ABSTRACT
Anthony Smith, former director of the British Film Institute and a leading proponent of public service broadcasting, gave vent to the ambivalence typical in European views on audience measurement:
The measurements have the same kind of precision as pre-election polling. Better methods are constantly being devised and tried out and complex though interesting technical arguments take place concerning the comparative reliability of methods. Happily, we have never chosen to use these systems to replace elections in the political sphere. It is one thing to sample an electorate and enquire about preferences in politics in a given week. It is quite another to sample television receivers, which may be switched on without anyone being in the room, or with people in the room who are not actually watching, who may be asleep or playing cards or listening to music on headphones. It is fortunate that in Britain the measuring of audiences has been carried out not to decide the levels of revenue for any system, but merely to satisfy curiosity and to feed institutional rivalries. In a truly competitive commercial system the measurement of audiences through crude methods of sampling is paramount; programmes, channels, whole companies even, can be swept away through the presence of unknown statistical bugs. In America vast quantities of investment in pilot programmes are regularly junked on the basis of ratings divergences well within the margins of statistical error. It is the extension of roulette into culture.
