Deck 7: Tests of Significance and Measures of Association

Question 1

A researcher studying the ratings of men and women toward the Democratic Party hypothesizes that women will give Democrats a higher rating than will men. The female mean rating is 52.5 with a standard error of 1.2 points. The male mean rating is 47.1 with a standard error of 1.6 points. Should the researcher reject the null hypothesis?

A) No, because the lower boundary for women overlaps the upper boundary for men
B) Yes, because the lower boundary for women overlaps the upper boundary for men
C) No, because the boundaries do not overlap
D) Yes, because the boundaries overlap

Answer 1

A

Question 2

Measures of association allow a researcher to determine ______.

A) how much of the sample lies above the upper boundary
B) how much of the sample lies below the lower boundary
C) the strength of the relationship between the variables
D) the significance of the relationship between the variables

Answer 2

C

Question 3

A PRE measure of association tells a researcher ______.

A) the significance of the variables
B) whether to accept the null hypothesis
C) whether to reject the null hypothesis
D) how much the odds of predicting the value of the dependent variable are improved

Answer 3

D

Question 4

Which of the following tests is based on the confidence interval approach?

A) regression
B) $\chi$ ²
C) eyeball
D) two-tailed

Question 5

Somers' d_yx is used to measure the strength of the association between two ______-level variables.

A) nominal
B) interval
C) ordinal
D) ratio

Question 6

In statistical notation $\chi$ ² refers to ______.

A) chi-square
B) the square of the independent variable
C) the square of the dependent variable
D) the square of the intervening variable

Question 7

Concordant pairs are cell values consistent with a(n) ______ relationship.

A) negative
B) neutral
C) inverse
D) positive

Question 8

A pair of cell values with a negative relationship are referred to as a(n) ______ pair.

A) discordant
B) concordant
C) tied
D) invalid

Question 9

The symbol H₀ represents the ______.

A) null hypothesis
B) alternative hypothesis
C) standard deviation
D) confidence interval

Question 10

A researcher studying partisanship and the likelihood of being a protester concludes there is no difference between Democrats and Republicans when, in fact, there is, has committed a ______.

A) faux pas
B) Type I error
C) Type II error
D) Type III error

Question 11

A ______ helps researchers decide whether an observed relationship between an independent variable and a dependent variable really exists in the population or whether it could have happened by chance when the sample was drawn.

A) measure of association
B) test of statistical significance
C) hypothetical claim
D) level of measurement

Question 12

The $\chi$ ² test of significance is used when both the independent variable and the dependent variable are measured ______.

A) at the nominal level
B) categorically
C) at the ordinal level
D) at the interval level

Question 13

A relationship between which of the following pairs of variables could be tested using $\chi$ ²?

A) income and age in years
B) weight and the number of cigarettes smoked
C) political party and frequency of church attendance
D) race and income (in dollars)

Question 14

Which of the following pair of variables would allow us to compare two sample means?

A) income (in dollars) and age (in years)
B) education level and partisanship
C) gender and partisanship
D) gender and income (in dollars)

Question 15

The ______ allows the researcher to determine the exact probability of a particular result occurring by chance.

A) test of significance
B) p value approach
C) standard error approach
D) confidence interval approach

Question 16

A researcher who concludes that a relationship exists between X and Y when it does not has committed a ______.

A) faux pas
B) Type I error
C) Type II error
D) Type III error

Question 17

A PRE measure of 1 indicates what?

A) no association
B) perfect association
C) a 100% improvement in the odds of predicting the dependent variable
D) no improvement in the odds of predicting the dependent variable

Question 18

A ______ marks the upper plausible boundary of random error and so defines H₀'s limit.

A) critical value
B) discordant pair
C) test statistics
D) confidence interval

Question 19

The ______ is stated to remind us that the observed results of a sample could be the result of random error.

A) standard error
B) probability
C) null hypothesis
D) skeptic's position

Question 20

When comparing two sample means, we can safely reject the null hypothesis if ______.

A) the confidence intervals overlap
B) the confidence intervals are higher than .05
C) the confidence intervals are lower than .05
D) the confidence intervals do not overlap

Question 21

When researchers compare two sample statistics, however, we have two sources of uncertainty: both sample statistics have random sampling error.

Question 22

Chi-square works by comparing the observed frequencies with the expected frequencies.

Question 23

The p value for a test statistic is based on the expected sampling distribution and can be found on a probability distribution table.

Question 24

The eyeball test of statistical significance is based on the confidence interval approach.

Question 25

Most political research begins by using the .10 level of significance to decide whether to reject a null hypothesis.

Question 26

An asymmetric measure of association does not take into consideration which variable is being used to predict the other.

Question 27

A sample statistics tells us exactly how many standard errors separate the sample difference from zero.

Question 28

A symmetric measure of association treats the independent variable as the causal variable and the dependent variable as the effect.

Question 29

To determine the degrees of freedom when using the $\chi$ ² test, we subtract the number of rows from the number of columns.

Question 30

A researcher who has two categorical level variables with one measured at the nominal level should use Somers' d_yx to test the association between them.

Question 31

The $\chi$ ² test returns a critical value that allows a researcher to determine the upper plausible boundary of random error that defines the limit of the null hypothesis.

Question 32

The assumption that there is no difference between the attitudes of males and females on abortion rights is called a null hypothesis.

Question 33

Cramer's V is a good PRE to use when comparing two interval-level variables.

Question 34

Lambda is designed to measure the strength of a relationship between two categorical variables, at least one of which is a nominal-level relationship.

Question 35

The $\chi$ ² test is the oldest statistical test still in use and takes into account tabular data.