It all depends on what you are using the groups to determine. With sporter weight barrels, many calibers heat them up enough so that deterioration in grouping will result when shooting more than three shots in a group. As long as you specify how many shots there are in a group, so that comparisons can be made "apples for apples", no one should complain.
I get a little tickled when I see some "testing" that is done. Fellows may be shooting an unbedded action, off of a bad rest, from a wobbly bench, with no way to know what the shot to shot differences in the wind are, using a scope that has parallax, with an unaltered factory trigger of who knows how many pounds pull weight, with a seating depth that then describe as about a sixteenth of an inch shorter than touching, but by God, they are trying five shots each by tenth grain intervals, of their powder charge. I guess that my point is that as long as they are happy and satisfied, who is to say that they are doing anything incorrectly. After all, they made it to the range, and are having fun.
Back in the day, when I knew a lot less, I think that I had as much or more fun.
BTW when doing initial testing of some new component for my sub .2 capable 6PPC, I use 2 shot groups, if the flags looked right, and there is paper between the bullet holes, more shots isn't going to make the group less ugly. After I have good 2 shot results, I progress to three, and then five shot groups, and may eventually shoot several of those, since benchrest group competition is about averages (aggregates, and grand aggregates) of five shot groups.
Call me cheap, but with such short competitive barrel life, and the cost of other components, it seems to me that there is good reason to try to be efficient when testing.