Experimenter Expectancy Effect

Limits of Goodwill

Although research participants bear goodwill toward researchers, they may not (and often cannot) fake responses to please the researcher as implied in the SPOPE thesis.

To begin with, research participants might give untruthful responses only when illegitimate features in the research procedure render it necessary and possible. Second, it is not easy to fake responses without being detected by the researcher, especially when measured with a well-defined task (e.g., the attention span task). Third, it is not possible to fake performance that exceeds the participants’ capability.

Nonexperiment versus Experiment

Faking on the part of research participants is not an issue when experimental conclusions are based on subjects’ differential performance on the attention span task in two or more conditions with proper controls. Suppose that a properly selected [Page 453]sample of boys is assigned randomly to the two conditions in Table 1. Further suppose that one group fakes to do well, and the other fakes to do poorly. Nonetheless, it is unlikely that the said unprincipled behavior would produce the difference between the two conditions desired by the experimenter.

Difference is Not Absolute

Data obtained from college or university students do not necessarily lack generality. For example, students also have two eyes, two ears, one mouth, and four limbs like typical humans have. That is, it is not meaningful to say simply that A differs from B. It is necessary to make explicit (a) the dimension on which A and B differ, and (b) the relevancy of the said difference to the research in question.

...