Ph.D Program
Instructions for the Comprehensive Examination in Clinical Psychology
Questions or comments, please send e-mail: John Kounios, PhD
Comprehensive Examination
Clinical Psychology Doctoral Program 2005
Link to Video : .MOV (25 MB) | .MP4 (32 MB) | .WMV (29 MB)
Procedures for Comprehensive Examination
Clinical Psychology Doctoral Program 2003
Link to Reading List (PDF)
Comprehensive Exams will take place during the week of June 23, 2003. Students will engage in one day of written examination and one day of oral examination. Two raters will score each written question. Two raters will score the student on the oral questions. A student will not have the same raters for all questions. All departmental faculty will be asked to participate in areas corresponding to their expertise.
Day 1 (written examination) will be comprised of the 4 questions involving 3 content domains:
" 2 General Questions (2 choices per question)
" 1 Clinical Question (2 choices per question)
" 1 Research Methods/Statistics Question
Students will have 80 minutes for each question and students will have choice within content domains.
Day 2 (oral examination)
" 1 Clinical Case including Assessment
" 1 Research Method Question (Design a Study Format)
The oral examination is designed to assess skills and content that are not covered in the written portion. Thus, it will provide a means to assess unique content that is not covered in the one written question and a different modality that will enable a greater elaboration of student's responses.
Students can choose the type of case for discussion (General Clinical, Neuropsychological, Health Psychology, Forensic Psychology)
Students should arrive 1 hour in advance to study the materials (30 minutes for the clinical and 30 minutes for the research methods). The Oral Examination will take 30 minutes to one hour.
Passing Scores will be determined on the following:
General will be averaged across two items. Clinical will average written score with oral score. Research Methods will average written score with oral score.
Sample Scoring Form
A minimal score of 3 for each content area is needed to pass a given content area. The General score will be based on the average ratings of the two questions. The Clinical score will be based on the average written and oral score. The Research Methods score will be based on the average of the written and oral score.
Each score of each question will be determined two ratings.
Example:
General Question 1.
Poor Passing Excellent
Rater 1 1 2 3 4 5
Rater 2 1 2 3 4 5
Average Rating 1 2 3 4 5
General Question 2.
Rater 1 1 2 3 4 5
Rater 2 1 2 3 4 5
Average Rating 1 2 3 4 5
General Question Score (Average of Q1 and Q2 ): ______
Comments:
Oral Scoring form
Clinical
Poor Passing Excellent
1 2 3 4 5
Avg. of 2 Raters:
