UNITED STATES DEPARTMENT OF AGRICULTURE 



1 BULLETIN No. 529 



Contribution from the Office of Farm Management. 

 W. J. SP1LLMAN, Chief. 



Washington, D. C. 



PROFESSIONAL PAPER 



April 5, 1917 



VALIDITY OF THE SURVEY METHOD OF 

 RESEARCH. 



By W. J. Spiixman, Chief, Office of Farm Management. 



CONTENTS. 



Introduction 1 



How farm records are obtained 2 , 



Accuracy of the farmer's knowledge 7 



Accuracy of cost-accounting methods 8 



Law of error 9 



Mistaken notions of accuracy 13 



INTRODUCTION. 



The distinguishing feature of farm-management investigations is 

 the application of the inductive method of reasoning to farm prac- 

 tice. In practically all farming communities can be found examples 

 of successful and of unsuccessful farms. It is assumed that a careful 

 analysis of the methods and business system of a large number of 

 farmers, all working under essentially similar soil, climatic, and 

 economic conditions, may be made to reveal the reasons for the suc- 

 cess of one and the failure of another. The essential difference 

 between the farm-management method and the laboratory method 

 of investigation lies in the fact that the laboratory investigator varies 

 his causes and studies the resulting variation in the effects produced. 

 The farm-management investigator has his experimental results 

 already produced for him. He merely collects the results of farm 

 experience, arranges them in such manner as to display the varia- 

 tions of a causal factor, and then studies the resulting variations in 

 the effects produced. Suppose, for example, it is desired, to know 

 what degree of soil fertility will result in the greatest profit to the 

 farmer under the conditions prevailing in a given locality. Having 

 analyzed the business of a large number of farms in the locality, the 

 farms are first grouped on the basis of their yields per acre, with 

 enough farms in each group to give reliable averages. The average 

 profit made by the farms in each group is then determined. Table I 



77589°— 17 



