Generating hypotheses (3-4 pages, 10 % final grade)
The purpose of this assignment is to encourage you to draw on your own interests in psychopathology to begin to develop a research topic. Towards this end, you will be asked to (1) review case studies about different psychological disorders; (2) use these case materials to identify interesting questions about a particular disorder; and (3) develop specific hypotheses, using the guidelines for good hypotheses presented in class.
Good clinical research often is sparked by personal experiences, contact with clients, and case studies that raise interesting questions about different mental disorders. For this assignment, you are encouraged to adopt a scientist-practitioners mindset. In practical terms, this means that you will reflect on the rich material presented in an individual case and then draw on these reflections to generate broader questions about a particular disorder (risk factors, associated features, nature and course of illness, prevention, treatment, etc.). Be sure to pick a topic that truly interests you, as this will help you think creatively when generating your hypotheses and keep you engaged in later assignments, as you will be asked to elaborate on this topic in Assignments 3 and 5.
Here are the steps you should take in completing Assignment 1:
1. There are four cases-study books on reserve for this course (at the Reserve Desk in the UgLi). Browse through these cases and select one that you find particularly interesting.
2. Generate two hypotheses about the clinical syndrome presented in your case study. As outlined in lab and in lecture, many different types of questions are important in clinical research: questions about risk factors, associated features, nature and course of illness, prevention, treatment, etc. Any of these questions is appropriate.
3. When outlining your hypotheses, be sure to follow the guidelines for good hypotheses that are presented in lecture.
4. Be sure to present material from the case study that helps us understand why this particular case captured your interest and why you think it supports your hypotheses.
ASSIGNMENT #2
Assessing behavior - (6 - 8 pages, 15 % final grade)
This assignment challenges you to think critically about the process of measurement in research by (1) designing a new observational measure with your lab section; (2) conducting your own observations using both this new measure and qualitative methods; (3) reflecting on the relative strengths and weaknesses of quantitative observations; and (4) comparing quantitative observational assessments with other types of measures, in particular, self-report questionnaires.
The work for this assignment will begin in your labs, when you will collaborate on a new observational measure for rating some kind of interesting behavior: social aggression, empathy, etc. To complete this assignment, you should participate in this group effort. Outside of class, you then should find some time to conduct your own observations using this measure in a field setting (a social gathering, a playground, etc.). Youre encouraged to go on these outings with your lab-mates and to discuss your experiences with each other, but you each should report on a different sample of behavior and write your own reports.Here are the steps you should take for writing up your report.
1. Give a brief summary (one page or so) of your observational measure and your goals in using this assessment: What behaviors are you trying to assess? Why is this behavior important? What are some of the challenges you face in trying to assess this type of behavior?2. Find someone to observe. Conduct your observations in a field setting that your lab has identified as a good place for using your new measure.
3. Write a qualitative description of a 10-minute segment of behavior. This means that you describe what you see in a narrative, telling us how the person behaved, what events preceded the target behavior, what were the outcomes of the behavior, etc. (Please note: You dont have to describe everything that your subject did, just the relevant behaviors. To keep this part of the task manageable for you, please limit this section to no more than 2 pages. )
4. Determine the score you would give this particular segment of behavior if you were coding it using the observational system that you developed in your lab. Tell us why you decided on this score.
5. Reflect on your experiences by answering the following questions:
6. In your opinion, does this numerical score say something meaningful about the behavior you observed? How does it compare to your qualitative narrative in assessing the target behavior in a meaningful way?
7. Given your experiences, do you think quantitative observations are useful in clinical research? If so, how? If not, why not?
8. If you tried to assess the same behavior with a self-report questionnaire, how do you think your results would differ? Compare and contrast the relative benefits of self-report questionnaires versus quantitative observations when trying to assess the behavior that is targeted by your labs observational system
Conducting a literature review
(6-8 pages, 10 % final grade)
This assignment is designed to give you some practice in conducting a literature review and in reviewing research articles. You are being asked to find two journal articles that report on research related to the topic you identified in Assignment 1. In your work on Assignment 3, you will begin to summarize, organize, and synthesize information from your two sources as you think ahead to designing your own study (Assignment 5). This assignment also encourages you to hone your psychology writing skills, so we encourage you to refer to the APA guidelines for the effective expression of ideas, on reserve for this course.
Refer back to Assignment 1, and select one of your two hypotheses. As you will be designing a study to test this hypothesis in Assignment 5, you first need to conduct a literature review to see what other work has been done in this area.
Using PsycInfo and other sources, conduct a literature review on your topic. Be sure to seek guidance from your GSI about your literature search and selection of keywords, as this can save you lots of time. It is especially important to touch base with your GSI if you find that your searches are turning up very few or too many citations.
