Deck 7: Give Me One Good Reason Why: Bivariate Correlation and Regression

Question 1

The most common form of simple regression uses:

A) two ratio-level variables
B) two ordinal-level variables
C) a ratio-level variable and a nominal-level variable
D) a ratio-level variable and an ordinal-level variable

Answer 1

A

Question 2

If b=3, the mean of X is 5, and the mean of Y is 10, what is the value of the constant?

A) -5
B) 0
C) 5
D) 10

Answer 2

A
Explanation:10 = a + 3(5). Solving for a, we get -5

Question 3

If the slope is 4, the mean of X is 7, and the constant is 3, what is the mean of Y?

A) -15
B) 0
C) 5
D) 31

Answer 3

D
Explanation:mean of Y = 3 + 4(7) = 31.

Question 4

When we want to visually examine the relationship between two ratio-level variables, the most common type of graph to use is:

A) the pie graph
B) the bar graph
C) the line graph
D) the scatterplot

Question 5

If the value of the covariance is negative, which of the following cannot be true?

A) the slope is negative
B) the correlation coefficient is negative
C) the slope is exactly zero
D) the constant is exactly zero

Question 6

With simple regression, if we want to know the size of the effect, the best statistic to look at is:

A) the slope
B) the correlation coefficient
C) r-squared
D) the constant

Question 7

We run a simple regression using an independent variable X and a dependent variable Y. We find that there is absolutely no relationship between the two variables. What will the regression line look like?

A) There can be no regression line in this situation.
B) It will be a completely vertical line at the mean of X.
C) It will be a completely horizontal line at the mean of Y.
D) It will take on a U-shape.

Question 8

With simple regression, the best-fitting line is the one that:

A) touches the most points
B) has the least squared distances from the points
C) has the least squared vertical distances from the points
D) has the least squared horizontal distances from the points

Question 9

Regular regression is also called OLS, or:

A) ordinary least squares regression
B) only linear situations regression
C) obvious latent summary regression
D) original longitudinal survey regression

Question 10

With simple regression, if we want to know the strength of the relationship, the best statistic to look at is:

A) the slope
B) the correlation coefficient
C) r-squared
D) the constant

Question 11

A researcher looks at two variables: the number of cats people own, and the number of dogs people own. She finds that the correlation coefficient is -.84. The best interpretation of this value is:

A) Few people with cats have dogs.
B) Few people with dogs have cats.
C) As the number of dogs people own goes up, the number of cats people tend to own goes down.
D) The relationship between cats and dogs is negative.

Question 12

With regard to positive signs and negative signs, which of the following is most accurate?

A) The correlation coefficient and the slope are always the same sign.
B) The correlation coefficient and the slope are never the same sign.
C) The correlation coefficient and the slope are usually the same sign.
D) The correlation coefficient and the slope are almost always the same sign.

Question 13

A correlation coefficient of -.40 is ________ a correlation coefficient of +.40.

A) not as strong as
B) .80 less strong as
C) just as strong as
D) twice as strong as

Question 14

With simple regression, if we want to know the proportion of variation in the dependent variable that is explained by its relationship with the independent variable, the best statistic to look at is:

A) the slope
B) the correlation coefficient
C) r-squared
D) the constant

Question 15

In the movie gross forecasting example in the textbook, the removal of the James Cameron films from the sample:

A) made the relationship weaker
B) made the relationship stronger
C) made the relationship perfect
D) had no effect on the strength of the relationship

Question 16

The textbook suggests that if you're using a large number of cases and you end up with a small r-squared value of, say, 0.02, you should:

A) use a smaller number of cases
B) start over with different variables
C) check your math
D) be unphased, as this can happen

Question 17

In the textbook example about health and work, which had the weakest correlation?

A) men's missed work with men's physical health
B) men's missed work with men's emotional health
C) women's missed work with women's physical health
D) women's missed work with women's emotional health

Question 18

In their article on U.S. politics, Healy and Lenz examine the correlation between:

A) campaign spending and vote share
B) negative advertisements and vote share
C) candidate age and vote share
D) economic growth and vote share

Question 19

In Healy and Lenz's article on U.S. politics, they find __________ between economic growth and vote share.

A) no correlation
B) a weak correlation
C) a moderately strong correlation
D) a perfect correlation

Question 20

Jaeger uses which dataset in his study of physical attractiveness?

A) The National Longitudinal Study of Adolescent to Adult Health
B) The Wisconsin Longitudinal Survey
C) The Panel Study of Health Dynamics
D) The General Social Survey

Question 21

In presenting his findings, Jaeger uses what technique?

A) a scatterplot
B) a regression equation
C) a correlation matrix
D) a crosstabulation

Question 22

With regard to BMI and physical attractiveness, which of the following best describes what Jaeger found?

A) Men's BMI was positively correlated with physical attractiveness
B) Men's BMI was negatively correlated with physical attractiveness
C) Women's BMI was positively correlated with physical attractiveness
D) Women's BMI was negatively correlated with physical attractiveness

Question 23

When you run a scatterplot in SPSS, you put the dependent variable on the:

A) x-axis
B) y-axis
C) z-axis
D) h-axis

Question 24

The textbook suggests that if you want to run a scatterplot in SPSS:

A) use a small number of cases
B) use a large number of cases
C) you should run a chi-square test as well
D) you should run a t-test as well

Question 25

Both crosstabs and simple regression can be used to look at relationships between two variables. Explain how you would choose which one of these to use.

Question 26

In the 2008GSS, we have the years of education and hours of television watched for five Native Hawaiians:
Using years of education as the independent variable and hours of television watched as the dependent variable, calculate the regression equation and correlation coefficient. Interpret your results.

Question 27

Among Buddhists in the GSS2008, does education affect how often one visits an art museum? Here are the regression results:
TIMES VISIT ART MUSEUM IN PAST YEAR = -3.94 + .31 (YRS. EDUC)
r = .62, r² = .38 n = 10
Interpret these results, and make predictions for a Buddhist who has 16 years of education, and a Buddhist who has 22 years of education.

Question 28

Explain why regular regression is sometimes called Ordinary Least Squares regression.

Question 29

Someone shows you some regression results they just developed. They have a negative slope and a positive correlation coefficient. Explain to them why they must have done something wrong.

Question 30

To assess the relationship between number of absences and grades, an instructor collects data from the 10 students in a class and runs a simple regression. She finds a perfect negative relationship: as number of absences goes up, grades go down. But then, the instructor remembers that there is an 11^th person in the class: someone who never came to class but still managed somehow to get a perfect grade in the course. Conceptually (not mathematically) describe what will happen to the slope and correlation coefficient.

Question 31

Using the relationship between education and income, explain what an outlier is, and why it is likely to affect regression results more dramatically in a small sample than a larger sample.

Question 32

Using only the American Indians in the GSS2008, we run regressions to see if education affects income, and we do so for men and women. The r² for men is .28 and the r² for women is .03. Interpret this difference.

Question 33

Explain why there is a diagonal of "1.00"s in every correlation matrix.

Question 34

Explain how the health and work example in the textbook uses regression to dispel an important gender stereotype.