Deck 14: Visualizing Causal Stories: Path Analysis

Question 1

On the left-hand side of a path model is the variable education, with no arrows pointing to it. Within the context of path analysis, we would call this variable:

A) an exogenous variable
B) an endogenous variable
C) an independent variable
D) a dependent variable

Answer 1

A

Question 2

On the right-hand side of a path model is the variable education, with no arrows pointing away from it. Within the context of path analysis, we would call this variable:

A) an exogenous variable
B) an endogenous variable
C) an independent variable
D) a dependent variable

Answer 2

B

Question 3

Within a path model, the variable education has an arrow pointing toward it, but also an arrow pointing away from it. Within the context of path analysis, we would call this variable:

A) an exogenous variable
B) an endogenous variable
C) an independent variable
D) a dependent variable

Answer 3

B

Question 4

Calculating an indirect effect requires:

A) multiplying the two direct effects together
B) adding the two direct effects together
C) subtracting one direct effect from the other
D) dividing one direct effect into the other

Question 5

In a path model, the beta on an arrow going from education to income is +.35, the beta on an arrow going from income to political views is +.40, and the beta on an arrow going from education to political views is -.20. All of the betas are statistically significant. The total effect of education on political views is:

A) +.55
B) +.95
C) -.11
D) -.03

Question 6

In a path model, the beta on an arrow going from age to religious service attendance is +.30, the beta on an arrow going from religious service attendance to happiness is +.20, and the beta on an arrow going from age to happiness is +.50. All of the betas are statistically significant. The total effect of age on happiness is:

A) +1.00
B) +0.03
C) 0
D) +.56

Question 7

Using GSS2008 data, we investigate the relationship between sex (coded Male=0, Female=1) and income. We get the following arrows and betas for a path model (all betas are statistically significant):
Beta for arrow from sex to income: -.17
Beta for arrow from sex to hours worked: -.26
Beta for arrow from hours worked to income: +.26
What percentage of the total effect of sex on income is indirect?

A) 17%
B) 24%
C) 29%
D) 71%

Question 8

Using GSS2006 data, we investigate the relationship between education and concern over global warming (concern is measured in a scale from 0=not concerned, up to 15=great concern). We think that support for science (measured on a scale from 0=no support, up to 9=full support) could play a mediating role in this. We get the following arrows and betas for a path model (all betas are statistically significant):
Beta for arrow from education to science support: +.25
Beta for arrow from education to global warming concern: +.10
Beta for arrow from science support to global warming concern: +.15
What percentage of the total effect is direct?

A) 10%
B) 20%
C) 37%
D) 71%

Question 9

Which of these variables cannot be endogenous?

A) age
B) political views
C) income
D) education

Question 10

Path analysis has some similarities to:

A) nested modeling
B) interaction effects
C) logistic regression
D) crosstabulation

Question 11

We are interested in how education affects the number of times one visits the public library in the past year. We think that the more educated you are, the more you use the library. However, we think that income also plays a role: the more educated you are, the more income you have, and the more income you have, the less you use the library (because you simply buy the books in which you are interested). We create a path model using GSS2008 data:
Beta on arrow going from education to income: +.39***
Beta on arrow going from education to library use: +.22***
Beta on arrow going from income to library use: -.14***
Here is a set of nested models created using GSS2008 data. The coefficients are betas:

$\text { Dependent Variable: \# of times visited public library in past year }$
$\begin{array}{llc}\text { Independent Variable } & \text { Model 1 } & \text { Model 2 } \\\text { Education (in years) } & \# \# \# * * * & .22 * * * \\\text { Income (in thousands) } & --- & -.14^{* * *}\end{array}$ What number goes in the place of the ###?

A) .17
B) .13
C) .22
D) .40

Question 12

In a path model, the beta on an arrow going from number of children to relaxation time is a statistically significant -.30. However, the total effect of number of children on relaxation time is zero. There is also an arrow going from number of children to hours worked, and then an arrow going from hours worked to relaxation time. Which of the following could be the betas for these effects (both statistically significant)?

A) -.50 and +.60
B) +.50 and -.60
C) -.50 and -.60
D) -.10 and -.20

Question 13

The punchline of the BMI path analysis example in the textbook was that BMI…

A) had a direct effect on happiness among women, but an indirect effect among men
B) had both direct and indirect effects on happiness for both men and women
C) had both direct and indirect effects on happiness for women, but only a direct effect for men
D) had both direct and indirect effects on happiness for men, but only a direct effect for women

Question 14

In the textbook example regarding the relationship between father's education and son's occupational prestige:

A) there were both direct and indirect effects
B) there was only a direct effect
C) there was only one indirect effect
D) there were multiple indirect effects

Question 15

A common measure used in examining the social-class relationships between generations is:

A) status attainment
B) occupational prestige
C) status structure
D) occupational level

Question 16

In their article on infertility treatment, which racial/ethnic group did Greil and his co-authors NOT study?

