Each consumer received six samples corresponding to six random
storage times. Panelists were asked to sniff the samples and report
whether they were acceptable or unacceptable. The CensorReg procedures
from S-PLUS (Insightful Corp., Seattle, WA; ver. 7) were used
according to the protocol of Hough et al. (16, 18). Censoring was
defined as follows: at a given storage time t two possible answers
could be given by the panelists. (a) The sample was perceived as
acceptable, indicating that it would be rejected beyond time t; thus,
the data are right censored. (b) The sample was perceived as
unacceptable, indicating that the panelist would start rejecting the
product before time t; thus, the data are left censored