1. Summarized the presenting problems and symptoms.
2. Determine the general diagnostic category of the case.
3. Described the assessment plan.
4. Summarized the unique feature of the case.
5. Case conceptualization.
6. Described the treatment approach.
7. Described any special problems.
8. Described empirically validated treatments.
9. Described expected outcomes of treatment.
10. Described ethical or professional issues.
General Clinical Score
(Rater 1+ Rater 2/2) _____
Research Design/Statistics
1. Importance, quality and clarity of the hypotheses.
2. Appropriate time-line.
3. Internal Validity of the study.
4. External validity of the study.
5. Sampling adequacy.
6. Accuracy of measures.
7. Integrity of treatment delivery for treatment outcome studies
8. Quality of treatment delivery
9. Power
10. Statistical Analyses
11. Effect Size
12. Clinical Significance
13. Ethical or Professional issues.
General Research Score
(Rater 1+ Rater 2/2)
The Clinical Section
The clinical examination first involves reviewing clinical case reports that are given to the student at the start of the study period. The student then chooses a case to focus on for the oral examination. The remaining cases in the packet are not used in the examination. Here is the outline for the oral examination:
- Summarize the presenting problems and symptoms of the case
- Determine the general diagnostic category of the case
- Describe your assessment plan
- Summarize the unique features of the case that may deviate from typical cases
- Specify your case conceptualization
- Describe the treatment approach you would use with the case
- Describe any special problems you think might arise during treatment and suggest treatment approaches to these problems
- Describe your understanding of empirically validated treatments that may apply to the case
- Describe the expected outcomes of treatment
- Describe any ethical or professional issues that are associated with the case
Sample Clinical Case
SANDRA
BACKGROUND
Sandra is a 43-year-old woman who has had several psychiatric hospitalizations for psychosis over the past 15 years. Six weeks ago the patient's favorite uncle died suddenly. He had helped raise her when she was a little girl and had always been supportive during periods when she was ill or under stress. Sandra was initially sad, withdrawn, and mildly depressed. Her husband had some difficulty getting her out of bed in the morning before he went to work.
Approximately 3 weeks ago her behavior began to change. She became brighter and more outgoing. Unfortunately, the changes in her behavior did not stop with a reestablished optimistic outlook on life. Within 2 weeks her energy knew no bounds. She began sleeping 2-3 hours a night and then awoke to clean the house. She decided to become more involved in her 16-year-old daughter's school and called the principal early one morning to volunteer her services. Her daughter was not pleased.
One day Sandra went to the supermarket and returned with three cases of frozen orange juice, explaining, "You can never be too healthy." Sandra's behavior continued to change and she became more irritable and irrational. She talked endlessly but less and less of what she said made sense. She began to be suspicious of her family and friends. She told her husband, "I've solved the secret of aging and now they're trying to steal it from me."
This pattern of changing behavior was consistent with her past psychiatric history. In each of her previous hospitalizations the patient became overtly psychotic and delusional. She was treated with electroconvulsive therapy (ECT), responded well, and was discharged with a diagnosis of Schizophrenia on low doses of antipsychotic medication. She still had some delusions but these generally remitted after 2-3 months and the medication was discontinued. The patient had recurring episodes of illness every 3-5 years. The onset of active illness was often correlated with the loss or separation from a loved one. The last episode occurred shortly after the death of her mother 5 years before. Between episodes, the patient returned to her usual activities and remained functional, without psychiatric symptoms. However, she stopped working as a nurse after her second hospitalization.
The patient's current psychotic delusions and irritability continued and increased. She took offense at the slightest comment and started yelling at her husband. He, in turn, tried to get her to see a psychiatrist but she refused. Finally, in desperation, he took his wife to the emergency room of a nearby hospital where she was admitted to the psychiatric ward. In the hospital the patient continued to be irritable, delusional, paranoid, and critical of everyone around her. She left bizarre poems and notes for her physician each morning: "E = MC2 = love makes the world go round and Jet power builds strong minds and bodies." The poems were decorated with chains of daisies drawn along the edges.
The Research Section
In this section, the student will be given a series of questions that request the student to design a study. The student is requested to choose a subset of the questions and formulate a response to each one. For example, here is a typical question:
Design a study examining the effectiveness of a response-cost procedure in the management of aggressive behavior among children in an inpatient treatment program for children with conduct disorder.
Such requests imply that you know the basics of the treatment under study. However, the examination team will not expect you to know the intervention in detail or the current literature about its effectiveness. The purpose is to examine your knowledge and ability in designing such a study, not in your knowledge about the treatment (e.g. response cost procedures for aggressive behavior). The evaluation form that guides the testing is based on the criteria domains described below. These focus on the research design and not the particular content of the study.
Prepare a design paying close attention to all of the relevant design issues. There are three points you must keep in mind:
1) Prepare a study that is concise and focused on a limited research question. Do not propose a large, comprehensive study that will be difficult to describe and defend.
2) Consider very carefully all types of studies presented in the questions before you choose which ones to answer. In general, choose the study type and research area that is the most familiar to you.
3) Manage your time! Do not spend 20minutes on the first criterion domain and then have no time for the others.
