|
Field
Intervention (Hidden)
|
Before
First intervention is a voice over of the video recorded lecture in male and female voices.
Second intervention is information provision about the hypothetical instructor about their education, teaching experience, research interests and professional services.
This makes it a 2 (male, female) X 2 (with, without info) design
|
After
First intervention is a voice over of the video recorded lecture in male and female voices.
Second intervention is information provision about the hypothetical instructor about their education, teaching experience, research interests and professional services.
This makes it a 2 (male, female) X 2 (with, without info) design
An update (April 22, 2024): Based on reviewer comments received, we will be assessing whether students perceived the voice used in the experiment as a real voice or a synthetic/digital fake. We are thus conducting an additional non-incentivized online survey with students from a different institution, where the surveyees will listen to a short 19 seconds audio clip and will rate it on the following measures: friendly, reassuring, pleasing, trustworthy, can enhance knowledge, is engaging, and is trustworthy. We also ask them to guess the gender of the voice and rate the voice on a naturalness scale - with 1 being extremely automated and 5 being extremely human like.
About 400 students will be sent emails, receiving survey with one of the following five different audio clips, randomly assigned (audio reading a short text from the lecture, spanning 19 seconds) - (1) modulated female clip from experiment, (2) modulated male clip from experiment, (3) same female voice as in the experiment reading the text but in a synthetic manner, (4) same male voice as in the experiment reading the text in a synthetic manner, and (5) original clip (which was modulated in the main experiment to male and female versions). The response rate is expected to be 30-40% given the unincentivized nature of survey.
A comparison of how subjects rate the five clips will indicate how subjects perceived the audios in the lecture. We expect that the rating of (1) and (2) will be closer to (5) since (1) and (2) were just the modulated female and male versions of original naturally read (5). However, we expect (3) and (4) to get naturalness rating much lesser than (1), (2) and (5) since they are synthetically read texts presented in same voices as in the experiment. If we find this, it would suggest that our experimental audios (1) and (2) were perceived more natural than synthetically recorded or a perceived digital fake audio (3) and (4) would sound.
|