Clinical Trial for the Efficacy of Estrogen Patches in the Treatment of Post-Natal Depression

Women who had suffered an episode of post-natal depression were randomly allocated to two groups; the members of one group received an estrogen patch, and the members of the other group received a dummy patch-c--the placebo. The dependent variable was a composite measure of depression, which was recorded on two occasions prior to randomization and six occasions after treatment. Lower scores indicate less depression. The data in the SPSS file consists of three derived depression scores for each women, Pre-treatment mean score, Post-treatment mean score, and Change (Pre Treatment-Post Treatment). These data can be analyzed in many ways (see Frison and Pocock, 1992, Statistics in Medicine, 11: 1685-1704). The complete research design forms a

Randomized pretest-posttest control group design

O: Observation

X: Treatment

R: Randomized to Treatment

Group Estrogen: Pre-Test Randomize Estrogen Treatment Post Test Group Placebo: Pre-Test Randomize Placebo Treatment Post Test

Note, the treatment is fully under control of the researcher and the patients are randomly assigned to the two treatment conditions. The dependent variable is measured twice during the study (before and after the treatment). That is, the patients were randomly assigned to either an Estrogen Treatment or the No Treatment Placebo Control condition, and following the treatment period depression was again measured (posttest). The independent variable is the drug treatment (with two levels: Estrogen Treatment and Placebo Control). Further the time of testing can be considered a repeated measures or within-subject's "treatment" (with two levels: pre- and posttest). Another way in which to depict a design like is like this follows:

Again, O: Observation; Treatment (Estrogen); Control (Placebo).

The SPSS file is set up this way, of course omitting the subscripts. In addition, a column of change scores is provided in the file.

This is a prototypical research design and is considered a "good" design because it controls for many alternative explanations. Again, patients are randomly assigned to either a treatment or control condition and the dependent variable is measured in both groups at baseline (i.e. prior to treatment for the experimental group) and post-treatment. The strength of the design lies in the fact that we can make strong conclusions because we can evaluate or test a number of statistical hypotheses and rule out alternative interpretations.

Estrogen and Depression Data Analysis Test Questions

- Caution, because there are multiple data sets in the SPSS file, only analyze one dependent variable at a time. Make sure to check sample size to ensure that you are analyzing the data as intended.

The Test has as its focus the "Big Question": Did the Estrogen Treatment "Work"?

The focal question, Did the Estrogen Treatment "Work"? This design allows us to answer this question from a number of perspectives.

1. Are the two Treatment groups different at posttest? \(\left(\mathrm{O}_{2}\right.\) versus \(\left.\mathrm{O}_{4}\right)\);
2. Ignoring the Placebo Condition, Did the Estrogen Treatment change depression scores \(\left(\mathrm{O}_{7}\right.\) versus \(\left.\mathrm{O}_{2}\right)\) ?

Part I (30 -35 points)

1. Did the Estrogen Treatment "Work"? State the statistical hypotheses and conduct the appropriate statistical analysis.

1. State the statistical hypotheses and specify your alpha level.
2. List and evaluate the assumptions of the statistical test.
3. Conduct an independent -sample t-test on the post-test depression scores and specify the statistical decision.
4. Interpret the \(95 \%\) confidence interval of the statistic that is the focus of this statistical hypothesis.
5. Calculate an estimate of effect size for this study. What did we learn from this statistic?
6. Why is an independent sample t-test appropriate?
7. Is this a true experiment (explain why or why not)?

2. Did the Estrogen Treatment "Work"? State the statistical hypotheses and conduct the appropriate statistical analysis.

1. State the statistical hypotheses and specify your alpha level.
2. List and evaluate the assumptions of the statistical test.
3. For the Estrogen Treatment condition only, conduct a paired / dependent-sample t-test on the change from pre-treatment to posttreatment depression scores and specify the statistical decision.
4. Interpret the \(95 \%\) confidence interval of the statistic that is the focus of this statistical hypothesis.
5. Calculate an estimate of effect size for this study. What did we learn from this statistic?
6. Why is a paired sample t-test appropriate?

Part II. Write a "stand alone" results section that summarizes the statistical analyses and that addresses the researcher's question, Did the Estrogen Treatment "Work"? Make sure to include a figure (e.g., error bar plot(s)) as part of the results section. Make sure to report and interpret the effect size for each analysis. (20 points)

Price: $24.32

Solution: The downloadable solution consists of 12 pages, 1232 words and 15 charts.
Deliverable: Word Document