But for any research, the researchers have to define the terms. Operational definitions are used in my field of research, for example. So, "hostility" might be defined as "how many times on person hits another" within the parameters of one research study.

True, a definition of "porn" might vary. In fact, we have even distinguished between "soft" porn versus "hard" porn. Thus, anybody analyzing the research would need to see how porn is defined within the parameters of the studies.