Scribd
Upload
95 views4 pages

Interpret The Beta. Note That Ethwhite Is Left Out As Reference, To Used To Compare With Asian and Black

This document contains questions and responses regarding regression analysis and hypothesis testing. Key points include: 1. A regression model is estimated to predict crime rates based on ethnicity, with Asian associated with 0.25 higher crime rates than white. 2. An F-test is conducted to test the joint significance of regression coefficients, failing to reject the null hypothesis of no relationship. 3. A Wald test examines the equality of two coefficients, failing to reject the null of equality, indicating there is a "triple effect". 4. Alternative models are considered by removing variables, choosing the better fitting model without the PTIME86 variable based on R-squared values. 5. Graphs of residuals are examined

Uploaded by

viettuan91
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
95 views4 pages

Interpret The Beta. Note That Ethwhite Is Left Out As Reference, To Used To Compare With Asian and Black

This document contains questions and responses regarding regression analysis and hypothesis testing. Key points include: 1. A regression model is estimated to predict crime rates based on ethnicity, with Asian associated with 0.25 higher crime rates than white. 2. An F-test is conducted to test the joint significance of regression coefficients, failing to reject the null hypothesis of no relationship. 3. A Wald test examines the equality of two coefficients, failing to reject the null of equality, indicating there is a "triple effect". 4. Alternative models are considered by removing variables, choosing the better fitting model without the PTIME86 variable based on R-squared values. 5. Graphs of residuals are examined

Uploaded by

viettuan91
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Q1:

a. Estimate equation.
Interpret the Beta.
Note that Ethwhite is left out as reference, to used to compare with Asian and
Black.
Estimate equation : write the name of every X. Then Substitute Beta.
.Explain the meaning of each Beta
Ethasian: = 1, then 0.25 units of crime to be Asian more than to be white
Ethasia = 0, then no chance for Asian to commit crime. It can only be black
and white.
E.g: The regression coe cient, b = 0.2744 means that if Xt increases in 1
unit, Yt will increase in 0.2744 units, i.e., for each extra hour of study the mark
increases in 0.2744 points.
Meaning of estimated intercept.
C. If other factors = 0, then the number of times a man could be arrested can
be 53.25%. Y becomes a constant model. This mean C is the minimum value
of Y, also is the starting point of the graph on the Y- axis, if other factors = 0.
Q2:
f- test: k = coefficient +1 = 7.
N= 1495
R2= in table
Tra bang F- test, k-1= 6 (ngang), doc. : infinity. = 2.10
Q2a:
R- squared= 0.0088148 (no rounding).
Thay ACT = sth not significant

Q3:
3B3 (PTIME86)= B4 (QEMP86)

Wald test: 3C4 = C5


F-Statistic= 2.520724, within AR= [0; 3.84]. Accept H0
H0 : 3B3 = B4: triple effect
H1 : 3B3 # B4: not have triple effect
P- value: 0.1126 > 0.05 (alpha), accept H0
Conclude: has triple effects.
Generate series Z.
Re- estimate the model, sub z: narr86 c avgsen tottime z ethasian ethblack
(show results)
Steps Calculate F*
RSSr = 1134.619
RSSu = 1132.7 (in table q1)
Ku = 7
Kr = 6
Thay so, tinh theo CThuc F*
3c: same. Conclude triple result

Q4:
Same hypothesis Ho. H1
Viet lai AR, thay so =[-1.96; 1.96]
Same conclusion.
4b (i): same. Sub ALCOHOL = PTIME86.

Estimate model again, without PTIME86.

narr86 c avgsen tottimeqemp ethasian

ethblack
r-quared=

0.072824

adj r-squared=
0.069711

Estimate model again, without QEMP86.

narr86 c avgsen tottime ptime86 ethasian

ethblack
r-quared=

0.050633

adj r-squared=
0.047445

Conclude: choose the model without PTIME86.


(ii): same as sample.
Q5: same. To plot the graph, Quick -> Graph-> AVGSEN RESID01
Do the same for the others. Then save the graph.
-

First plot, scatter: with homoscedasticity: AVGSEN, TOTTIME

- Second plot: copy the steps


Conclude: same as sample

Q6:
Auxiliary regression:
^i=a0 +a1 AVGSEN i + a2 TOTTIM Ei +a3 PTIME 86i +a 4 QEMP 86i +a5 ETHASIAN i +a6 ETHBLACK i +a 7 AVG
U

Note: The same as sample ETHASIAN2, ETHBLACK2 not included


H0: a0=a1=.=a25=0 -> Homoskedasticity
H1: at least one of the coefficients not equal to 0
y A.R = [0,2

( ,1)

] = [0,2(0.05 ,251) ] = [0,36.4150]

Run White tests on Eviews


R-squared =
0.025161

LM=nR2=37.615695

Conclusion: not in the range, reject H0


b) Same as sample
Q7:
The same as sample, substitute ACT -> QEMP86
Q8:
Run eviews based on instructions
Same conclusion as sample
Q9:
Same
Q10:
Using DW tables for n=1495 and k = 6(excluding intercept)
Dl = 1.90842
Du = 1.92181
DW statistic =
1.969219

it is in the range of no autocorrelation

You might also like