Yes/no and two-interval forced-choice tasks with listener-based vs observer-based responses

Lori J. Leibold, Emily Buss

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


Observer-based procedures are used to assess auditory behavior in infants, often incorporating adaptive tracking algorithms. These procedures are reliable, but effects of modifications made to accommodate infant testing are not fully understood. One modification is that observation intervals are undefined for the listener, introducing signal-temporal uncertainty and increasing the likelihood that listener response bias will influence estimates of performance. The effect of these factors was evaluated by comparing threshold estimates obtained from adults using two tasks: (1) single-interval, yes/no and (2) two-interval, forced-choice. Detection thresholds were estimated adaptively for a 1000-Hz FM tone in quiet and for a word presented in two-talker speech masking. Trials were initiated and judged by the observer (observer-based) or the listener (listener-based). Thus, listening intervals were temporally uncertain in observer-based procedures and temporally defined in listener-based procedures. Thresholds were higher for observer-based relative to corresponding listener-based procedures. The magnitude of this difference was similar across the yes/no and two-interval tasks, and was larger for masked word detection than tone detection in quiet. Listeners adopted a conservative criterion when tested using the observer-based, yes/no procedure, but modeling results suggest that signal-temporal uncertainty accounts for the largest portion of the threshold difference between observer-based and listener-based procedures.

Original languageEnglish (US)
Pages (from-to)1588-1596
Number of pages9
JournalJournal of the Acoustical Society of America
Issue number3
StatePublished - Mar 1 2020

ASJC Scopus subject areas

  • Arts and Humanities (miscellaneous)
  • Acoustics and Ultrasonics


Dive into the research topics of 'Yes/no and two-interval forced-choice tasks with listener-based vs observer-based responses'. Together they form a unique fingerprint.

Cite this