A) whites
B) blacks
C) Asians
D) Hispanics

Question 17

According to their survey results, Greil and his co-authors found that, among women who had experienced fertility problems, white women were around ______ as likely to have had fertility tests compared to other women in the study.

A) half
B) twice
C) four times
D) ten times

Question 18

One of Greil's main findings regarding racial-ethnic differences in fertility treatment was that a(n) _________ explanation is too simplistic.

A) economic
B) attitudinal
C) social
D) interpersonal

Question 19

According to Greil's research on fertility treatments, which racial-ethnic group had particularly large indirect effects?

A) whites
B) blacks
C) Asians
D) Hispanics

Question 20

In her research on the causes of distress, Song focuses on which type of capital?

A) financial
B) social
C) cultural
D) human

Question 21

In her research on distress, Song focused not only on the respondent's number of contacts, but also on the ________ of these contacts.

A) level of distress
B) availability
C) gender
D) occupational prestige

Question 22

In her research on distress, Song found that age had:

A) a direct effect on distress, but not an indirect effect
B) an indirect effect on distress, but not a direct effect
C) both a direct and indirect effect
D) neither a direct nor indirect effect

Question 23

In her research on distress, Song found that education had:

A) a direct effect on distress, but not an indirect effect
B) an indirect effect on distress, but not a direct effect
C) both a direct and indirect effect
D) neither a direct nor indirect effect

Question 24

With regard to building a path model using SPSS, the textbook is particularly concerned about:

A) using the betas rather than the unstandardized effects
B) using the same cases in all of the necessary regression models
C) using only paths with statistically significant effects
D) using regular regression and not logistic regression

Question 25

A proposed path model has five variables, two of which are exogenous. How many regression models must you run in order to complete the model?

A) 1
B) 2
C) 3
D) 5

Question 26

Explain what path analysis gets you that regular regression does not.

Question 27

Explain why path analysis uses betas instead of unstandardized slopes.

Question 28

Here are arrows and betas for a simple path model using GSS2008 data (using only respondents who have 12 years of education or less). All betas are statistically significant. The variables are:
Sex: coded Male=0, Female=1
Attend: # of times respondent attends religious services per month
Abindex: 0=no support for abortion, up to 7=full support for abortion
Beta on arrow from sex to abindex: -.10
Beta on arrow from attend to abindex: -.35
Beta on arrow from sex to attend: +.14
By talking about the direct, indirect, and total effects, tell the story of this model.

Question 29

A claim: "We know that age affects political views: older people tend to be more conservative. However, income plays a role in this: older people tend to have higher incomes, and those with higher incomes tend to be more conservative." Here are betas and arrows for a simple path model using GSS2006 data (only white men aged 60 and below). Political views is measured from 0=extremely liberal to 6=extremely conservative. All betas are statistically significant. Use this path model to comment on the claim.
Beta on arrow from age to political views: +.09
Beta on arrow from income to political views: +.16
Beta on arrow from age to income: +.35

Question 30

Earlier in the textbook, there was a nested model showing that, even after controlling for hours worked, there was still an income difference between men and women, although the difference was now smaller. Reframe this nested model as a path diagram, and describe what it is saying.

Question 31

There appears to be absolutely no relationship between Variable A and Variable B. But then we build a path model with these two variables and a variable that connects them: Variable C. There is a statistically significant beta of +.30 going from Variable A to Variable B. The statistically significant beta from Variable A to Variable C is +.60. What is the beta on the arrow going from Variable C to Variable B?

Question 32

Here are the arrows and betas for a simple path model using GSS2006 data (all of them are statistically significant):
Beta on arrow from education to prestige: +.44
Beta on arrow from education to income: +.35
Beta on arrow from income to prestige: +.25
From this information, build a set of two nested models.

Question 33

Someone claims: "Education affects occupational prestige only through income: the more education one has, the more income one will make, and it is income that determines occupational prestige." Here are the arrows and betas for a simple path model using GSS2008 data (all of them are statistically significant):
Beta on arrow from education to prestige: +.43
Beta on arrow from education to income: +.39
Beta on arrow from income to prestige: +.22
By talking about the direct, indirect, and total effects, respond to the claim.

Question 34

Someone claims that black women and Hispanic women use infertility services less frequently than white women do because they lack economic resources, and infertility services are very expensive. How do Greil and his co-authors refute such a claim?

Question 35

What relationship does Song find between age and psychological distress? Describe both the direct and indirect effects.