Has UAC Ever Actually Been Put Through Controlled Tests

I am a unix / linux admin, so I understand the concepts of running under a restricted user account to protect critical system files and structures. I follow that UAC is Microsoft's version of this, with UAC acting a bit like a graphical version of "sudo". From articles I have read from Microsoft and others, the purpose of UAC isn't just to protect the user from themselves--which can be valuable in its own right in some circumstances--but also to limit the ability of any malware that may accidentally be acquired so that the OS itself is protected and any damage is restricted to that specific user's account and directories.has UAC ever been put through controlled testing to see if it actually works as designed? Has it been shown to stop malware from being able to make changes to the OS or system structures like the registry? What is its overall level of effectiveness in this--very effective, somewhat effective, or not very effective?