Question 1

Use the following description and table to answer the next few questions.
A state education department conducted a study of 1000 primary schools to find out how much parental engagement matters for improving academic performance .They measured parental engagement at each school (on a 1-5 scale), and used regression analysis to examine the effect of engagement on the percent of students meeting state standards in English language arts .To get a better estimate of the true causal effect of engagement, they controlled for poverty and school size (enrollment) because of a concern that these variables may be influencing both parental involvement and academic performance .The results of various specifications using regression analysis are shown below.
Regression analysis of percent of students meeting ELA standards (various specifications)
 $$\begin{array} { | l | c | c | c | c | } 
\hline & \text { Spec 1 } & \text { Spec 2 } & \text { Spec 3 } & \text { Spec } 4 \
\hline \begin{array} { l } 
\text { Parental engagernent (1-5 } \
\text { scale) }
\end{array} & 13.1 8  ^ { * } & 8 . 2 3^ * & 14.18 ^ { * } & 9.14 ^ { * } \
\hline \text { Poverty (\%/ of students) } & & - 0 . 4 9 ^{*} & & - .47 * \
\hline \text { Enrollment (\# of students) } & & & .013^*& .009^* \
\hline
\end{array}$$ 
NOTE: N=1000 primary schools; dependent variable is percentage of students meeting standards in English language arts (ELA); * p < .05
-How did controlling for poverty change alone the estimated causal effect of parental engagement?

Accepted Answer

A)  Reduced the effect
B)  Increased the effect
C)  Eliminated the effect
D)  Did not change the effect  A

Question 2

Violent behaviors among high school students, such as carrying a weapon to school, are difficult to predict .Studies examining the determinants of these behaviors thus have weak explanatory power .The statistic that represents this idea is called

Accepted Answer

A)  Constant
B)  Slope
C)  P value
D)  R square D

Question 3

A researcher may choose to utilize multiple regression over stratification since:

Accepted Answer

A)  It is possible to add several control variables at the same time
B)  Control variables do not play into the equation
C)  It permits up to three control variables but no intervening variables
D)  None of the above A

Question 4

A shortcoming in the use of stratification would be:

Accepted Answer

A)  Stratifying data by more than two or three control variables makes the results difficult to interpret
B)  It may produce imprecise results
C)  Stratifying generates little data from numerous subgroups
D)  All of the above

Question 5

Alcoholics Anonymous (AA), a voluntary sobriety program, compared AA members to similar non-members based on age, gender, marital status, and number of drinks per week, producing a predicted probability of joining AA .What type of design is this?

Accepted Answer

A)  Control variables implemented through multiple regression
B)  Qualitative research
C)  Lab experiment
D)  Propensity score matching

Question 6

Use the following description and table to answer the next few questions.
A state education department conducted a study of 1000 primary schools to find out how much parental engagement matters for improving academic performance .They measured parental engagement at each school (on a 1-5 scale), and used regression analysis to examine the effect of engagement on the percent of students meeting state standards in English language arts .To get a better estimate of the true causal effect of engagement, they controlled for poverty and school size (enrollment) because of a concern that these variables may be influencing both parental involvement and academic performance .The results of various specifications using regression analysis are shown below.
Regression analysis of percent of students meeting ELA standards (various specifications)
 $$\begin{array} { | l | c | c | c | c | } 
\hline & \text { Spec 1 } & \text { Spec 2 } & \text { Spec 3 } & \text { Spec } 4 \
\hline \begin{array} { l } 
\text { Parental engagernent (1-5 } \
\text { scale) }
\end{array} & 13.1 8  ^ { * } & 8 . 2 3^ * & 14.18 ^ { * } & 9.14 ^ { * } \
\hline \text { Poverty (\%/ of students) } & & - 0 . 4 9 ^{*} & & - .47 * \
\hline \text { Enrollment (\# of students) } & & & .013^*& .009^* \
\hline
\end{array}$$ 
NOTE: N=1000 primary schools; dependent variable is percentage of students meeting standards in English language arts (ELA); * p < .05
-Poverty and school size are presumed common causes that, if omitted from the analysis, result in a biased estimate of the causal effect of parental engagement on the percent of students meeting ELA standards .If this is true, then the best estimate of the true causal effect of parental engagement available in the table is

