When you manually test on screen readers, do you test all the different versions, or would one work the same?

We manually audit using different screen readers. Across our team, we use Voice Over, JAWS, and NVDA to test on Chrome, Safari, and Firefox. The majority of the time, when we're testing on multiple devices/browsers, the content is coded correctly, but we're trying to ensure it works across different combinations. For example, if we encounter sound discrepancies in one testing environment, another team member with a different browser and assistive technology will double-check that we are getting repeatable results.

JAWS and NVDA are comparable in how they function. Voice Over is a little bit different. For example, Voice Over can detect if something is coded inaccessibly and try to fix it. Typically, testing with a single type of screen reader is not suggested. While auditing, you want to ensure the code is accessible and communicates the same information through different assistive technologies, browsers, and operating systems.