Track down your two references in Hatcher (or elsewhere) and make copies for yourself. Both of these references should be research reports that are published in professional journals. (See list of suggested journals in this syllabus.) NOTE: Chapters and review articles are not appropriate for this assignment. Your references should report on findings from a single study or a series of studies. You also will need to turn in copies of these references with this assignment.
Select your two references thoughtfully, as they will guide your thinking for Assignment 5. In general, they should provide background information on work that has been conducted in your area of interest and suggest measures, methods, and interesting questions for future research. You are strongly encouraged to seek feedback from your GSI in selecting your references! This early guidance can save you lots of time and many headaches on this assignment and on Assignment 5. When you meet with your GSI, bring in your hypothesis and a printout of abstracts from your PsycInfo search.
Write a summary of each of these journal articles using the following outline format on the next page.
GUIDELINES FOR ASSIGNMENT 3:
I. Introduction:
II. Participants:
Procedure:
IV. Measures:
V. Results:
Discussion:
Data Analysis; Results Section
(2-4 pages, 10 % final grade)
The purpose of this assignment is to gain some hands-on experience with data, statistical software, and analyses. In the process, we hope that you will come to regard statistics as tools to be used to your advantage, rather than as an exercise that inspires fear and loathing. This assignment asks you to answer 5 questions using SPSS statistical software. (We will provide the data set.) Your GSIs will walk you through each of these analyses in your lab sections, which will be held in computer labs outfitted with SPSS.
IMPORTANT: It is critical that you attend your lab sections on descriptive and inferential statistics to complete this assignment. If youd like to have some extra help, you also can request to sit in on another lab section taught by your GSI. In labs, you will receive a handout describing how to use SPSS software to run these analyses.
REMINDER: This assignment is to be turned in DURING LECTURE on the Monday before Thanksgiving, November 20, 2000.
Brief overview of the study:
This study was conducted because the researcher was interested in examining the effects of maternal depression on child development. The researcher recruited depressed mothers from the University of Michigan Mood Disorders Clinic. All of these women were in weekly outpatient psychotherapy, and they received information about the study through flyers distributed by their primary therapists. Comparison participants were recruited from the Ann Arbor community through newspaper advertisements. All participants were compensated $25 for their time. Data were collected during home visits, when trained research assistants conducted interviews with mothers and children separately.
Variables in the data set:
subject: This is the identification number for each participant.
sex: Childs sex. 1 = male; 2 = female.
kidage: Childs age
momdx1: Maternal diagnosis of depression. 0 = no depression diagnosis, 1 = depression diagnosis.
momdx2: Maternal diagnosis of anxiety. 0 = no anxiety diagnosis, 1 = anxiety diagnosis.
cdi: Childrens Depression Inventory (CDI), a self-report questionnaire on depressive symptoms completed by children. Higher scores indicate more severe depression.
cbcl_i: Child Behavior Checklist (CBCL), anxiety/depression subscale. This questionnaire was completed by mothers. Higher scores indicate more severe depression.
social: social problem solving. Higher scores indicate better problem-solving skills.
gpa: Children's grade-point average.
For this assignment, please run the following analyses then report your findings, using no more than four or five sentences for each question. Be sure to follow APA guidelines for reporting statistical analyses.
1. First, describe the children in the sample. How many boys and girls are there? What is the age range of kids in this sample? What is the mean age of the kids in this sample? How many kids lived with depressed moms, and how many kids lived with non-depressed moms?
2. Clinical researchers often struggle with issues of overlap among different clinical syndromes. Often it is difficult to study "pure" cases of a particular disorder, because individuals can receive more than one diagnosis. When two disorders tend to co-occur, this is known as comorbidity. In this sample, conduct a chi-square analysis to determine whether maternal depression is co-morbid with anxiety among the women in this sample.
3. This researcher expects that children of depressed mothers are more likely to experience symptoms of depression. She used two different measures to assess depressive symptoms in children: The Childrens Depression Inventory (cdi), which is a questionnaire completed by children, and the Child Behavior Checklist (cbcl_i), which is a questionnaire completed by mothers. Each of these measures asks about similar symptoms: suicidal thoughts, sad mood, guilty feelings, social withdrawal, etc., but they rely on different perspectives (mother-report versus self-report). Test the hypothesis that children of depressed mothers will be more depressed than non-depressed mothers using first the CDI, and then the CBCL. To answer this question, you will conduct two different t-tests, one with CDI as the dependent variable, and one with CBCL as the dependent variable. For each of these analyses, maternal depression (momdx1) will be the grouping variable.
4. The researcher expects that the more depressed children are, the more likely they are to have problems in school. In particular, the researcher expects that childhood depression will be related to academic problems. Determine the correlation between childrens depression (again using the CDI) and childrens grade point average.
