I like real world tests. In other words, the testing should reflect the intended use of the knife, and even some unintended uses that might fall within the realm of reason. Chopping concrete blocks means little to me, but cardboard and blister plastic are very realistic substrates for a utility blade. If the knife is intended for use in the outdoors, then wood, hide and rope are very appropriate cutting mediums. I can learn, and possibly make comparisons with my own knives, if the substrates are similar.
Speaking of comparisons, I love them. Two knives dont necessarily need to be pitted against one another, but comparisons between certain aspects of knives are greatly helpful. One doesnt need to directly pit a small Sebenza against a Camillus EDC to expound upon the nuances of each design, but an understanding of the differences and similarities would make it all that much easier for people the decide if the Seb is right for them.
I also value comments that only come though extended testing. I knife can seem quite good during the first couple weeks of use, but time reveals those small weaknesses in design. What is just a minor annoyance at first can turn into a major drawback through extended testing. Some pocket clips wear through my blue jeans over a coarse of months. Hot spots are often revealed only after very long days of use. Scale materials are affected by the most unlikely substances. Ive been using my Spyderco Military for over a year now, and have yet to discover everything about it, good and bad.
And always remember that a review is about communication. You dont have to use fancy words or lofty statements, but you do have to keep the reader from falling asleep. Keep things interesting, and try to use a grammar and spell checker. You cant pick out all the mistakes, but youll get the majority of them. Now if the magazines would only use a grammar and spell checker.
In fact, get a photo of someone holding the knife while you perform this test 
