Recognition of Audified Data in Untrained Listeners
Alexander, Robert Lewis, II
MetadataShow full item record
The effective navigation and analysis of large data sets is a persistent challenge within the scientific community. The objective of this experiment was to determine whether participants who received no training were able to identify audified data sets at a rate above chance in a forced-choice listening task. Nineteen participants with various levels of musical and scientific expertise were asked to place audio examples into one of the five following categories: Digitally Generated Sound - White Noise, Solar Wind Data, Neuron Firing Data from a Human Brain, Seismic Data (Earthquake Activity), and Digitally Generated Sound - Sinusoidal Waveform. At no time were participants made aware of the accuracy of their responses during the experiment. Participants who had never been exposed to audified data sets were able to recognize audification examples at a rate that was 23 percentage points above chance performance; however, the sample size of individuals with no previous exposure to audified data was not large enough to determine statistical significance. When controlling for previous exposure to any of the provided listening examples, all participants performed well above the statistical likelihood of chance responses in the recognition of digitally generated white noise and sinusoidal waveforms. This indicates that participants with no previous exposure to audified data were able to discriminate between audified data and digitally generated sounds.