### HINTS FOR REVIEW PROBLEMS FOR FINAL EXAM    M358K, Sp 06  Smith

#3. For Step 2, it may help to review the sampling distribution of p-one-hat.

#5
For part (b), yoiu should be able to rephrase the question in the first paragraph in the form, "Are the distributions of variable Y the same in populations 1 and 2?" You need to figure out what variable Y is and what the two populations are.

For part (c), remember that when you have a two-way table with just two categories in each categorical variable, the chi-squared test is equivalent to a certain test for proportions. In other words, you should be able to rephrase the question in the first paragraph in the form, "Is the proportion of population A that has property X the same as the proportion of  population B that has property X?" You need to figure out what A, B, and X are.

#6: Exactly one of parts d, f, j, and n is each of the following:
Two-sample t-test
Matched pair one sample t-test
One sample t-test (not matched pairs)
None of the above -- there is a confounding factor that makes the study design inappropriate.

Hint for parts b, c, g, and i: When is the chi-squared approximation good ehough to use? (See p. 597, paragraph below the box.)

#13: Try using the formula for slope.

#14: Do you have an SRS or something that might reasonbly be considered one?

#15(b): What about the shape of a distribution can tell you which numerical summary is more appropriate?