LECTURE 14.1


SINGLE-N DESIGNS

I. Classic examples of the single-n design

A. Fechner

B. Ebbinghaus

C. Wundt and Titchener

D. Pavlov

E. Skinner

II. Advantages of the single-n design

A. More intensive study of each subject

B. Valuable when subjects or equipment is limited

C. Avoids two hazards of group data

1. What is true of group average may not be true of any individual subject

a. Average family has 2.3 children

b. Illustration of how if each subject learns in an all-or-none fashion that the group data could make learning appear to be a gradual, continuous process

2. Effects that might have a negligible effect on any individual subject may be declared statistically significant by a powerful large subject design

III. Logic of single-n design

A. Same as between subject designs in terms of

1. Covariation

2. Treatment preceding effect

B. Different from between subject designs in dealing with extraneous variables

1. Between subject designs: Unwanted variability unavoidable, best to make it random variability and let statistics factor it out

2. Single-n design: Variability caused by some factors; let's control those factors, thereby eliminating unwanted variability

C. Unwanted variability eliminated by:

1. Using only one subject to eliminate between subject variability

2. Establishing stable baseline in that subject

D. Establishing a stable baseline

1. If variables that control behavior aren't known, this may be a trial-and-error process and the investigator may try to isolate the subject from as many variables as possible (Skinner box).

2. If variables that control behavior are known, may simply involve keeping these variables constant (e.g., amount of reinforcement)

IV. The A-B design and its extensions

A. The A-B design, usually not recommended because change might be a coincidence due to maturation, history, etc.

B. The A-B-A (reversal design)

1. Makes coincidences less likely explanation for effect

2. Carry-over effects

3. Manipulation may not be reversible (brain surgery)

4. For ethical reasons, may not want to reverse treatment

C. Extensions of the A-B-A design

D. Similarities between the A-B designs (and their extensions) and the within subject designs

1. Order effects and coincidence as threats to internal validity

2. Sensitization as a threat to construct validity

3. Multiple treatments as a threat to external validity

V. The multiple baseline design

A. Internal validity

B. External validity

VI. Conclusions about single-n designs

A. How they differ from quasi-experimental designs

B. External validity


Lecture 14.2

Lecture 14.3

Back to Chapter Menu