A great deal here comes down to the simple fact that most of what's measured in a typical lens test has almost nothing to do with how that lens will perform in real life.
First of all, most lens tests emphasize resolution. This gives some idea of the largest print you could produce from a picture and still have it look sharp -- but doesn't tell you much (if anything) about how it will look when sized down to fit on the screen of a computer or tablet -- and the last I heard, that's the primary way of viewing a large (and growing) majority of pictures.
Second, for most people almost none of that matters anyway. Lens testing is normally done at the lowest ISO the camera supports. Many (most?) people routinely use an ISO considerably higher than that, immediately reducing their best resolution to quite a bit worse than what the testers would rate as really poor performance.
Third, even when/if the do shoot at minimum ISO, most people can't plan on getting even close to the resolution of test anyway. Neither autofocus systems or hand focusing will get you even close to the resolution shown in a lens test. Since the testers can't focus accurately enough, they don't even try -- instead they just bracket the focusing, taking multiple pictures (moving the camera minutely from one to the next) and choosing the sharpest one.
Along with that, they are (of course) doing quite a bit to ensure the sharpest possible picture -- mounting the camera extremely solidly, pre-firing (or locking up) the mirror, using a cable release, etc. Many people will buy a camera, use it for years, and discard it without ever, even once, taking a single picture with nearly the care that's considered the bare minimum for a lens test.
Fourth, there are rather specific accepted rules about how lens tests are done, some of which keep the tests from having much to do with how the lens will work in real life. Just for example, it's generally accepted that you focus for maximum sharpness at the center of the frame, and then measure the results for the rest of the frame with that same focus. This will show a lens that has field curvature as having extremely "soft" edges/corners. When you take real pictures of three dimensional subject matter, you may easily see results that are exactly the opposite of what the test seems to indicate (i.e., a lens that looks bad in the test looks great in real pictures, while one that looks pretty good in the test doesn't look nearly as good in real pictures).
Bottom line: it's not just a matter of subjective opinion versus objective measurements, differences in standards, or anything of that sort at all. In reality, most of what you see in a typical lens test tells you next to nothing about how good of pictures that lens will produce in real use.
Most of the time when people get pictures that aren't sharp, it's caused by missed focus or camera movement (or both). Blurry pictures from an 18-200 lens will most likely stem from the fact that these lenses are fairly slow (small maximum aperture). Unless you have quite bright light, this will lead to either boosting the ISO or hand-holding at too slow of a shutter speed. Either will typically lose far more sharpness than lack of lens resolution.