if there are three or more levels associated with that particular factor
a multiple comparison test is appropriate to find out where the signifi- :
cant differences lie amonf the main effect means. If a first order is sig-
nificant, researcher will often disregard the. results related to the main
effects of the two factors that are involved in the interaction. Instead

they can average the data across the variable that is not involved in the

interaction then apply tests of interaction.