PRODUCTIVITY OF MUNICIPAL ENTERPRISES 385
ship differ so widely in the various enterprises, that a statistical examination must lose in thoroughness and range if it seeks to include several classes under a single investigation. From the point of view of method, therefore, it would seem better to make a separate statistical study for each industry affording the requis- ite information. While administrative considerations may have prevented, it would seem theoretically better had the bureaus of labor decided to make one report upon private and municipal water works and a second upon private and municipal lighting.
Since the methods to be employed must differ in detail with the various enterprises, it seems best to limit the further discus- sion to a single industry which may be deemed typical. For this purpose gas lighting has been chosen as the one upon which the greatest amount of trustworthy information is available.
If private and municipal gas plants are to be compared with reference to their economic productivity, certain units of measurement should be fixed at the start. Now it is of the essence of a scientific unit that for the purpose in hand one unit should be the approximate equivalent of any other. Yet in popular discussions this prime requisite of sound statistical work is often neglected. Thus a ton of coal is a unit frequently employed, but to the gas manufacturer it is grossly inaccurate, because the quality of the coal its gas-yielding power, and the character of the residuals is extremely variable and is almost as important to him as its quantity, but far less susceptible of measurement. So, too, a thousand cubic feet of illuminating gas is a popular unit, but the quality of the gas its constituents, candle power, and degree of purity is almost as variable as that of coal.
In discussions of the economic productivity of municipally owned gas works a comparison of the present condition of municipal and private plants should be held subordinate to a study of the changes that are in progress and the tendencies to change which are inherent in the various systems. The empha- sis has often but wrongly been laid upon a comparison of the present condition, partly because that is the more easily made,
