It seems I came too late and missed the flames, bummer! haha. So I'm gonna ignore common sense and dip my spoon as they say here in Mexico:
Ste: it's not a matter of whether we believe your experience or not. The problem here is that, despite everything else, it's a subjective observation. How many times have I said "in my experience this happens this way", and when I get to actually measure it objectively, I discover it was not really so. That's why we have science (and I expect you above everyone else to understand what I'm talking about).
In other words, in the question of whether 8MB vs 16MB makes a noticeable difference, the only truly reliable answer is through a systematic, thorough benchmark. Of course, the reliability of those performing it can be put to judgment, but if they're thorough enough and elaborate clearly what they did, I think their results are better than subjective opinion - unless they're flat out lying about it.
Now, on to the actual issue.
I'm at work right now so I can't read the Tom's Hardware article, but I looked at the test results posted in the overclockers.net forum. Here are my observations:
1) Notice that the 8MB HDD graph has large, regular spikes in it. This usually means there was something in the background requesting HDD access during the test. The 16MB, on the other hand, is smooth in this respect. This alone might be a source of bias large enough as to make the test invalid. This might also explain the higher CPU usage.
2) Burst Rate is not, in my opinion, a good indicator of performance, since what you actually "feel" is the overall time it takes to transfer a file - which depends of course on the average transfer rate. In the test, the 16MB HDD shows a 12.4% increase in this parameter. Ignoring the aforementioned observation, I'd say this difference is noticeable, and considering the price difference between the disks, I'd say it's worth it.
Oh, and Havoc, it's spelled touché (I believe that's what you were trying to say). Do u c h e indeed means shower, and I don't know of any "bad" version of it - so it's weird that it's being caught by the profanity filter.