The Secret Master The DZNeP-Scene Is Actually Uncomplicated!

Матеріал з HistoryPedia
Версія від 08:56, 2 травня 2017, створена Drawer9parade (обговореннявнесок) (Створена сторінка: Indeed two of the UBC-17 observers who were excluded from this analysis had non-monotonic psychophysical functions for the 65 Hz stimuli, indicating that they w...)

(різн.) ← Попередня версія • Поточна версія (різн.) • Новіша версія → (різн.)
Перейти до: навігація, пошук

Indeed two of the UBC-17 observers who were excluded from this analysis had non-monotonic psychophysical functions for the 65 Hz stimuli, indicating that they were simply guessing the appropriate response. These observers may have had an undiagnosed low-frequency hearing loss. The fact that the recalibration judgments of 1000 Hz stimuli interleaved with the 65 Hz judgments for these observers remained normal, as did their other psychophysical functions, indicates that the problem was only with the 65 Hz stimuli. As mentioned earlier, West et al. (2000)��s observers performed 200 judgments or 50 judgments of the 65 Hz stimuli; perhaps more practice with these stimuli in our experiments would have led to a closer replication. Nevertheless, this failure of convergence of the 65 Hz exponents between most of the present 17-stimulus experiments and the 200- and 50-stimulus experiments of West et al. (2000) needs to be investigated further, and indicates that the minimal implementation of CS might not be good enough for scientific purposes in some cases. Moreover it emphasizes the problem that arose in comparing exponents across different protocols: given the regression, practice, and fatigue effects that are ubiquitous in psychological experiments, and the extraordinary cross experiment and cross-individual precision of CS within a given protocol, perhaps not only response scales but even protocols must be standardized in order to achieve the desired level of reproducibility of experiments in psychophysics. The numbers of valid subjects listed in Table ?Table22 are relevant to the question of standardization of the protocol. There is a clear trade-off of precision against efficiency in CS, as in any other experimental method. For the 52-stimulus protocol only four of a total of 54 (7.4%) 500 and 5000-Hz runs had rRP2 hand, for the 17-stimulus protocol, 39 of the total of 205 (19.0%) 65, 500, and 5000-Hz runs, and 12.7% of the corresponding recalibration runs, failed to meet our fairly liberal criterion. For scientific purposes even the 52-stimulus protocol is relatively quick, gives adequate precision, and probably represents a good compromise between efficiency and precision. The numbers just listed, DZNeP however, make it clear that, although ideal for the clinic, the 17-stimulus protocol will result in a fair number of imprecise measurements, and that measures should be taken to improve results when this occurs. Thus, in our opinion the best approach for the clinic would be to build the psychophysical function in stages and measure rRP2 for each stage (possibly with decreasing numbers of trials) until the desired precision is obtained, similarly to the way the psychometric function is built up using adaptive techniques (e.g.