mere statistics, the unemployed, homeless, hospital waiting lists, alcoholics and other drug addicts, the mentally ill are all made invisible.

From their numerical platform, obsessive compulsive neurotic politicians claim to manage the risk and the uncertainty of our society with the arbitrary use of statistical models; with opinion polls, market research, focus groups, socio-economic classifications, performance measures, efficiency audits, cost-benefit analyses. Through statistics, society becomes well-behaved, tidy, controllable.

Instead of continually re-evaluating every uncertain situation and depending on the talent of individuals, political decision-making becomes a matter of controlling the future by labelling it with numbers. However the approach of searching for the right numerical label to represent the future, is no different to numerology, astrology, or the personality tests found in women's magazines.

Statistical targets are delusions. For "any observed statistical regularity will tend to collapse once pressure is placed on it for control purposes". Goodhart's Law says  you cannot mix up cause and effect. Any observed regularity in society is an effect, but the moment it is measured and then used as the basis of control, the false assumption is made that the regularity is itself the cause of the observation.

At a recent conference on 'Managing Uncertainty' I was expected to drone on about distributions, expectations, formulae. However, I just stood there, .. and stood there. For a whole minute I just stood there. I shot frantic and terrified glances at the audience. At first there was silence, then a few murmurs, then a growing rumble of concern. When the chairman rose to help, I banged the table: "now that is uncertainty, it has nothing to do with statistics".

Life is a game of poker; and there's a lot of bluffing going on. Players who put all their faith in the statistical distribution of the cards are going to lose their shirts. Unfortunately, our so-called leaders will lose our shirts as well.

When politicians talk of a 'project' that will perfect society, when they pronounce that "the figures don't lie", I hear Mark Twain: "it's not the figures lying, it's the liars figuring." When the liars are hiding behind their wall of statistics, it's time to man our barricades.

Ian Angell is professor of Information Systems at the London School of Economics and Political Science.