You are interested in the outcomes of the children in your workload in general functioning and school performance, and whether they are related. As a result you decide to collect some school information from the children's parents and compare it to their GAF.
1a. What is the measurement scale of these variables?
1b. What measure (exactly) you would you likely use to measure the relation between these variables?
1c. Why? Describe what this procedure measures.
2. Use Excel (or SPSS) to run the statistic you decided on in the previous question. Print the spreadsheet or output and staple it to the back of this HW (make sure you type your name in a cell or in the output before printing). Report the result and interpret exactly what this means in the above scenario.
3. As the counseling supervisor, you are concerned that you have had several sexual assault victims who have not reported good outcomes from therapy in your clinic. You decide to conduct a study comparing their outcomes using their Beck Depression Inventory scores (BDI scores range from 0 to 63-most severe) both before and after 4 months of treatment.
Your results are:
Before After
44 50
51 47
55 44
56 51
45 27
52 34
27 13
48 25
54 31
26 19
3a. Calculate the mean, median, variance, and standard deviation of both groups.
3b. What type of statistical test do you think would be most appropriate for this data, and why?
4.Conduct this test using Excel or SPSS. Print the spreadsheet or output (respectively) and staple it to the back of this HW (make sure you type your name in a cell or in the output before printing).
Interpret this result in detail, as it applies to the above scenario.
5.Calculate Cohen's d (by hand-show your work). What does this measure tell us? Interpret the value.
6.You are now quite curious whether your sample was a bit more depressed than the national average of women in this population at the beginning of the study, which may have led to problems in the generalizability of your experiment. You have heard that the expected BDI within 1 month of such an assault is approximately 42, and would like to know if your sample differed significantly from this average. What statistical procedure would be most appropriate for such an analysis?
6a. Conduct that statistical procedure by hand, taking care to show all work and important values along the way.
6b. What is the relevant tcrit (alpha of .05 -from the t-table, df = N-1)?
6c. Compare your tobt to this value, and state the result of your experiment, and what it means.
7. You find that there is a negative relationship between outcome after a sexual assault (again measured as Beck Depression Inventory score) and the number of days that passed before the assault was reported. You also believe that several other factors play relate to the outcome of the client, such as the gender of the counselor, age of the client, and whether the perpetrator is punished. What statistical technique could you use to try to predict a client's outcome using these variables? Write the appropriate equation (without numbers).
8. You want to know if your male and female counselors differ in how effective they are with clients. You decide to compare the two groups on several variables, including (1) how clients rate their overall experience, (2) how much the client reports feeling valued by the counselor, (3) likelihood of recommending the center to friends or relatives with problems. What type of technique could you use to do this analysis?
9.You are interested in ways to deter domestic violence among prior abusers, and contact three professionals from three different agencies who use very different approaches to the abusers, and request permission to conduct follow-up on the individuals. One focuses on punitive methods, a second on group therapy, and the last on individual psychotherapy. You then record the number of violent relapses or arrests that have occurred during the last year among each of the three groups.
Punitive Group Individual
6 1 1
4 0 3
5 2 0
3 1 1
2 3 0
9a.What statistical procedure would be most appropriate for these data? What would your null hypothesis be? Your alternative hypothesis?
9b.If you were to run the statistical test mentioned above, and found it to be significant, what would your next step be, and why?
9c.Should you include measures of effect size for these steps as well? If so, describe a measure that you could use for each step, and how they could be interpreted.
10. Describe three possible correlations (at least one positive and at least one negative) that you can think of that probably exist in real life that we have not discussed in class or the text. You are welcome to use any research you hear about (in the news, etc.). For each correlation describe whether it is most likely positive or negative, and, for one of the three, mention at least one possible 3rd variable. (These should be unique for each student)