Accepted Answer

A)  13.18
B)  8.23
C)  14.23
D)  9.14

Question 7

Use the following description and table to answer the next few questions.
A state education department conducted a study of 1000 primary schools to find out how much parental engagement matters for improving academic performance .They measured parental engagement at each school (on a 1-5 scale), and used regression analysis to examine the effect of engagement on the percent of students meeting state standards in English language arts .To get a better estimate of the true causal effect of engagement, they controlled for poverty and school size (enrollment) because of a concern that these variables may be influencing both parental involvement and academic performance .The results of various specifications using regression analysis are shown below.
Regression analysis of percent of students meeting ELA standards (various specifications)
 $$\begin{array} { | l | c | c | c | c | } 
\hline & \text { Spec 1 } & \text { Spec 2 } & \text { Spec 3 } & \text { Spec } 4 \
\hline \begin{array} { l } 
\text { Parental engagernent (1-5 } \
\text { scale) }
\end{array} & 13.1 8  ^ { * } & 8 . 2 3^ * & 14.18 ^ { * } & 9.14 ^ { * } \
\hline \text { Poverty (\%/ of students) } & & - 0 . 4 9 ^{*} & & - .47 * \
\hline \text { Enrollment (\# of students) } & & & .013^*& .009^* \
\hline
\end{array}$$ 
NOTE: N=1000 primary schools; dependent variable is percentage of students meeting standards in English language arts (ELA); * p < .05
-How did controlling for school size alone change the estimated causal effect of parental engagement?

Accepted Answer

A)  Reduced the effect
B)  Increased the effect
C)  Eliminated the effect
D)  Did not change the effect

Question 8

An educational researcher is interested in matching student scores along several variables across schools in a particular district .She will most likely want to use

Accepted Answer

A)  Corresponding score aptitude
B)  Propensity score matching
C)  Tendency score matching
D)  Relative score matching

Question 9

Use the following description and table to answer the next few questions.
A state education department conducted a study of 1000 primary schools to find out how much parental engagement matters for improving academic performance .They measured parental engagement at each school (on a 1-5 scale), and used regression analysis to examine the effect of engagement on the percent of students meeting state standards in English language arts .To get a better estimate of the true causal effect of engagement, they controlled for poverty and school size (enrollment) because of a concern that these variables may be influencing both parental involvement and academic performance .The results of various specifications using regression analysis are shown below.
Regression analysis of percent of students meeting ELA standards (various specifications)
 $$\begin{array} { | l | c | c | c | c | } 
\hline & \text { Spec 1 } & \text { Spec 2 } & \text { Spec 3 } & \text { Spec } 4 \
\hline \begin{array} { l } 
\text { Parental engagernent (1-5 } \
\text { scale) }
\end{array} & 13.1 8  ^ { * } & 8 . 2 3^ * & 14.18 ^ { * } & 9.14 ^ { * } \
\hline \text { Poverty (\%/ of students) } & & - 0 . 4 9 ^{*} & & - .47 * \
\hline \text { Enrollment (\# of students) } & & & .013^*& .009^* \
\hline
\end{array}$$ 
NOTE: N=1000 primary schools; dependent variable is percentage of students meeting standards in English language arts (ELA); * p < .05
-The independent variable of interest in this study is ...

Accepted Answer

A)  Percent meeting ELA standards
B)  Parental engagement
C)  Poverty
D)  Enrollment

Exam 13: Multiple Regression to Disentangle Causal Effects

Violent behaviors among high school students, such as carrying a weapon to school, are difficult to predict .Studies examining the determinants of these behaviors thus have weak explanatory power .The statistic that represents this idea is called

A researcher may choose to utilize multiple regression over stratification since:

A shortcoming in the use of stratification would be:

Alcoholics Anonymous (AA) , a voluntary sobriety program, compared AA members to similar non-members based on age, gender, marital status, and number of drinks per week, producing a predicted probability of joining AA .What type of design is this?

An educational researcher is interested in matching student scores along several variables across schools in a particular district .She will most likely want to use