Single N Design Activity/Experimental Psychology (PSYC/CHLD 3404)

Small N designs generally involve a large number of observations on a single or small group of participants.
This design is historically significant and still important today in the experimental analysis of behavior (behaviorism), psychophysics (learning how senses work in conjunction with physical stimuli), and clinical psychology. An advantage of small N designs is that they provide good ideas for future research, they are a good method of studying rare phenomena, and they can avoid some of the ethical and practical problems with control groups in clinical settings. That is, it may be unethical or not practical to use a control group when people need help with a psychological problem. Unfortunately, there are also drawbacks to this type of design. For example, researchers cannot draw causal conclusions because there is no random assignment and there are other issues with validity.
Small N experiments are created using the same logic as within subjects or repeated measures designs. Typical designs include the ABA or reversal design, where
A = observation or baseline phase
B = introduce treatment or manipulation
A = withdraw treatment or manipulation and measure baseline again
The second “A” or baseline observation is necessary to interpret the effect of the treatment. If the behavior doesn’t revert back to baseline levels then you won’t be sure that the treatment was caused by the change, treatment, or manipulation.
There are two problems with the ABA design. The first problem is ethical. You may not want the client to finish in the baseline phase. In fact, you probably want them to be treated. The second problem is more practical. Some treatments are irreversible. Thus, when you return to the baseline phase, the behavior remains at treatment phase levels.
The ABAB design addresses the first problem. There are also extensions of the reversal design – the alternating treatment design. This involves A B A B BC B BC whereby one introduces a second treatment or manipulation to determine if it has an effect in addition to B alone.
Part 1. Read the case1 below and then answer the questions that follow with your team.
The client was a 26-yr-old housewife who was seen at the Anxiety Disorders Program for complaints of a severe fear of frogs. The onset of her fear seemed to stem from a traumatic incident 18 months earlier when she was mowing thick grass in the yard of her home on a riverbank. She apparently ran over a group of frogs, as she was suddenly surrounded by a number of the hopping animals and at the same time several were chopped up by her lawn mower. Bloody pieces of frog spewed from the machine. She became so upset that she had to stop mowing the grass and go inside her home. She had not been able to mow the yard since then. Her life became a misery. In the evenings she could hear croaking frogs at the riverbank along the back yard. On several occasions frogs found their way into her home, forcing her to leave the house until they had been removed by a neighbor or her spouse. She had occasional upsetting dreams about frogs which led to insomnia. She was happily married and was in her third trimester of pregnancy when interviewed. She fulfilled DSM criteria for simple phobia.
1 The case was taken from Thyer, B. A., & Curtis, G. C. (1983). A repeated pretest-posttest single-subject experiment: A new design for empirical clinical practice. Journal of Behavior Therapy and Experimental Psychiatry, 14, 311-315. The treatment was not taken from the article.
The therapist used exposure therapy plus reward. This involves gradually increasing one’s contact with frogs under guided conditions. The clients rated her anxiety level using the subjective units of distress scale (SUD). At baseline, the client was seated in a room and told that a live frog was in the next room. She rated her distress from 0 (complete calm and relaxation) to 100 (complete panic or as frightened as she has ever been). The assistant then entered the room at a distance of 15 ft with a frog in an aquarium on a cart. Ideally, the assistant would come 2 or 3 ft closer and closer until the client was touching the frog. During the treatment phase, the client was gradually exposed to the frog and was provided with a brief reward – a two-minute neck massage provided by a professional masseuse.
Research design: The therapist used an ABAB design with several observations per treatment phase. The data are displayed below. The x marks the amount of fear the patient experienced as the frog was at different distances from her (from 0 to 130). Note: OR = outside room, DT = direct touch, HOLD = hold, UC = unlimited contact, and numbers = feet away.
A1 A2 100
100
X
X
X 90
X
X
X
Could not continue 90
X
X
80
X
80
X
70
X
X
70
X
X
60
60
X
X
50
50
X
40
40
30
30
20
20
10
10
0
0
OR 15 12 10 8 6 4 1 DT HOLD UC OR 15 12 10 8 6 4 1 DT HOLD UC B1 B2 100
100
90
X
X 90
80
X
80
70
X
X
X
X
70
X
60
60
50
X
X
50
40
X
40
X
X
X
X 30
X
30
X
X
20
20
X
X
X
X
10
10
0
0
OR 15 12 10 8 6 4 1 DT HOLD UC OR 15 12 10 8 6 4 1 DT HOLD UC
1. Are there any ethical issues or concerns related to the research?
2. Can you think of a second way to operationalize the dependent variable?
3. Will the research generalize? Of the possible problems below, which will present the largest threat to external validity?
a. Ecological validity—it may not happen the same way in real life.
b. Robustness—it may not replicate in different settings
c. Robustness II—it may not replicate with different people
d. Robustness III—it may not replicate using different operational definition of the measures.
e. Are the findings relevant and useful for others’ with phobias?
Provide a rationale for your answer:
4. Are there any threats to internal validity? If so, which of the following will present the largest threat to internal validity?
a. History
b. Regression to the mean
c. Mortality/attrition
d. Practice effects
e. Maturation
f. Experimenter bias/demand characteristics
g. Selection
h. Instrumentation
Provide a rationale for your answer:
5. Was the treatment effective?
a. Yes
b. No
Provide a rationale for your answer:
6. Can you infer “cause” using this design? Why or why not?
7. In general, what are a few things that must be present to infer cause? Please answer as if you were talking with a non-psychologist.
8. Is there a possible third variable (or confounding variable) that is responsible for the pattern of results above? List as many possible third variables below.
9. Imagine that I extended this design by including your third variable alongside the treatment and then varied it so that design was as follows (where the 3 is your additional variable):
A B A B (B+3) B (B+3)
Could you then infer cause? Why or why not?
Part 2. Design a study
Imagine that you were hired to treat a young person with obsessive compulsive disorder. Specifically, the young person was leaving class multiple times per period to wash his hands. Research evidence suggests that the following therapies have been effective:
? Gradual exposure to germs
? Reward/reinforcement for non-washing behavior
? Cognitive behavior therapy
? Drug therapy
Design a single N study using one or more therapies.
1. Describe your study (operational definitions should be discussed in addition to procedures) and chart what progress you expect (by charting the frequency of hand washing on the Y axis and time over ABA intervals on the X axis).
2. What internal validity threats do you foresee in your treatment design? Develop methods to eliminate validity threats.
3. One possible problem with single N designs is that it is difficult to establish a stable baseline. How would you develop a stable baseline?
4. Are there any potential third variables that might help explain your results?