Unformatted text preview: STA 3024 Homework 2 This form is only for people whose ﬁrst name starts with any letter from ”p” to”t” .
Suppose we want to investigate a potential association between gender and political aﬃliation.The table below shows the data from a random sample of potential voters.
Gender
Females
Males Poltical party/Gender
Democrats
Independents
780
992
551
878 Republicans
600
554 1. For our sample data, calculate the conditional proportions of a voter being in a given party
knowing its gender. (Your answer should be a 2 × 3 table of conditional percentages.)
2. Now let’s use this data to conduct a chisquared test with α = 0.1.
(a) Calculate each cell’s expected count for the chisquared test. (Your answer should
be a 2 × 3 table of expected counts.)
(b) State the assumptions made by the chisquared test. For each assumption, state
whether or not it is satisﬁed by our data.
(c) State the hypotheses for the chisquared test.
(d) Calculate the value of the test statistic X 2 .
(e) In the following step, we will compare this observed X 2 value to a chisquared distribution. What df should this chisquared distribution have?
(f) Find the pvalue as best you can (i.e., “between 0.025 and 0.05”) using a chisquared
table or any other method of your choice.
(g) Use this pvalue to make a decision, and interpret this decision in the context of the
actual variables. (Your interpretation should be understandable by someone who
doesn’t know anything about this data or about statistics in general.)
(h) Would you have needed to go through the chisquare test of independence procedure
if the table was based on all possible voters ?Explain. ...
View
Full Document
 Spring '08
 TA
 Statistics, ChiSquare Test, Normal Distribution, Chisquare distribution, Pearson's chisquare test, Fisher's method

Click to edit the document details