THE ART OF SCIENTIFIC INVESTIGATION 



Many of the anecdotes cited in Chapters Three and Four and 

 in the Appendix also provide illustrations of the role of observa- 

 tion in research. 



Some general principles in observation 



In discussing the thoroughly unreliable nature of eye-witness 

 observation of everyday events, W. H. George says : 



" What is observed depends on who is looking. To get some 

 agreement between observers they must be paying attention, 

 their lives must not be consciously in danger, their prime neces- 

 sities of life must preferably be satisfied and they must not be 

 taken by surprise. If they are observing a transient phenomenon, 

 it must be repeated many times and preferably they must not 

 only look at, but must look for, each detail."*^ 



As an illustration of the difficulty of making careful observa- 

 tions, he tells the following story. 



At a congress on psychology at Gottingen, during one of the 

 meetings, a man suddenly rushed into the room chased by another 

 with a revolver. After a scuffle in the middle of the room a shot 

 was fired and both men rushed out again about twenty seconds 

 after having entered. Immediately the chairman asked those 

 present to write down an account of what they had seen. 

 Although the observers did not know it at the time, the incident 

 had been previously arranged, rehearsed and photographed. Of 

 the forty reports presented, only one had less than 20 per cent 

 mistakes about the principal facts, 14 had from 20 to 40 per cent 

 mistakes, and 25 had more than 40 per cent mistakes. The most 

 noteworthy feature was that in over half the accounts, 10 per 

 cent or more of the details were pure inventions. This poor 

 record was obtained in spite of favourable circumstances, for 

 the whole incident was short and sufficiently striking to arrest 

 attention, the details were immediately written down by people 

 accustomed to scientific observation and no one was himself 

 involved. Experiments of this nature are commonly conducted 

 by psychologists and nearly always produce results of a similar 



type. 



Perhaps the first thing to realise about observations is that not 

 only do observers frequently miss seemingly obvious things, but 

 what is even more important, thev often invent quite false 



98 



