16 SAFETY AND PRODUCTION 
tance, the insurance interests, following the trend of progressive think- 
ing on insurance policy—namely to prevent losses instead of merely 
to settle claims—presented the problem for solution. 
While it was suggested that the primary purpose of the investiga- 
tion be the proof or disproof of the thesis on the relationship of safety 
and production efficiency, Mr. Whitney developed a number of other 
questions which began to assume importance as the study progressed. 
These are as follows: 
1. Have industrial accidents under the conditions of intense pro- 
duction of the past few years increased (a) in frequency, (b) in 
severity ? 
2. Are these increases in the frequency and severity of industrial 
accidents greater or less proportionately than the increase in produc- 
tion ? 
3. Can industrial accidents be controlled under modern conditions 
of production? 
4. If so, how? 
3. Does safety interfere with production? 
6. Does a positive correlation exist between the safety perform- 
ance and the efficiency of production? 
7. Is the safe factory the efficient factory. and the efficient factory 
the safe factory? 
8. Is safety an executive responsibility? Is the key to the new 
safety movement the executive? 
9. What degree of improvement can be made in safety performance? 
What reduction can be expected in the costs of industrial accidents? 
These questions naturally fall into four groups, of which the first 
includes questions 1 and 2; the second, questions 3, 4 and 5; the third, 
questions 6 and 7; and the fourth, questions 8 and 9. They will be 
taken up in this order and grouping in the presentation of the informa- 
tion obtained by the Committee. 
It was at once apparent to the Committee that American industry 
was so extensive and so ramified that, with the time and money at its 
disposal, a complete survey of the situation would be impossible. The 
most that could be done was to secure a somewhat extensive assay of 
facts and conditions, and draw conclusions therefrom, with full realiza- 
tion, however, that this process would be that of sampling and that 
there would be lacking in many instances positive knowledge that the 
sample was representative of the whole. Where figures could be secured 
for an entire industry or for the greater part of an industry, the con- 
clusions would be credible, but elsewhere the precise degree of repre- 
sentation would be in doubt. In any event, however, the data obtained