As a research analyst for the Public Policy program at the University of Maryland, you are asked to deter-

mine whether electronics should be banned from the classroom. Proponents of the rule change argue that

electronics are distracting. Opponents of the rule change, on the other hand, argue that electronics are useful

for taking notes, for example, and therefore should be permitted. Using the following code, generate (simu-

lated/fake) data to help in answering the question. The code will generate data for two separate variables:

test scores and hours of electronics use in the classroom.

The following code should be copied directly into R (For some reason copying and pasting the code from

this PDF doesn't always work well. So I have posted the code on the exam assignment page for your conve-

nience.). The only thing you should change is the seed value. You must use your birthdate as the seed value.

For example, if my birthday was on April 22, 1984, I would use the number 04221984 inside the parenthesis

in set.seed() as such: set.seed(04221984).

set

. s e e d ( 0 4 2 2 1 9 8 4 )

t e s t

s c o r e=

rnorm

( 1 5 0 ,

mean

=90 ,

sd

=10)

e l e c t r o n i c s

h o u r s=

rnorm

( 1 5 0 ,

mean

=20 ,

sd

=5)

summary

(

lm

( t e s t

s c o r e

e l e c t r o n i c s

h o u r s ) )

Part A (5 pts)

Using the output from the regression in R, make table to show the results.

Part B (5 pts)

Explain the meaning of statistical and substantive significance. Comment on both the

statistical and substantive significance of your results (from Part A).

Part C (5 pts)

Is the model a good fit? How can you tell?

Part D (5 pts)

Interpret any other information that you think is relevant in the table and have not discussed so far.

Part E (5 pts)

Based on the results you obtained, what is your conclusion? Should you or should you not ban electronics

from the classroom? In answering this question, make sure that you use proper statistical terminology and

as much information from the table that you created as necessary to completely answer the question.

2

Part F (5 pts)

Why might the model specification not be accurate? Is there a variable(s) that we could add to the regression

model that might help us obtain a better or more reliable conclusion?

Part G (5 pts)

Is a linear specification appropriate here? To answer this question, I do not need you to run any tests. All

I want you to do is tell me what the assumptions for a linear model are and how (answer this using your

words and your intuition - no need to run any code) you would go about testing these assumptions.