Use repeated measurements and context rather than treating one number as an absolute diagnosis.
WI-FI VS ETHERNET
Run two user-confirmed measurements on the same device: first Wi-Fi, then Ethernet. The browser cannot verify the physical transport, so the switch is explicitly yours to confirm.
0bps cannot reliably detect whether the browser is on Wi-Fi or Ethernet. Run A while you confirm Wi-Fi is active, then switch to Ethernet and run B on the same device.
COMMON QUESTIONS
Can the browser prove that I switched from Wi-Fi to Ethernet?
No. Browser APIs do not reliably expose the physical access medium. The workflow asks you to confirm the switch and then compares the two measured runs.
How should I make the comparison fair?
Use the same device and keep VPN, background traffic and test location as similar as possible. Only change the access medium if that is what you want to compare.
Browser APIs and network paths can limit what a web tool can observe.