STATISTICS AND ADMINISTRATION. 447

employment of substitutes is not entered in the books, and other factors
boo often introduce errors of importance. Further, the employer is
only able to give particulars of the individual when he is actually
earning money. There is at present little record of periods of absence,
and the record of a man’s earnings for a month or two may afford
an entirely misleading indication of his average annual income. Finally,
even if accurate information regarding the earnings of the individual
could be secured from the employer’s books, it would in itself have
a very limited value. We believe that in some cases employers
might find such information useful as a measure of the success of any
endeavour to raise the standard of living by increasing regularity of
employment, but it would be of practically no value as a measure of the
standard of living itself. This depends on a large number of factors
lying outside the knowledge of the employer. It depends in the first
instance on the income, not of the individual, but of the family, and even
that income affords little indication of the measure of comfort. In order
to secure this, it is necessary to have full information regarding the com-
position of the family and the various claims on its income.
Income and Expenditure.

Information regarding the income of the workers must ordinarily
be associated with information regarding its expenditure, and both are
best, collected by means of family budget enquiries. Here again we find
that comparatively little progress has been made. Enquiries of value
have been conducted by the Bombay Labour Office in Bombay, Ahmed-
abad and Sholapur and by the Labour Statistics Bureau in Rangoon.
In some other centres a certain amount of work has been done. As a
rule this has been directed more towards measuring movements in the
cost of living than towards giving any full indication of the standard
itself. One or two of these smaller enquiries have been successfully
carried through. In a number of cases enquiries have been instituted by
students of economics, social workers, officials and other agencies, without
any clear recognition of the difficulties involved and without much regard
bo statistical principles. The two commonest errors were the failure to
give adequate training to the investigators of the basic facts, and in-
difference to the vital factor of sampling. As a result of these and other
mistakes, a fair amount of honest effort has been misdirected.

Training of Investigators.

Anxious as we are to see a great extension of economic enquiries
bearing on the standard of living, we must emphasise the difficulties in
the way. The collection of statistical material from the workers on any
extensive scale requires special qualifications. For an untrained investi-
gator to descend on the workers’ homes and collect such particulars
as he can in a casual visit 18 valueless. The preliminary difficulties
have been faced already by the Bombay Labour Office, and they have
evolved a technique which can be studied with advantage by others
Who propose to embark on similar enquiries. We recommend that, where-
ever possible, investigators should undertake a course of training with