These benchmark is more complicated than you think
Firstly, we have to define which VPN server is tested.
Sometimes you have VPN with a foreign IP but this is just a "virtual IP" localised in your country. That mean you could get a Romanian IP physically localised in UK.
Lot of VPN use these bullshit technology which is not protecting user from metadata retention. If you ping a VPN server very far with a ping less than 75 ms that's bullshit. I know that's my VPN provider ActiVPN is having real servers in real physical location.
Secondly,
it's impossible to compare 4G and DSL network based on ping, for the reasons it takes different network route to the VPN servers and it depend on processor power an app loaded on the smartphone. If smartphone get a lot of app running, software will be less focused on VPN app, so ping will be increase.
Thirdly, on smartphone, there is not WebRTC protection which leak your real IP when you're surfing., so be careful. That's a bad idea to use it for web surfing, but it is prefect for eavesdropping protection including metadata.
Quaterly, comparing ping based on WiFi with a smartphone and a MBP, could be biaised by WiFi signal level on each devices, because we know that stronger WiFi signal is, stronger speed is, so the signal on each devices must be the same to make these test, is that right ?