If you want to test that respondent is able to hear, you can add one screener question with some sound and ask question like which sound is this (say guitar, piano...).

---

Thanks @bansalpeeyush29 But the issue is not whether respondents are able to hear, which can be easily sorted out with a screener.

The key is that I need to know how volume level affects their responses. So their device volume information during the audio play would be crucial for my study. I could ask them to set their device volume to let's say 50% during the study. But it would be better if I could double-check whether they cooperated or not.

---

Hello @chunkyu_lee



I guess that won't be possible as any browser made by Microsoft/Google/Apple/Mozilla/Opera doesn't allow this to avoid exploits through File.IO and limit developer's access to user's machines.

---

This is something you'd have to do in a controlled environment in order to trust your data is valid. For example, do you have any sort of physical location that you can control, like computer lab at universities or a conference room in a company?



Depending on the population requirements you have, you could even try doing this in some sort of public (or willing private domain), like park (or mall/coffee shop).



The only way you would be able to trust your data is to verify the volume levels before and after each response session. This is not an element I would trust to self-reported data.

---

@Akdashboard Thank you for the suggestion. Yes I will be running a lab experiment but I'd have to wait for mid-semester to recruit undergrad participants. I was trying to do a pretest online but you're right the validity would be really low. I guess I'll just give up the pretest.

---