Читать книгу Abnormal Psychology - William J. Ray - Страница 126

Science is a way of asking questions about the world. The quality of the answers we receive is influenced by several factors, one of the most important being the experimental design that we use. Somewhat like a blueprint, the experimental design directs the procedures and gives form to the experiment. In essence, an experimental design is a plan for how a study is to be structured.

In an outline form, a design tells us what will be done to whom and when. To be evaluated favorably, a design must perform two related functions. First, it must provide a logical structure that enables us to pinpoint the effects of the IV on the DV and thus answer our research questions. Second, it must help us rule out confounds as an alternative explanation for our findings.

Imagine a study in which a clinical psychologist was interested in determining if teaching children with autism to look at a person’s face would increase their interaction with others. Thus, the question asked would be whether looking at a face (IV) results in the participant interacting more with others (DV). After the subject was taught to look at a face, he or she could be placed in situations in which other individuals were present. The researcher could measure the number of interactions in which the child engaged.

If we were to diagram the design of this study, it would be as follows:

Select the group → Teach facial focus → Measure number of interactions

If we performed the experiment with just a single group, what could we conclude? We could determine if our participants had a certain number of interactions. However, this would not help us determine if this was related to the facial focus procedure. Such a design would not be much help in pinpointing the effect of the IV on the DV, nor would it rule out confounds.

A stronger design would use a control group. This design would appear as follows:

Experimental group → Teach facial focus → Measure number of interactions

Control group → No treatment → Measure number of interactions

Since the control group had not received the treatment, we would have stronger evidence that the treatment was related to differences in number of interactions.

null hypothesis: a statistical hypothesis that is tested to determine if there are differences between the experimental and control groups; the null hypothesis states that there is no difference

confound hypothesis: a conceptual question that asks if results of an experiment could have been influenced by a factor other than the independent variable (IV)

research hypothesis: the formal statement of the manner in which the dependent variable (DV) is related to the independent variable (IV)