November 16, 1917] 



SCIENCE 



473 



best but fragmentary and fails to yield all 

 the knowledge which we need for guidance 

 in the complex social conditions of the 

 present. 



This perception has led to the search for, 

 and the emphasis upon, other methods of 

 social research and investigation. Chief 

 among these has been statistics. Statistics 

 has had many enthusiastic advocates as 

 the method of the social sciences, both 

 among economists and sociologists, a recent 

 advocate going so far as to say that the 

 statistical method bears much the same rela- 

 tion to the social sciences that the experi- 

 mental method bears to the physical sci- 

 ences." There can be no doubt that 

 statistics presents the one means of measur- 

 ing social facts upon a wide scale, and so 

 of rendering our knowledge of mass move- 

 ments exact. In so far as exact measure- 

 ments are needed in the social sciences 

 (and they are needed not less than in other 

 sciences), the statistical method must re- 

 main a highly important part of the 

 methodology of the social sciences. It is 

 greatly to be regretted, therefore, that as 

 yet we possess adequate statistics of only 

 very small sections of our social life; and 

 it is manifestly our duty as students 

 banded together to promote scholarship in 

 the social sciences to do all that we can to 

 promote the accurate collection and study 

 of social statistics. However, apart from 

 the fact that statistical methods have still 

 to be enormously developed before they are 

 susceptible of application to the general 

 problems in the field of the social sciences, 

 it is evident that there are many problems 

 in political science, jurisprudence, sociol- 

 ogy and other social sciences which by 

 their nature are not amenable to statistical 



3 See the suggestive articles on ' ' The Experi- 

 mental Method and Sociology" by Professor F. 

 Stuart Chapin in the February and March, 1917, 

 issues of The Scientific Monthly. 



treatment. It is noteworthy, moreover, 

 that the natural sciences have made but a 

 subordinate use of statistics. It is true 

 that they have other instruments of pre- 

 cision, but the experimental method, so far 

 from closely resembling the statistical 

 method, is rather mere observation under 

 controlled conditions. It would seem, 

 therefore, that the nearest approach to it 

 in the social sciences would be the direct 

 observation of social life under mentally 

 controlled conditions. It is true that 

 social conditions can rarely be fully con- 

 trolled, but observation hy trained observ- 

 ers can be, and the results can be checked 

 up with the aid of the historical, compara- 

 tive, and statistical methods. 



A little over a dozen years ago the prac- 

 tical needs of social workers for more ac- 

 curate and scientific knowledge of the 

 social conditions in the communities in 

 which they worked led to their instituting 

 programs of social investigation which they 

 called "social or community surveys." 

 One of the first and most extensive of these 

 "surveys" was the well-known "Pittsburgh 

 Survey." A great number of these sur- 

 veys have now been made in widely scat- 

 tered communities, and the movement has 

 become specialized, so that now we have 

 surveys of different sorts, such as "health 

 surveys," "educational surveys," "in- 

 dustrial surveys," "agricultural surveys," 

 etc. It will be noted that the movement 

 arose entirely to meet practical needs, and 

 that there was no thought of making a con- 

 tribution to scientific methods of studying 

 the social life. At first, the movement was 

 narrow. The "survey" was confined 

 largely to the material aspects of the social 

 life, such as sanitation, housing, wages, etc. 

 Moreover, the survey was supposed to be 

 an entirely local and community affair, and 

 though statistical accuracy was emphasized, 

 but little attention was paid to history and 



