Correlations and Comparisons

Solution Preview

---

See the attached files.

Individual Assignment: Analyze and Report Correlational and Comparative Data Sets

? Analyze the following two data sets:

o In Statistics Alive, Module 35, p. 410-411, data are presented for 15 persons regarding high school and college GPA.

o In Statistics Alive, Module 20, p. 227-228, data are presented for customer assistance and retail store purchases.

I've shown you how to do these assignments (see below).

? Prepare a report on the two different types of analyses: correlation and comparison, using the questions associated with the two data sets. Address the following for each analysis:

o Answers to each module, such as the summary of statistical test, the results of the statistical analysis, the degrees of freedom, p value, one-tailed or two-tailed
o Meaningfulness of study outcome, such as significance of the statistical finding
o Conclusions that would follow from these results

See the answers to the modules for things like the summary of the tests and the results, etc.

A correlation analysis looks at the relationship between two variables and determines if they have a linear relationship (if you graph them, they fall in a line).

A comparison looks at two different groups and determines if they are significantly different from one another.

Module 20, p. 227-228

PRACTICE

1. Assume that our research hypothesis for the above study is nondirectional. That is, while
we believe that the type of treatment given will affect the level of depression, we have
no idea which treatment-medication or counseling-will be more effective. Look up the
critical t for this two-tailed hypothesis at both the .01 and .05 levels. At which level
(if either) can we reject the null hypothesis?

The null hypothesis is that the two treatments are equally effective. The alternative hypothesis is that they are not equally effective.

If the observed t-value (the one we find while doing the statistical test) is higher than the critical t-value for the level we want (either 0.01 or 0.05), we can reject the null hypothesis and assume that the treatments have different levels of effectiveness.

If the observed t-value is less than or equal to our observed t-value, then we cannot reject the null hypothesis.

Critical values of t depend on the sample size. Find them at one of these ...