Sociology 592, Exam 2, November 8, 2002 – Page 1
Sociology 592 - Research Statistics I
Exam 2 Answer Key [ROUGH DRAFT]
November 8, 2002
(10 points each, 30 points total.)
You have been asked to serve as a statistical consultant for several proposed projects.
For each of the following, your employers want you to tell them:
Which of the cases we have studied their problem falls under (e.g. one sample tests, case I,
nonparametric tests, case II, tests of association).
Briefly explain why.
the null and alternative hypotheses
whether a Z, T, chi-square, or F test is appropriate; where applicable, also tell what the degrees of freedom
for the test are.
You DO NOT have to give the formula for the test statistic, nor do you need to specify the
If values for population parameters are not specified (e.g.
) assume they are unknown; and if two or more unknown
's are involved, assume they are equal.
The Federal Reserve Board is uncertain what effect, if any, the recent elections have had on the public’s confidence in
A week before the election, 80 Americans were asked to rate their confidence in the economy, on a scale ranging
from a low of 0 to a high of 30.
A week after the election, another random sample of 80 Americans will be asked to do the same
Two Sample Tests, Case II,
are unknown but assumed to be equal (or else
One Way Anova will work.)
Use either T test with df = 158 or F test with DF = 1, 158
President Bush feels he has a mandate for his tax policies.
His advisors, however, are cautioning him that support for
his policies is not universal.
In particular, they warn that middle class voters are much less favorable to his tax policies than are
upper class voters.
One hundred middle class voters and 100 upper class voters will be asked whether they approve or
disapprove of Bush’s tax policies.
Two Sample Tests, Cave V, Difference between two proportions.
Obesity is a growing health concern. A medical researcher knows that race and social class are related to each other,
i.e. whites have higher incomes than do nonwhites.
She also believes that obesity may be related to race and/or social class.
Sixty individuals will be selected at random.