Criteria domains for structuring the proposal. You will be asked to make responses in each of these areas and you will be rated by the commitee in each of these domains:
- Significance and clarity of the study hypotheses and intedependent variables
- Appropriate time-line and other practical aspects of conducting research
- Threats to the internal validity of the study.
- Threats to the external validity of the study.
- Sampling adequacy.
- Choice of dependent measures.
- Integrity of treatment delivery.
- Quality of treatment delivery.
- Power.
- Statistical analysis.
- Clinical Significance.
- Ethical concerns
- Minority and cultural diversity issues
A General Summary of Threats to the Validity of Research Designs
Here is a brief and very concise summary of the major issues involved in the analysis of all research projects. As you ponder the research problems presented to you, use this summary as a general guide. Keep in mind that many of the defects cited below are interdependent and not mutually exclusive. A great general reference in this area is Research Design in Clinical Psychology by Alan E. Kazdin.
Here are the defects that might characterize any study:
1. Poorly formed hypotheses. How clearly do the investigators express the research questions and the independent variables? Of course, this has consequences for every aspect of the study design.
2. The presence of extraneous factors that confound the experimental comparison. This general heading covers a multitude of internal threats to the validity of an experiment. A good example is represented by the use of a placebo control in most drug treatment studies. The expectation that we should get better when simply administered a pill or injection works as an extraneous factor and confounds the interpretation of the effectiveness of a new drug. Hence, the control group is given an ineffective pill or injection to control for the placebo effect.
If a pre-post comparison is made, extraneous factors may influence the dependent measures during the time interval between the pre and post comparisons. These are usually referred to as history-related effects.
3. Sampling recruitment biases. If samples are recruited in some nonrandom fashion then study outcomes may be only generalizible to a limited population. If investigators are unaware of the bias then the results are usually incorrectly interpreted.
4. Sampling selection biases. Here, the investigators use unreliable or invalid assessment instruments to select subjects. If an investigator selected a sample of depressed subjects for an experiment of a new antidepressant but used an invalid measure of depression to select the subjects, then the study findings may not be generalizible to the population of depressed people.
If a study has matched subjects on some extraneous variable, such as age, education intelligence etc., the measures used for matching may be unreliable or invalid. This results in a selection bias.
5. The dependent measures do not correspond to the hypotheses under study. Judgments concerning this factor require a sophisticated theoretical model of the constructs assessed by the dependent measure and the hypotheses of the study. As an extreme example imagine a study in which the outcome of a new treatment for depression was assessed using all of the scales of the MMPI.
6. Statistical power is too small. Power is expressed as an interaction of effect size and sample size. Most studies have too little power because of small sample size.
7. Sample size is too small. This results in low power and poor generalizibility of the results.
8. Poor choice of statistical tests. This is usually uncontroversial and most investigators use conventional methods. However, you should be familiar with the features of the common statistical procedures.
9. Effect size is not examined or characterized. Statistical effect size and clinical significance are two distinct but related concepts. The effect size is simply the magnitude of mean difference divided by the variance in the samples. This mean difference may be statistically significant but still too small to meet the expectations of the practitioners using the treatment or the patients receiving it. The reason there are so many headache treatments available on the market is because the FDA does not have a mechanism for evaluating effect size. Anything that shows statistically significant results gets to the market. The clinical significance required of a headache treatment might be one that expects the treatment to eliminates the pain. A treatment that reduces the pain but does not eliminate it, may have attained statistically significant findings on a self-report rating scale of pain. However, since the field expects a treatment to eliminate the pain, the treatment does not meet an effect size standard of clinical significance.
10. The independent variable levels and type are not generalizible to the manifestations of the independent variables in the real world. For example, the dose of saccharine given to laboratory rats that resulted in cancer does not exist in the real world. In addition, the generalization of findings from rats to humans also may be influenced by differences between the two species. Medical research studies have generally discovered that rats are a lot tougher than we are. Independent variable manipulations that use animals may not be generalizible to humans.
Questions or comments, please send e-mail: J. Michael Williams, PhD