THG LCD test method for response time

http://www.tomshardware.com/2005/0 [...] page5.html

"Again we'll recall our test method: the curve shows the different latency values for various levels of gray. A black-white alternation is shown on the curve by a point at X-axis 255, a black-gray alternation produces a point at X-axis 125, a black-gray alternation shows as 50, etc."

I understand the method explained here for testing response time and that it gives a better representation than that reported by the manufacturers under optimal conditions.

What I'm curious about is considering this method relys on greys instead of white, could further variation from these results be expected in reality when colours are used?

Would these variations be substantial?