Client Description; Targeted Problem and Goal
I have chosen for my single subject design a moderately developmentally delayed adult male who is 37 years of age. This man's name is "Hank," and he lives in a group home an the campus of the West Seneca Developmental Center in West Seneca, New York.

The problem that we will be addressing in our experiment will be Hank's poor self-care skills. All of the other residents in the home have at least one roommate, some of them are even in suites of three due to overcrowding. Hank has been placed by himself on a temporary basis as his room is always messy and unkempt. He always fails to make his bed and his clothes are strewn all over the floor. Hank, like all the other clients who live in this group home, has been judged capable of learning and of being integrated into the society. The center feels that before Hank can progress any further, this matter must be rectified.

The goal of our experiment will be to teach Hank that he will have to make his room more presentable by picking up after himself and making his bed on a daily basis without being told that this is what he has to do by a member of the staff.

Measuring the dependent variable.
We will have three dependent measures during all the phases of the experiment. First we will observe the frequency of two problem indicators: 1) Are there clothes on the floor? and 2) Did he make his bed today? Every Monday through Friday for the entire duration of the experiment, two trained observers will enter Hank's room while he is eating his dinner in the dining room. Both of them will make separate observations about the condition of his room. They will note simply "yes" or "no" for the two behaviors that we are trying to change. These observations will then be compiled into a checklist form which will become a crucial measurement in our assessment of to determine if our treatment has worked. We will use a third measurement method in this experiment: self-report. Hank has been involved in one-to-one counseling with the same psychologist at the center on a weekly basis for the last ten years. They discuss topics such as how Hank feels, how his workshops are going, how he spends his free time, whether he writes letters to his family, if he has made any new friends-- all of which Hank talks about willingly. Hank has always been very honest and all of his existing records, that we were allowed access to, showed that over their ten-year therapist-patient relationship, Hank has been honest and trustworthy. Access to Hank's private files was obtained through permission of his mother, who understood clearly after the treatment was explained and what we wanted to accomplish, that this would be in the best interest of her son. The research team will instruct the psychologist, who agreed to participate in the experiment, to ask Hank on a weekly basis (at their session) if he (1) picked his clothes off the floor Monday through Friday during the week and (2) if he made his bed that week?. The psychologist will then be asked to turn in a log which contains all the answers that Hank has given about each specific task that has been targeted. These are all measures of discrete responses. A discrete response has a clear beginning and end, so that a distinct instance can be counted (Skinner, 1966). Frequency measures have been used for such diverse behaviors as the number of: correct academic responses (Harris and Shuman, 1974) and aggressive acts (Horton, 1970) among others. The two trained observers, plus the psychologist's report from Hank will serve as reliable and valid measures of the dependent measure. The above-mentioned studies have demonstrated that frequency measures when used help to ensure the accuracy of the attained data.

Treatment sequence & explanation.
My design for this experiment is A - B - A - C - A where the letters signify the following:

Strategy B is the token-economy system. Hank is only moderately delayed and has been on the token-economy system before. He will be informed that if he performs the desired behaviors, he will receive tokens which can be used for extra desserts, more TV time, healthy snacks and longer visitation periods. His psychologist will be an additional enforcer of this behavior throughout the experiment. Hank will be given a token if he exhibits both of the desired behaviors. He can accumulate a maximum of 20 tokens for this period. In this treatment, Hank is being awarded for desired behavior,

Strategy C is the response-cost (withdrawal of tokens) which means that Hank loses his tokens for not meeting the two criteria we have set as guidelines for success. In this treatment, Hank is being penalized for not exhibiting desired behavior. Researchers will take back the tokens at the end of each day when the results, (for that day) have been assessed.

Both treatments have been cited in social science research. The token-economy system was used by Zimmerman and Russell (1969) and Kazdin (1971). The response-cost treatment was used by Kazdin (1971). These studies were cited in the textbook entitled Behavioral Assessment - A Practical Handbook by Michael Hersen and Allan S. Bellack, 1976, pp. 355-356, published by Pergamon Press in Elmsford, New York.

