The question of whether or not the prostheses of the kind Pistorious has cause him to run faster than he otherwise would is fundamentally unanswerable, which is to say it is a FUQed question. A fundamentally unanswerable question.
In the language of causal inference popular in medical and social sciences...
To scientifically answer this question simply requires more data. In the most ideal, fake world, we'd want to be able to measure the same people racing both with and without prostheses under controlled conditions. The average difference between the with prostheses and without prostheses times would be the causal effect of the prostheses. The more data we have that is representative of some group, the more confident we can be in estimating a causal effect for that group.
Of course, we can't chop off people's legs and try. So the next best thing would be to compare groups who are on expectation equally as fast/slow who do have prostheses (the "treatment") and those who don't (the "control"). If we could do that, we could get a pretty good estimate of the effect of Pisty's Prosthetics.
But, we have an N of 1 for the treatment group--Pistorious. Which is problematic, because frankly, we have no empirical, replicable data to base our conclusions on.
N of 1 isn't always totally problematic. We could learn a lot, for example, if someone had an accident while in great shape, got proesthses, and instantly ran *a lot* faster. Then we have good treatment (post-accident) and control (pre-accident) measures. If we believe that such a person would be equal in every way except his prostheses before and after the accident, then we have a really good "case-control" type study. That might be another way to make a scientific conclusion. That hasn't yet happened, and the sample size would still be troubling.
That leaves us with the methods used so far by people like those on Science of Sport blog, which is basically story telling using the data that is available. This is the best we can do if we want a scientific appraisal of the effect of Pisty's blades. For example, one compelling argument they've made is that Pistorious speeds up throughout the race, and no other world class runner does. The able-bodied runners slow down because they fatigue. Thus, we conclude that Pistorious is at an advantage because he doesn't fatigue in the same way. Still, this is one data point, and itself is not convincing either way for a scientific conclusion. There may be outliers who defy the norm and race in such a way where they speed up. Maybe Pistorious is like that and has no advantage. Which may be UNLIKELY, but we'll never know.
At this point in time, the decision to decide whether or not Pistroius should be able to compete at the top level is much more an ethical one than a scientific one. Until there a lot more Pistoriouses out there for which we can try to piece together evidence about we're ultimately speculating.
That said, I don't think he should be able to compete in the Olympics and World Championships. All the problems I talked about above also make it unclear whether or not Pistorious is really "competing at the same thing" as the others. We don't have those doubts for able-bodied athletes.
Until we knew the effect of prostheses were on expectation zero, we should be cautious. (I would argue that if the effect is positive for some amputees, and negative for others, but washes out overall -- a zero effect -- then the amputees should be able to run. There's also sorts of human variance that determines performance!)