I'm doing data analysis from an experiment in which participants compared two pieces of music. The pieces of music varied in whether they were composed with standard Western diatonic scales with registrally ascending pitches or with "scales" consisting of nondiatonic pitch-class sets in an altered order. The pieces of music were presented in pairs, and listeners had to rate how similar the two pieces sounded.
The experiment was conducted via online survey software that randomized the pieces of music that were presented. Participants used their own computers, listening and providing ratings at the time and location of their own choosing.
The pieces of music were approximately 24 seconds long, and the duration that participants spent on each music presentation page was recorded by the survey software.
We need to ensure that participants listened to each piece of music an acceptable amount of time.
Here's the question: What is an acceptable minimum duration and an acceptable maximum duration?
Might some listeners listen for a second and then click to the next page so that they can finish quickly, not really hearing the music and providing uneducated answers? Might some listeners have such good discrimination that they can tell within seconds whether the pieces use different music systems?
What is an acceptable maximum duration? If a person listens to a piece and then listens to the second piece within 10 seconds, are they making a more valid judgment than someone who listens to the first piece then gets a snack and surfs the web for five minutes before listening to the second piece? There should be some maximum cutoff duration, that when exceeded, disqualifies the data from that judgment.
Regarding the minimum, I'm going to test whether judgments after 5 seconds of listening are more accurate than those that occur after only 1 second of listening. Perhaps that will help me make the decision.
Regarding the maximum, I am choosing a semi-arbitrary cutoff of 56 seconds, which is twice the duration of the longest stimulus. That allows some time for delays in downloading or other problems, while preventing participants from listening to a presentation more than twice or waiting too long before listening to the second piece or making a judgment.
Do these sound reasonable? Is there research literature to support these methods or to suggest something better?
The experiment was conducted via online survey software that randomized the pieces of music that were presented. Participants used their own computers, listening and providing ratings at the time and location of their own choosing.
The pieces of music were approximately 24 seconds long, and the duration that participants spent on each music presentation page was recorded by the survey software.
We need to ensure that participants listened to each piece of music an acceptable amount of time.
Here's the question: What is an acceptable minimum duration and an acceptable maximum duration?
Might some listeners listen for a second and then click to the next page so that they can finish quickly, not really hearing the music and providing uneducated answers? Might some listeners have such good discrimination that they can tell within seconds whether the pieces use different music systems?
What is an acceptable maximum duration? If a person listens to a piece and then listens to the second piece within 10 seconds, are they making a more valid judgment than someone who listens to the first piece then gets a snack and surfs the web for five minutes before listening to the second piece? There should be some maximum cutoff duration, that when exceeded, disqualifies the data from that judgment.
Regarding the minimum, I'm going to test whether judgments after 5 seconds of listening are more accurate than those that occur after only 1 second of listening. Perhaps that will help me make the decision.
Regarding the maximum, I am choosing a semi-arbitrary cutoff of 56 seconds, which is twice the duration of the longest stimulus. That allows some time for delays in downloading or other problems, while preventing participants from listening to a presentation more than twice or waiting too long before listening to the second piece or making a judgment.
Do these sound reasonable? Is there research literature to support these methods or to suggest something better?
