Subject: Re: Threshold of increasing/decreasing tones From: Wolfgang Ellermeier <we@xxxxxxxx> Date: Thu, 16 Feb 2006 10:19:58 +0100Dear Stefan Kerber, while John Neuhoff's response points to true perceptual effects in SUPRATHRESHOLD rising and falling levels, you might also want to consider simple RESPONSE BIASES. Already in 19th century psychophysics, you find discussions of "habituation effects" and "anticipation effects" biasing threshold measurements like the one you are using. There is a risk that you will never know, in any given subject, what mix of these effects will occur, in some circumstances favouring your 'ascending', in others your 'descending' runs (as you seem to imply is the case with your data). The standard methodological way out of this dilemma has been to use 'unbiased', or 'criterion-free' forced-choice (adaptive) procedures, as described in Levitt (1971). Of course, they require short tone bursts, rather than the continuous tones used in your paradigm. A discussion of the pros and cons of the 'classical' threshold procedures (like the method of limits that you are using) can be found in the methods chapter of our (German, sorry list!) textbook on hearing [Hellbr\"uck, J. & Ellermeier, W. (2004). // <http://www.hogrefe.de/buch/isbn/3-8017-1475-6.html>H\"oren. Physiologie, Psychologie - Pathologie. G\"ottingen: Hogrefe.] . Hope, these hints are helpful, Regards, Wolfgang Ellermeier -- ___________________________________________________________ Wolfgang Ellermeier e-mail: we@xxxxxxxx http://www.soundquality.dk http://www.acoustics.aau.dk phone: +45 9635 8713 FAX: +45 9815 2144 Forskningsprofessor Research Professor Afdeling for Akustik Departm. of Acoustics Aalborg Universitet Aalborg University Fredrik Bajersvej 7 B5 Fredrik Bajersvej 7 B5 DK-9220 Aalborg Ø DK-9220 Aalborg O Danmark Denmark _____________________________________________________________