TIME INTERVALS IN TELEPHONIC CONVERSATION 283 



item which is sometimes imposed by the configuration of toll circuits, 

 namely lockout} Lockouts did occur on the circuit carrying the con- 

 versations which provided the data for this study, due to the special 

 circuit arrangement employed at the time, but their occurrence will 

 not be treated in this paper. 



The specific information desired was the probability that a con- 

 versational element would have any given duration /. The true 

 probability in each case can be approximated, except for a scale factor, 

 by a distribution curve of experimental data depicting the fractions of 

 the total number of observations of the quantity which lie within each 

 of a regular progression of time cells of a chosen width. It is in this 

 form that the data will be presented. 



Data Source 



Some preliminary data were obtained from observations on local 

 inter-ofifice calls between various members of the stalT of Bell Telephone 

 Laboratories. Members of the non-technical groups were included 

 among the talkers, some of whom were women. Equipment added 

 to the telephone circuit for the purpose of recording was held at a 

 minimum and its presence and action were not noticed by any of the 

 talkers. The recording means were less elaborate than those employed 

 in the later investigation, and the principal interest in the preliminary 

 test lies in the fact that the results have shown themselves in good 

 agreement with those obtained later with different talkers and widely 

 different circuit conditions. 



The conversations which provided the material for the main part 

 of this study were those which took place between male talkers on a 

 circuit used as a tie line by the Western Electric Company and running 

 from the company's Hawthorne plant at Chicago, Illinois, to the New 

 York office. This is a circuit which is used wholly for transaction of 

 company business, and most of the users hold at least minor executive 

 positions in their organizations and have a background of scientific or 

 formal business training. It is recognized by the authors that this 

 somewhat restricted class of talkers may not be representative of tele- 

 phone users in general. It is also recognized that the manner of 

 telephonic conversation may be different for long distance and local 

 calls. However, analyses of data from other tests which have been 

 made on the tie line, involving a wide range of delays and types of 

 voice operated devices, have indicated that the talkers endeavored to 

 converse in the same manner regardless of the circuit configuration. 



^ A lockout is the simultaneous blocking, by voice-operated devices, of both direc- 

 tions of transmission of a two-way communication system. For a discussion of this, 

 and other possible definitions, see the companion paper by Mr. A. W. Horton, Jr. 



