LECTURE 10.3: FIELD EXPERIMENTS

I. Why conduct field experiments?
Real people in real settings doing real-life behaviors

II. Examples of using random assignment in the field

III. Why not conduct field experiments?

A. Ethics

1. Informed consent

2. Unintentional harm

3. Privacy

4. Debriefing

B. Practical obstacles

1. Random assignment harder to execute than when you have sign-up sheets

2. Preserving independence of observations may be difficult

3. Difficulties in manipulating the independent variable

a. Standardization

b. Getting participants' attention

c. Manipulation check

4. Collecting dependent variable

a. Often use less sensitive measures in field

b. Recording DV not automatic

C. Power may be hurt because of problems discussed above (lack of standardization, insensitive dependent measure, participants not noticing the independent variable)

D. There is no guarantee that a field experiment will have more external validity than a lab study

E. Construct validity may be a problem because of

1. Not using manipulation check

2. Not using best available measure

3. Potential for experimenter bias

IV. The field experiment with interacting groups (workers in a company, students living in a dorm, students in a class)

A. Problems:

1. Internal validity:

a. "Powers that be" may not allow random assignment (but random assignment is one way of fairly distributing scarce resources).

b. Treatment-related mortality

2. Power, because independence is very difficult to maintain.

3. Construct Validity:

a. Diffusion of treatment

b. Compensation

c. Demoralization

B. Notable successes


Back to Chapter 10 Main Menu