Part A:
READ THE FOLLOWING ARTICLE from the Detroit Free Press of several years ago and, then, answer the questions below.
An exhaustive audit of a fraud-tainted cancer study has confirmed the original results: Breast cancer is as effectively treated with a breast-saving procedure as with surgery in which the whole breast is removed.
A report Tuesday by experts at the National Cancer Institute said that removing the tumor and then treating with radiation is as likely to result in long-term survival as is mastectomy--removal of the entire breast.
The breast-saving procedure, called lumpectomy, has been in use for some time, but its validity was questioned earlier this year when NCI officials announced that a major study that supported lumpectomy was based, in part, on fraud.
Officials learned that a researcher at St. Luc's Hospital in Montreal had falsified data for some of the patients that he had enrolled in the study.
NCI removed the St. Luc's data from the study last spring and did a cursory recomputation that found the fraud did not affect the conclusion about lumpectomy.
The agency also undertook a detailed audit in which researchers reexamined hospital data for 1,554 patients in the studies and were able to find the original charts to verify data for 1,390 of the enrollees.
The audit found that breast cancer patients who chose lumpectomy and radiation had a 10-year survival rate of 71 percent versus a survival rate of 66 percent for those who chose mastectomy. The results, from a statistical viewpoint, show no significant difference between the two therapies, experts said.
An independent study by the General Accounting Office came up with similar results.
Questions:
1. What is the dependent variable in the study?
2. What is the independent variable in the study?
3. Which group(s) are the experimental group(s) in the study?
4. Which group is the control group?
5. Construct the crosstabulation that results from this study--the crosstabulation that lets you test whether the independent variable has an effect on the dependent variable.
Part B.
Let's say that you are working for a community mental health agency. You
develop a support group for drug addicts; the support group meets weekly. The
aim of the
support group is to reduce drug use among these drug users.
After the support group has been in existence for one year, the director of the
community mental health agency decides to evaluate the success of your program.
He creates a questionnaire and asks the recovering drug addicts: "How often do
you use drugs
now?"
1 Several
times a day
2 Several
times a week
3 Several
times a month
4 Several
times a year
5 Never
List three things that are wrong with this program evaluation (evaluation research)
design. Focus on the whole design—not just the questionnaire.
1.
2.
3.