Here are some tests you can do at the store if they have LCD panels hooked up to computers- whether or not the results concern you or not is entirely up to your own tastes:
•Open an explorer window- hopefully, the choice of desktop image is such that the explorer window stands out color-wise
•Now use the mouse cursor to drag the title bar of the window in tight circles and slow to medium pans across the width of the screen
•Follow the movement of the window with your eyes
•Do you see a subtle trail that follows behind the sharp border of the window as it moves? You may also see a shift in color along the moving border, as well.
•Now follow the movement of the text in the title bar as you move the window- does the text stay sharp, focused, and legible or does it collapse into a blur once you exceed a certain speed of motion? This is essentially testing the ability to maintain screen detail under motion.
•An alternative to the moving explorer window is if there happens to be a screensaver mode with long sweeping and predictable motion. Watch the detail of things in motion- especially the leading and trailing edges of the object.
If you do the same tests on a CRT nearby, you will strain to see any trail in the wake of moving window border, and text will remain rock solid clear under motion, within the realm of what is capable by the Windows screen rendering mechanism. You will (or should?) find that any LCD panel you try will fall down on these tests. It doesn't matter if it is 21, 16, or a blistering 8 ms panel. The LCD panel will show the artifact and the CRT will not. The very notion that the "8 ms" rating of the LCD panel should theoretically make it twice as fast as a typical CRT, yet cannot skirt the artifact test should tell you (or at least suggest?) how very little the transient response ratings are worth when they are spec'd for LCD's. They may be useful when judging amongst LCD panels, but are utterly meaningless when compared to "typical 16+ ms" CRT devices. When they say "16-20 ms" for a CRT, they mean it actually performs in that range. When they say "16 ms" for LCD panels, it is the best value from a specific shade to another shade, and all other shade transitions get considerably worse as you move away from that "best shade" figure. Worst case values for LCD's could fall as far out as 30-40 ms, even though it is rated at "16".
Now, my saying this is not to tell you to buy one or the other- just to give you another tool to play with in your own personal shopping evaluations of in-store demos. As prefaced at the beginning, the effects you see may or may not be important to you. Maybe you don't see the artifact behavior as having any relevant impact to the way you plan to use your display. Maybe it will irk you on principle that displays that perform like this can be brazenly spec'd as "8 ms", for the sake of all that is good in this world. Maybe it will be an interesting thing to make note of, but generally inconsequential in the big scope of things. Maybe you find that this "test" is rather contrived compared to actual useage patterns, but do note that, deductively, the artifact behavior will present itself under any other motion conditions that you will encounter in normal use that happen to be very similar to the test conditions, not just the test condition itself. It will affect everyone differently based on what is important to them.
I only bring this issue up, simply to bring attention to its existence. If an LCD panel will suit your needs, then it suits your needs, and I say more power to you.