Chapter 4: Measurement and Units of Analysis
Similar to the threats posed by extraneous variables, a rival plausible explanation (RPE) is an alternative factor that may account for the results you observed in your research, other than what you might have been expecting. Threats to internal validity are considered RPEs. While it is true that most RPEs can be eliminated through careful research design (Palys & Atchison, 2014), it is important to acknowledge that some cannot.
For example, imagine that you plan a research project to study a downtown Vancouver community’s level of satisfaction with a safe injection centre that has been operating for a year in the community. You carefully design and plan your research project to eliminate threats to internal validity. Your research includes a mail-out survey to every community household registered on the Province of British Columbia’s most recent voters’ list. You also mail the survey to all community businesses. Shortly after your survey is mailed out there is a serious violent incident at the safe injection centre. A client has attacked and seriously injured a staff member at the clinic, but he was able to disappear from the clinic without being apprehended. This individual is still on the loose. How do you think this incident will affect the members of the community and the local businesses? How might this incident affect how your survey participants fill out the survey, as it relates to their feelings related to the centre? How might their survey answers differ, had the survey taken place before this incident, when there had been no such incidents? It is quite likely that this event will impact or “colour” the responses of your participants. In other words, there is now a strong likelihood that you have an RPE as to why the research participants have reacted negatively to the safe injection centre.
RPEs are serious, and while it is true that careful research design can eliminate threats to internal validity, the incident as outlined in the previous paragraph demonstrates how an RPE can sink a research project. As a researcher you spent a lot of time designing and planning your research, but essentially the findings are null, in this case, because you are not getting the true feelings of the community. Their feelings will have been negatively influenced by this recent incident. The researcher must decide how significant and how likely it is that the RPE influenced the results, in order to decide whether or not to scrap the research project.
While the preceding is an example of a blatant RPE, some are less obvious. Researchers must always consider the likelihood that an RPE explains the results of their findings when analyzing data. Less blatant RPEs (i.e. weather, postal strikes, a new government policy, recent media attention to an incident related to your research) must be discussed in the limitations section of the research findings.