. Give three threats to internal validity. Explain how each harms the research process, and evaluate which one you think would be the most damaging.

Question

What will be an ideal response?

Answer 1

Ans: Internal validity refers to the degree to which the research was conducted in a manner that allows us to rule out alternative explanations. Often, researchers see this as the quality of the research—how much control is taking place and how well thought out is the design.—Campbell lists a number of threats to internal validity, and researchers need to think about these before designing a study. The three I am going to highlight are history, maturation, and instrumentation. History refers to something happening between experimental treatments. This something is out of the researcher’s control, but we can hopefully address it by having a large enough sample size to balance out the effects. An example might be conducting a study on trust in government and between the testing of group one and group two, a news report comes out that the governor committed fraud and is under investigation. With this one, sample size will not do it. Sometimes, you have to start again—or wait awhile for the next group. Maturation involves changes to participants that affect their performance. In one study on aging, the participants are contacted every 7 years to take measures of cognitive ability. When testing one participant who had been part of the study her whole life, she saw me at the door and said, “you again?” She was tired of doing the study and what looked like a decline with age actually came from fatigue. Finally, instrumentation involves changes in equipment and/or observers that affect the measurements that are made. Several years ago, one of my students was conducting her Research Methods project and told one group of participants that they would need to remember the emotional pictures when she came back in 2 weeks—and not the other group. This was not her manipulation and destroyed her study. Researchers need to be consistent across groups. I think all of them can be damaging to research, but the last one is under the researcher’s control to a large extent, and so that is the one I would worry about the most.