I dont trust toms with anything.
You know its pretty funny with all Tomshardware hatred and everythng. I personally don't go Tomshardware with their results but the way I see it, the only problem with them is that they don't put identical configuration to compared sometimes. Other than that, I think they are not skewing the results or anything.
If you look at their first Core Duo review where Core Duo laptop gets significant less battery life than Pentium M based one, if you look at the video card, Core Duo has X1600 mobility with 512MB of Video RAM while the Pentium M based one uses like X300 with 128MB Video RAM.
If you look at benchmarks with P4 chips, you see they overclock P4 chips while leaving AMD's alone, and that's their fault.
But I don't think they are really skewing the results.