Well I have been in the QA group for a few weeks now. Mostly I have been running the standard test procedures on the nightly drops for F28 and running the canned regression tests on new versions of the Kernel. I've occasionally tested updates when I could recognize what a particular update was and had an idea how to test it. I've also done some tagging of packages.
For now I'm limiting my testing on fedora releases to Workstation Live because that is what I am familiar with as a user / maintainer (for a small group of users). I haven't had any use for VMs and I am quite familiar with PC hardware so I always do the testing on bare metal. I always use the delete all option on the disk drive.
I have dedicated a hardware PC for testing (Lenovo M58P with an E8400 processor). It is a fairly old dual core 3Ghz 4Gb machine with a 160Gb disk. It's main virtues are that it's in good condition and it was available. My experience with virtual machines pre-dates PCs and I haven't needed one in PC land so I feel more comfortable running on real hardware. Though I understand that Fedora uses virtual machines internally, that's "invisible" to me as a user, so I'm not counting that. The other thing I get from running direct on hardware is that when I see a potential bug there are fewer things to question and investigate. Experience has taught me to question almost everything, so if I used a VM I would be concerned about if the a problem was with the VM or the interface with the VM.
When I test a drop, after checking the ISO's checksum, I burn it to DVD and use the DVD to boot my test machine. I always choose the (I want to make space available) and pick the (Delete All) followed by (Reclam Space). I do this to avoid any possible issues with the last installation. After the installation is complete and the PC has been restarted, I set the display to a resolution that works with the monitor using the Settings application. Then I install Gnome-Tweaks and set the fonts to sizes that I am used to.
For the standard testing I generally run the ones that Coconut does not run; though there have been exceptions when I run more of them. After The standard tests, I use a bash script to load some additional applications. and check their basic functions in a manner similar to the standard tests.
If you'd like to comment or send some advice, please feel free to do so.