"It does exactly what it says on the tin"

[Massimo Grassi <massimo.grassi@xxxxxxxx>]

Dear list members,

when we use an adaptive procedure such as the classic staircase we track 
a specific point of the psychometric function. For example, if we use 
the 2-down 1-up we track a threshold that corresponds to 70.7%.

I'm looking for a paper that investigates whether the adaptive procedure 
(e.g., the classic staircase) "does exactly what it says on the tin". In 
other words, whether the threshold estimate returned by the procedure 
elicits *that* percent of positive responses.

I remember that a few years ago I read a paper that was investigating 
exactly this! To the best of my memory, it was an '80s paper that was 
published by P&P (or JEP:HPP?). However, in these days I searched that 
paper in any possible direction with no success.

Maybe out there there is someone with a better memory.

Thank you all in advance for the help,
m
--
http://www.psy.unipd.it/~grassi/
http://www.springer.com/978-1-4614-2196-2