Question

# Regression Analysis 1. At the end of the Regression Analysis with Categorical Data lecture, there was...

Regression Analysis 1. At the end of the Regression Analysis with Categorical Data lecture, there was a prompt about a multiple regression analysis conducted to examine the factors influencing police arrests. There are two competing theories of when the police make arrests: Situational Threats: police only make arrests when protestors use violent or illegal tactics. When demonstrators step out of line, the police respond accordingly. Non-Behavioral Threats: while the tactics protestors use are certainly important, the police are more aggressive with some types of demonstrations due to factors unrelated to police behavior, such as protestor ideology or event size. • The situational threats approach emphasizes protestor tactics, so we use the following measures to test for it: 1. Protestor Violence (l=yes, O=no) 2. Property Damage (l=yes, O=no) 3. Civil Disobedience (1 =yes, 0= no) • The non-behavioral threats approach points to non-behavioral factors. In this case, we use the following measures: 1. Protest Size (in thousands) 2. Government Target (l=yes, O=no) 3. Radical Protestors (1 =yes, O=no) • For

Regression Analysis

1. At the end of the Regression Analysis with Categorical Data lecture, there was a prompt
about a multiple regression analysis conducted to examine the factors influencing police arrests.
There are two competing theories of when the police make arrests:

Situational Threats: police only make arrests when protestors use violent or illegal tactics.

When demonstrators step out of line, the police respond accordingly.

Non-Behavioral Threats: while the tactics protestors use are certainly important, the police
are more aggressive with some types of demonstrations due to factors unrelated to police
behavior, such as protestor ideology or event size.

• The situational threats approach emphasizes protestor tactics, so we use the following
measures to test for it:
1. Protestor Violence (l=yes, O=no)
2. Property Damage (l=yes, O=no)
3. Civil Disobedience (1 =yes, 0= no)
• The non-behavioral threats approach points to non-behavioral factors. In this case, we
use the following measures:

1. Protest Size (in thousands)
2. Government Target (l=yes, O=no)
3. Radical Protestors (1 =yes, O=no)
• For data, we examine 11,436 protest events taking place in 20 US cities between 1996

and 2006.

Here are the results:

 Estimate Std. Error t value P-Value Intercept 0.05 1.05 0.05 0.96 Protestor Violence 9.67 1.21 7.97 < 1.82e - 15 Property Damage 18.43 1.58 11.65 < 2e - 16 Civil Disobedience 12.72 0.78 16.24 < 2e - 16 Protest Size (1000s) 0.03 0.01 3.57 0.000355 Government Target -0.24 0.97 -0.25 0.80 Radical Protestors 5.66 0.6 9.43 < 2e.- 16 Value OF P-Value F(obtained) 116.8 (11430, 6) < 2e - 16 R2 0.07259

The mean for each independent variable is:

Protestor Violence                   0.07
Property Damage                    0.09
Civil Disobedience                  0.12
Protest Size (1000s)                21.58
Government Target                 0.41

brief -you do not need to write a lot to answer these questions.

1. Which theory is supported? Why? (5 points)

1. Does this regression model significantly improve our ability to predicty relative to
just guessing the mean? Justify your answer by drawing on the appropriate statistic(s)
in the table. (5 points)

1. Interpret the value of R2 from the table. (5 points)

1. Pick one partial slope and state the null and research hypotheses for it. (5 points)

1. What type of protest does the intercept refer to? (5 points)

1. What is the expected number of arrests for an event with violence, civil disobedience,
and radical protestors (other variables held constant at their means)? (10 points)

1. What is the expected number of arrests for an event without violence, property
damage, or civil disobedience and with radical protestors and a government target
(other variables held constant at their means)? (10 points)

1. Complete the table to indicate whether each of the partial slopes is statistically significant
at 95% confidence. (5 points)

i)

Variable                             Significant?

Intercept                                 Yes

Protestor Violence
Property Damage
Civil Disobedience
Protest Size (100s)
Government Target

If you Satisfy with Answer, Please give me "Thumb Up". It was very important to me.

#### Earn Coins

Coins can be redeemed for fabulous gifts.