Essay --

This research project investigates whether ensemble statistics are also possible in the sound intensity domain. It has been established that people represent multiple objects as ensembles in many types of stimuli in visual domain. Also, some studies investigated this ability in auditory domain on frequency. The present study aims to show that people can also represent a set of varying decibels as an ensemble. This ability might be crucial given that in daily life people are exposed to multiple levels of sound intensities and extracting summary statistics can be beneficial to guide attention to more salient sounds, like rising sound which can indicate an approaching target, among many others. To investigate whether people represent different decibels as ensembles this study used dynamic exposure in temporal sequences and to get finer grained responses used a response method in which participants generated their own instead of choosing among alternatives. Since the relationship between decibel increase and perceived loudness increase is not linear but it follows Stevens’ power rule, it is expected that participants’ responses will provide a better fit with Stevens’ exponent for loudness in perceiving the average loudness of sounds. â€Æ' In daily life, along with individual stimuli, people encounter with crowds of stimuli and they are able to perform in the face of the abundance of stimuli. Given that sensory systems’ capacity is limited to a certain amount of information, (Alvarez & Cavanagh, 2004), instead of representing every single object individually, there may be another way for representing group of items (Ariely, 2001). Ensemble representations, which can be defined as representations that give the compressed summary statistics... ...rentiating softer sounds would not be likely, a type of dynamic display should be used. For this study, discrete presentation of individual sounds in following temporal sequences will be used. Also, as for the response selection, people will generate their own responses. This, as also suggested by Haberman and Whitney (2009), allows getting a finer grained response from the participants. Hence, this study proposes that as it is the case for many other properties of various stimuli types, humans should be able to extract summary statistics of a set of sound decibels, which is a fundamental property of our acoustic environment, but this averaging should follow Stevens power function on loudness. This research project first explores whether this is the case, whether people can find the average decibel of sounds that are presented dynamically in temporal sequences.