Question 1

-Refer to the table. What is E1?

Accepted Answer

A)  10 
B)  50 
C)  55 
D)  70

Question 2

Which element of research design influences tests of significance far more than it influences measures of association?

Accepted Answer

A)  sample size 
B)  level of measurement 
C)  how the EPSEM sample was obtained 
D)  whether the association is positive or negative

Question 3

How would a researcher reveal if variables are associated when the variables are arranged in a bivariate table?

Accepted Answer

A)  by subtracting their scores horizontally 
B)  by computing percentages in the direction of the dependent variable 
C)  by computing percentages in the direction of the independent variable 
D)  by adding their scores vertically

Question 4

In a 2 × 2 table, what is the range of phi?

Accepted Answer

A)  1 to 100 
B)  0 to 1.00 
C)  0 to 1.00. 
D)  -1.00 to +1.00

Question 5

Suppose the maximum difference calculated in a bivariate table is 23 percentage points. How should we describe the relationship?

Accepted Answer

A)  as weak 
B)  as moderate 
C)  as strong 
D)  as causal

Question 6

Suppose there is a perfect relationship between being convicted of a crime (X) and being sentenced to jail (Y) in a sample of criminal defendants. Which of the following statements can we conclude?

Accepted Answer

A)  Just as many defendants not convicted of a crime were sentenced to jail as were those who were convicted. 
B)  All cases in each column of a bivariate table would be located in a single cell. 
C)  There is maximum variation in jail sentencing for each category of crime conviction. 
D)  We are unable to predict jail sentencing on the basis of crime conviction.

Question 7

How is phi similar to Cramer's V?

Accepted Answer

A)  Both first require calculating ÷2. 
B)  Both are measures of proportional reduction in error (PR
E) . 
C)  Both provide a meaningful percentage of the explained variance in Y. 
D)  Both are useful for 4 × 4 bivariate tables.

Question 8

-Refer to the table. Which of these statements would best describe the results if we drew a paralleled pie chart to depict the association between workplace type (unionized or non-unionized) and job satisfaction?

Accepted Answer

A)  Two pies would be drawn, one for satisfied workers and one for unsatisfied workers; the pie for unsatisfied workers would be larger. 
B)  Two pies would be drawn, one for satisfied workers and one for unsatisfied workers; both would have identical pie slice sizes. 
C)  Two pies would be drawn, one for unionized workers and one for non-unionized workers; they would have identical pie slice sizes. 
D)  Four pies would be drawn, the biggest slices of which would correspond to non-unionized workers.

Question 9

How do we find E2?

Accepted Answer

A)  by subtracting the respective column totals from row totals, and dividing by n 
B)  by multiplying E1 by n 
C)  by subtracting the column totals from the largest cell frequency in each column, and then adding subtotals together 
D)  by subtracting the largest cell frequency in each column from the column total, and then adding the subtotals

Question 10

Suppose chi square computed on a bivariate table is greater than zero. Which of these statements describes the variables?

Accepted Answer

A)  They are not associated, and they are independent. 
B)  They are associated, and they may be dependent. 
C)  They are associated, and they are independent. 
D)  They are not associated, and they are not independent.

Question 11

A researcher assesses the relationship between countries' development (highly developed versus less highly developed) and whether or not the country has ever elected a female head of state. The lambda for this association is 0.28. The researcher is better off knowing about country development by what percentage?

Accepted Answer

A)  72% 
B)  52% 
C)  28% 
D)  8%

Question 12

Suppose we wish to understand the relationship between two variables: social class and film genre preference. Which of the following would be a reason to prefer lambda as a measure of association?

Accepted Answer

A)  Lambda allows us to understand whether our sample was drawn randomly from a well-defined population. 
B)  Lambda gives us clear criteria for knowing whether the association is weak, moderate, or strong. 
C)  Lambda is a symmetrical measure of association, so we will get the same value even if we were to reverse the X and Y variables in our calculation. 
D)  Lambda allows us to quantify exactly how much knowing a person's social class helps to predict their film preference.

Question 13

Lambda is considered an asymmetric measure of association. What does this mean?

Accepted Answer

A)  Lambda can help reject a hypothesis, but it does not directly provide information about strength of association. 
B)  Lambda provides information about positive relationships, but not negative associations. 
C)  Lambda provides information about negative associations, but not positive relationships. 
D)  The value of lambda may vary depending on which variable is considered independent and which is considered dependent.

Question 14

In what sense do measures of association help with prediction?

Accepted Answer

A)  They show whether a relationship is causal. 
B)  They give us a better estimate of a case's score on the dependent variable than we would obtain by guessing. 
C)  They tell us whether an observed relationship is likely to have occurred by random chance. 
D)  They allow us to determine whether a theory is true or false.

Question 15

Which of the following types of bivariate tables are appropriate for calculating phi?

Accepted Answer

A)  2 × 1 
B)  2 × 2 
C)  4 × 0 
D)  Bivariate tables of any size are appropriate for calculating phi.

Question 16

In his thesis project, a student analyzes a 2 × 3 table with level of education (university degree versus less than university degree) predicting frequency of strenuous exercise ( weekly). He finds that lambda = 0.48. What should his interpretation be?

Accepted Answer

A)  Knowing how frequently a person exercises helps reduce our error in predicting education level by about half. 
B)  Knowing a person's education helps reduce our error in predicting exercise frequency by about half. 
C)  We can assume that there is a moderate relationship, but lambda lacks a PRE interpretation. 
D)  It is inappropriate to calculate lambda for a 2 × 3 table because it is not symmetrical. He should change the table to be 2 × 2 or 3 × 3 before computing lambda.

Question 17

-Refer to the table. What is lambda?

Accepted Answer

A)  0.82 
B)  0.45 
C)  0.33 
D)  0.25

Question 18

Researchers are comparing Canadians and Americans on support (yes versus no) for subsidized child care. A phi of 0.21 was calculated for the relationship between nationality and support for the proposed childcare policy. What can we conclude?

Accepted Answer

A)  Canadians are more supportive than Americans for subsidized child care. 
B)  There is a positive association between nationality and support for the policy. 
C)  There is a moderate association between the variables. 
D)  There is a strong association between the variables.

Question 19

Tests of significance reveal if results are non-random. What should a researcher use to investigate the strength and direction of relationships?

Accepted Answer

A)  ANOVA 
B)  percentages 
C)  measures of association 
D)  chi square tests

Question 20

Which one of the following patterns describes a positive association between two variables?

Accepted Answer

A)  As one variable increases in value, the other also increases. 
B)  As one variable increases in value, the other stays the same. 
C)  As one variable decreases in value, the other increases. 
D)  Neither variable decreases.

Exam 8: Measures of Association for Variables Measured at the Nominal Level

-Refer to the table. What is E₁?