ARTICULATION TESTING METHODS 817 



characteristics of the testing crew, the control of numerous haphazard 

 factors, and the control of the practice or experience that the crew- 

 acquires in the testing of circuits. 



The departure from normal, in acuity of hearing of prospective 

 crew members, should be measured with a good audiometer. In our 

 laboratories the 2-A audiometer is used for this purpose. Only those 

 individuals whose average hearing loss departs from normal in the 

 speech range of frequencies (100 to 8,000 cycles per second) by less 

 than 5 db (decibels) are selected. Although of normal hearing, some 

 observers of a crew usually obtain higher values of syllable articulation 

 than do others. The averages of the columns of Table III are a 

 typical set of results for nine observers who have passed such a hearing 

 test and who have had a year or more of experience in observing. 



Observer A. H. obtained the highest percentage, namely 59, and 

 observer M. W. obtained the lowest percentage, namely 38. In 

 general this order would be preserved in a series of tests, although 

 haphazard variations in a single test might change the order. The 

 spread in observations is of an order of 20 per cent. More extended 

 tests have shown that the spread tends to decrease as the observed 

 percentages approach or 100. In order to make a replacement in 

 the observing personnel from time to time without causing a probable 

 change of more than 2 per cent in the average percentage, it is necessary 

 to use an observing crew of 8 to 10 persons. Our experience has 

 shown that men and women show no characteristic difference when 

 acting as observers. 



The ability of prospective crew members to enunciate the sounds 

 in a normal way is determined in the following manner. An extensive 

 series of tests on various voices have yielded data which are arbitrarily 

 used as a basis for determining normalcy. These tests were made 

 with a simplified list consisting of common English words which will 

 be described in a later paragraph (see Table XVH). Tests were 

 made under three conditions; namely, direct transmission through 

 the air in a quiet, well damped room, transmission over a circuit 

 which uniformly transmitted the frequency range from 100-4,500 

 cycles, transmission over a circuit having a carbon transmitter. A 

 diagram of this latter circuit is shown in Fig. 7. The sounds were 

 observed by a crew of experienced observers. Table V gives the 

 results of tests that were made upon 21 male and 23 female voices, 

 the personnel being selected from various departments of the Labora- 

 tories. The average articulations of the simple consonant sounds are 

 shown. The data are given separately for men and women. 



