RT Journal
A1 Johnson AF
T1 VOice analysis and behavior pattern
JF JAMA
JO JAMA
YR 1969
FD July 28
VO 209
IS 4
SP 562
OP 562
DO 10.1001/jama.1969.03160170058021
UL http://dx.doi.org/10.1001/jama.1969.03160170058021
AB To the Editor:— 
    I should like to comment on the article "Voice Analysis for Detection of Behavior Pattern" by Friedman, et al (208:828, 1969).The usefulness of the proposed test in classifying subjects rests on the fact that the test procedure gave high results (greater than mean + 2 SD for the type B subjects) in 84% of type A subjects. In this instance 84% is 16 out of 19 subjects, a result compatible with a percentage anywhere from 60.4 to 96.6 (95% confidence limits) on the supposition that the patients were a random sample from a type A population. The result is, of course, more uncertain for the available sample of patients used because of possible bias.A further problem arises because the test examiner knew the behavior type of six of the type A subjects used.A fairer assessment for the test procedure would be ten out of