Edgeworth's Test of Interpersonally Dependent Demand

It turns out that, in the late 19th and early 20th centuries, economists were very much concerned with whether or not their models should incorporate the ways in which one agent's consumption can affect another's utility. In creating a literature review for Hongyi's and my work on conspicuous consumption, I stumbled across one of the earliest empirical tests of this idea, conducted by Edgeworth himself:

The conception may be illustrated by the common supposition that at social gatherings which are cheered by alcoholic beverages the consumption of liquor per head is likely to be greater the more numerous the company. An opportunity of testing this belief is afforded by the varying size of the dinners at a certain Oxford college, whose members are thought to be susceptible to the influences of good fellowship.

Finding that the dinner's population had no effect on wine drunk per person, Edgeworth concluded that indirect, interpersonal demand effects had no merit, and he never revisited the concept.