Controlled experiments are required with development to the same contrast index. This is important not only to evaluate image structure, but curve shape and film speed. The addition of sodium chloride to a developer may reduce film speed.

A controlled test with reduced agitation is also required. A claim was made in this thread that stand development with Rodinal yields finer grain than continuous agitation. I haven't seen evidence supporting this assertion and I'm not sure what mechanism would cause this to be the case.