Distinguish among internal, external, and construct validity

Table 1.2 will help. However, be aware that students will not always understand these concepts the first time through. Fortunately, they will be re-exposed to these concepts many more times because these concepts are an organizing theme of the text. In the beginning, you might focus your discussion of construct validity on measures (using examples of invalid tests such as tests found in magazines and early intelligence tests). You will find that students easily understand external validity, especially if you use examples of studies that use rats as participants. In discussing internal validity, students will have a tendency to assume that a study has internal validity. You can prevent this by:

1. Telling students that the study has a flaw and their job is to find it and
   
2. Introducing obvious confounds such as having the treatment group being composed of seniors and the comparison group being composed of sophomores.

A good source of puzzles is:

Huck, S. W., & Sandler, H. M. (1979).

data based conclusions. New York: Harper and Row.

Another group-work activity that some faculty find useful is to have students work on end-of-chapter exercises 5-20. Alternatively, you could have students bring in conclusions made in ads or news reports that have questionable validity.

In addition, you could have students read the following article:

Seechrest, L. & Walsh, M. (1997). Dogma or data:

Students will need some help going through the article (and you probably only want to cover selected sections of the first 3 1/2 pages), but it makes the following points:

1. It is useful to look at studies in terms of internal, construct, and external validity (the organizing theme of this text)
2. Brief definitions of internal, construct, and external validity.
3. Several good points about external validity, such as
   1. Large sample size does not guarantee external validity.
   2. We should have specific reasons for doubting the generalizability of research. As the authors state, "Humans may doubt the generalizability to their own lives about the foraging strategies of rats, but they would not eat anything that causes a rat to drop dead."

Back to Research Design Explained Home Page