CH. 6

CURRENCY STANDARDS

83

of price they produce in one or more commodities, then allthe individual variations of prices will be correctly balancedoff against each other in the geometric mean, and the truevariation of the value of gold will be detected.

Bowley, Elements of Statistics (5th ed.), p. 198 :

Thus if we start out to measure prices in general ... inorder that the resulting index-number should be subject tothe analysis of the law of error, the samples should be randomand independent in their fluctuations from the general move-ment ; dependence increases the number of samples necessaryfor an assigned precision. ... If the number of independentquantities is at all considerable, any reasonable system ofweights is likely to give as good a result as the conditions ofthe problem allow.

Edgeworth, Papers relating to Political Economy,vol. i. p. 247 :

We have seen that, upon the supposition of a changein the supply of money, Jevons’ method of combining thevariations of prices without regard to the correspondingvolumes of transactions is by no means so absurd as hasbeen thought by some. The case is, as if we wanted todiscover the change in the length of shadows, due to theadvance of day. If the objects casting shadows were un-steady-waving trees, for instance—a single measurementmight be insufficient. We might have to take the mean ofseveral shadows. Now for our purpose the breadth of theupright object casting the shadow would be unimportant.The “ wide-spreading beech ” and the mast-like pine wouldserve equally well as a rude chronometer.

Again, p. 256:

There seem to be combined in popular thought two elementswhich we have sought to distinguish in analysis, namely, theconception of an objective mean variation of general prices,and the change in the power of money to purchase advantages.

We have, in short, to this way of thinking, atypical problem in the combination of observations,where each individual observation is subject to adisturbing factor which it is our business to eliminate.