I like Macs, and the tests showed Mac in a good way, but those "tests" were pointless and all over the map. Really, there is a large difference between SL and Win7. These types of tests need to be revamped, and conducted later when both OSes have a little time to mature.
Keeping in mind no test will ever be perfect, you have to make some decisions about what you want to test. Do you want to test SL & Win7 performance on Apple hardware, or do you want to test performance on *equivalent* hardware? Are you testing the hardware or the OS? Are you going to test the OS or test Apps running on the OS? Testing with Apps shows the performance of the App on a particular platform, not the performance of the OS. Choose what you want to test for.
If you want to test the OS, then boot/shutdown times are good. Disk access, network performance, I/O, graphics performance and other subsystems are what you look at.
Separately you can test App performance... but seriously, you need to find 3rd party Apps, but those where the developer has optimized for BOTH platforms, not just ported generic code from one to the other.
Based on Windows to Windows testing conducted elsewhere, Win7 is faster than Vista but slower than XP. MacOS X has been shown to be consistently faster in some functions than Windows, but behind in some others (like graphics performance ironically).
What do you need your computer to do? Games will likely always be faster on Windows cause developers optimize for DirectX. Other Apps such as video editing run exceptionally well on Mac. When serious developers optimize for multi-core processing that MacOS does very well, those apps run better on Mac. When developers start to take advantage of the full 64bit, GCD, OpenCL, etc. features of SL, then those particular apps will really spank similar/equivalent apps on Win7.
There is no blanket "faster" system. It's case by case, over time.
Couldn't agree more.