These treatments were chosen, both B and C because they have both been successfully used in cases involving behavior modification of mildly, mentally retarded adults.

The design is classified as a withdrawal design because both interventions or treatments (strategies) initiated can be and are withdrawn without some of the treatment remaining. You can both give out and take back tokens. Withdrawing tokens may be viewed as a disappointment or even a source of anger, but it is not unethical not dangerous to the client's well-being.

In the assessment of the threats to internal validity, I will list the possible threats and explain whether or not they warrant any concern.

HISTORY - This has not been eliminated; but the subject is not very exposed to outside events, and, if present, it is highly unlikely that they would influence room cleaning. Subject has been in an institutionally oriented setting all of his life.

MATURATION - This has been largely eliminated because Hank is already 37 years old, he is already an adult. Not much growing up will occur during the short duration of this experiment - 11 weeks. If he were maturing out of the problem, we would have noticed this during the baseline phase.

TESTING - This has been eliminated because unobtrusive forms of data collection are being used and the self-report will just seem like another question that Hank's psychologist asks him. There will be nothing to make him nervous or upset.

INSTRUMENTATION - This will be eliminated because there will be no variation in the way that the questions can be answered. "YES" or "NO" does not leave much to interpretation of results.

REGRESSION TO THE MEAN - This is also eliminated because Hank was not targeted because his housekeeping skills have recently suffered; he has been a consistent poor about picking up after himself.

SELECTION - The selection threat does not apply when there is only one subject in the experiment. Obviously, he is not representative of all people with poor housekeeping abilities.

MORTALITY (ATTRITION) - This was not eliminated as a threat to internal validity - only minimized because of the short duration of the experiment. Eleven weeks is a relatively short duration compared to twenty years, isn't it? Hank is a relatively young man, in good health, and not very likely to leave the facility as he has no one capable of caring for him, and his mother can not afford to hire someone on her limited income. The nature of the experiment is daily behavior patterns which he lacks, as stated in TESTING all forms of measurement are unobtrusive.

DIFFUSION OF TREATMENT - This is simply not a threat in single subject designs as there is no control group that the treatment can diffuse into.

PRACTITIONER INTERACTION EFFECT - It is possible that the manner in which the staff interact with Hank will have a significant impact on his progress or lack thereof. This can be tested in future experiments by changing the person who gives or takes back the tokens.

CAUSAL-TIME ORDERING - "One important limitation of this design is that it is not possible to determine which is most effective if both treatments show some effect (Barlow and Hersen, 1984)." If a person having a similar problem arises, we could reverse the order of the treatment strategies to see if more progress came with the first or second strategy.

Now to address the threats to external validity. It seems to me that many of the have been eliminated or are not very strong. There are no artificial settings, this is Hank's "Home Sweet Home." Demand characteristics are alleviated because, once again, he is not in any special setting that makes him feel like he should behave and act in a certain way. He is not representative of all possible subjects who reside in this home and definitely not representative of the entire population. Reactivity will not be a problem, as stated before because Hank will never be exposed to a team of "research scientists" or "specially trained personnel breathing down his neck and watching every move that he makes. Direct replication of this protocol is possible because of the simple nature of the design and because many residents exhibit problems with self-care skills. This experiment was practical, inexpensive and did not require special staff training procedures. There was no potential harm that could befall Hank by participating. Ethical standards were used whenever possible to obtain permission for researchers to view Hank's medical records and the findings have not been associated with his true identity.
 

		11/ 1	11/ 2	11/ 3	11/4	11/5	11/ 6	11/7	11/8	11/9	etc.
Observer1 (Day worker)		Y or N	Y / N	Y / N	Y / N	Y/ N	Y / N	Y / N	Y / N	Y / N
Bed made?	Yes = 1	0	1	0	1	1	0	1	1	1
Clothes in draws/closet?	No = 0	0	0	0	0	0	1	1	1	1
Observer2 ( at Night)
Bed made?
Clothes in draws/ closet?
Psychologist's report
Bed made?
Clothes in draws/closet?