5. The researcher is primarily interested in maternal depression as a risk factor for childhood depression. However, she appreciates that most (if not all) clinical syndromes are multiply determined. She reflects on her clinical work with depressed children, and she recalls that many of them seem to have poor social problem-solving skills and that they often experience an episode of depression after experiencing rejection or victimization by peers. She wants to examine the relative influences of social problem solving and maternal depression on childrens risk for depression. Use the CDI as your dependent variable, and conduct a regression analysis to determine the relative influences of problem solving (social) and maternal depression (momdx1).
P.S. Most everything (for our purposes) that you need to know from STAT 402
1. Statistics are tools, which have two main uses: (a) to describe data and (b) to make inferences about whether the pattern of results did or did not arise by chance.
2. An important distinction among types of variables is between those that are:
categorical - no meaningful ordering of different values. The do not have an intrinsic quantitative meaning (e.g., sex, race, handedness, hometown, DSM diagnosis)continuous - a meaningful ordering of different values with an intrinsic quantitative meaning (e.g., height, weight, population of hometown, severity of depressive symptoms, income, age, GPA)
3. To describe categorical data, you want to count the numbers (or report percentages) of individuals who fall into each category (e.g., how many boys and how many girls are in your sample).
4. To describe continuous data, you want to calculate a measure of central tendency (e.g., mean, median, mode) and a measure of variability (variance or standard deviation, range, min/max).
5. You can also combine these approaches to description when data are both categorical and continuous for example, saying that the 53 men in your sample on average are 26.50 years old (SD = 4.22) and that the 47 women on average are 29.33 years old (SD = 5.06). In general, when reporting data, it is usually sufficient to use only two decimal points.
6. To decide whether a given pattern of results did or did not arise by chance, you calculate an inferential statistic. The statistical procedure that you choose depends on the nature of the variables (categorical or continuous) and the question posed. In all cases, the statistic you calculate has a numerical value. When you use computer programs like SPSS/PC to calculate inferential statistics, the level of significance or p-value is automatically provided in the output. By convention, the null hypothesis is rejected if the significance level is .05 or smaller.
In writing about the results of inferential tests, by convention you report the specific value of the inferential statistic and the specific significance level. Computer programs may report the significance level of an inferential statistic as .000, but this is a misleading result of rounding-off. A significance level cannot literally be zero or less than zero, so when you report this, bump it up to .001.
7. Degrees of freedom essentially correspond to your sample size minus the number of groups or contrasts. The bigger the sample size, the more degrees of freedom; the more degrees of freedom, the smaller the value of the inferential statistic needs to be to attain a given level of significance. This is why large sample sizes are more powerful than small sample sizes.
8. Here are some of the commonly-used inferential statistics and the sorts of data and questions for which they are appropriate. (Computer programs sometimes calculate these statistics in different ways, and so the exact values may differ slightly. Unless you have a reason not to do so, rely on the first of the alternative statistics reported in the program's output; this statistic typically reflects the conventional formula.)
chi-square statistic (X2) helps to determine whether two categorical variables associated. For example, are depression diagnoses comorbid with anxiety disorders?How to report: X2 (df) = x.xx, p < .05 (or whatever)
t-test helps determine whether two groups (defined by some categorical variable) differ in mean scores on some continuous variables (e.g., are children of depressed mothers more likely to report depressive symptoms on the CDI questionnaire?)
How to report: t(df) = x.xx, p < .001 (or whatever)
analysis of variance (ANOVA) - do more than two groups defined by some categorical variable differ in mean scores with respect to some continuous variable (e.g., are there life expectancy differences among people from different states?)
How to report: F(df,df) = x.xx, p =.30 (or whatever)
correlation coefficients help determine whether two continuous variables are associated (e.g., are childrens depressive symptoms associated with their grade-point averages?)
How to report: r(df) = .xx, p=.01 (or whatever)
linear regression analyses answer questions about the relative influences of two or more continuous or categorical variables on a continuous independent variable (for our purposes: what are the relative effects of maternal depression and social problem solving on childrens depression scores?). When reporting on regression results, you usually report on the standardized coeffiecients (or beta weights) for each of your predictors.
How to report (for each predictor): B = .xx, p < .05 (or whatever)
Research design (8-10 pages, 20 % final grade)
This assignment challenges you to synthesize and apply much of the information that we have covered during the term. This assignment also encourages you to hone your psychology writing skills, so we encourage you to refer to the APA guidelines for the effective expression of ideas, on reserve for this course.
Refer back to Assignments 1 and 3 and select one of your hypotheses. Now, its your turn to design a study to test the hypothesis that you select. REMEMBER: There is no one design or measurement strategy that is best. Whats most important is that you make thoughtful decisions about your research design (participants, setting, measures, etc.) and that you carefully consider potential confounds and ways to control for these.
I. Introduction:
II. Participants:
III. Procedure:
IV. Measures: