CH07 Linear Regression
CH07 Linear Regression
CH 7 - Linear Regression
1. _____ is a statistical procedure used to develop an equation showing how two variables are related.
a. Regression analysis b. Data mining
c. Time series analysis d. Factor analysis
2. A regression analysis involving one independent variable and one dependent variable is referred to as a
a. factor analysis. b. time series analysis.
c. simple linear d. data mining.
regression.
3. The population parameters that describe the y-intercept and slope of the line relating y and x, respectively, are
a. .
b. y and x.
c. a and b.
d. a and B.
5. In the simple linear regression model, the ____________ accounts for the variability in the dependent variable that
cannot be explained by the linear relationship between the variables.
a. constant term b. error term
c. model parameter d. residual
6. In a linear regression model, the variable that is being predicted or explained is known as _____________. It is denoted
by y and is often referred to as the response variable.
a. dependent variable
b. independent variable
c. residual variable
d. regression variable
8. In the graph of the simple linear regression equation, the parameter represents the ___________ of the regression
line.
a. slope b. x-
intercept
c. y- d. end-point
intercept
Copyright Cengage Learning. Powered by Cognero. Page 1
CH 7 - Linear Regression
9. In the graph of the simple linear regression equation, the parameter is the ___________ of the regression line.
a. slope b. x-
intercept
c. y- d. end-point
intercept
10. In a linear regression model, the variable (or variables) used for predicting or explaining values of the response
variable are known as the ________________. It(they) is(are) denoted by x.
a. dependent variable b. independent variable
c. residual variable d. regression variable
11. In a simple linear regression analysis the quantity that gives the amount by which the dependent variable changes for a
unit change in the independent variable is called the
a. coefficient of determination. b. slope of the regression line.
c. correlation coefficient. d. standard error.
12. A ___________ is used to visualize sample data graphically and to draw preliminary conclusions about the possible
relationship between the variables.
a. contingency table b. scatter
chart
c. Gantt chart d. pie chart
13. A procedure for using sample data to find the estimated regression equation is
a. point estimation. b. interval estimation.
c. the least squares d. extrapolation.
method.
14. When the mean value of the dependent variable is independent of variation in the independent variable, the slope of
the regression line is
a. positive. b. zero.
c. negative. d. infinite.
15. The difference between the observed value of the dependent variable and the value predicted using the estimated
regression equation is known as the
a. constant term. b. error term.
c. residual. d. model parameter.
16. The _______________ is the range of values of the independent variables in the data used to estimate the regression
model.
a. confidence interval b. codomain
c. experimental region d. validation set
Copyright Cengage Learning. Powered by Cognero. Page 2
CH 7 - Linear Regression
17. Prediction of the value of the dependent variable outside the experimental region is called
a. interpolation. b. forecasting.
c. averaging. d. extrapolation.
18. Prediction of the mean value of the dependent variable y for values of the independent variables x1, x2, . . . , xq that are
outside the experimental range is called
a. dummy variable. b. overfitting.
c. extrapolation. d. interacton.
19. The ____________________ is a measure of the error that results from using the estimated regression equation to
preduct the values of the dependent variable in the sample
a. sum of squares due to regression b. error
(SSR) term
c. sum of squares due to error (SSE) d. residual
20. The least squares regression line minimizes the sum of the
a. differences between actual and predicted y values. b. absolute deviations between actual and predicted y
values.
c. absolute deviations between actual and predicted x d. squared differences between actual and predicted y
values. values
21. What would be the value of the sum of squares due to regression (SSR) if the total sum of squares (SST) is 25.32 and
the sum of squares due to error (SSE) is 6.89?
a. 31.89 b. 19.32
c. 18.43 d. 15.32
22. The ___________ is a measure of the goodness of fit of the estimated regression equation. It can be interpreted as the
proportion of the variability in the dependent variable y that is explained by the estimated regression equation.
a. residual b. coefficient of determination
c. dummy variable d. interaction variable
24. What would be the coefficient of determination if the total sum of squares (SST) is 23.29 and the sum of squares due
to regression (SSR) is 10.03?
a. 2.32 b. 0.43
c. 0.19 d. 0.89
CH 7 - Linear Regression
25. Regression analysis involving one dependent variable and more than one independent variable is known as
a. simple regression. b. linear
regression.
c. multiple regression. d. none of these.
26. The process of making estimates and drawing conclusions about one or more characteristics of a population through
analysis of sample data drawn from the population is known as
a. inductive inference. b. deductive inference.
c. statistical inference. d. Bayesian inference.
27. The process of making conjecture about the value of a population parameter, collecting sample data that can be used to
assess this conjecture, measuring the strength of the evidence against the conjecture that is provided by the sample, and
using these results to draw a conclusion about the conjecture is known as
a. postulation. b. hypothesis testing.
c. statistical inference. d. empirical research.
28. __________________ refers to the use of sample data to calculate a range of values that is believed to include the
value of the population parameter.
a. Interval estimation b. Hypothesis testing
c. Statistical d. Point estimation
inference
30. The scatter chart below displays the residuals verses the dependent variable, x. Which of the following conclusions
can be drawn from the scatter chart given below?
CH 7 - Linear Regression
a. The residuals have a increasing variance as the dependent variable
increases.
b. The model captures the relationship between the variables accurately.
c. The regression model follows the standard normal probability distribution.
d. The residual distribution is consistently scattered about zero.
31. The scatter chart below displays the residuals verses the dependent variable, t. Which of the following conclusions
can be drawn based upon this scatter chart?
a. model is time-invariant.
b. model captures the relationship between the variables accurately.
c. residuals are not independent.
d. residuals are normally distributed.
32. The scatter chart below displays the residuals verses the dependent variable, x. Which of the following conclusions
can be drawn based upon this scatter chart?
CH 7 - Linear Regression
independent variable.
d. The residuals are normally distributed.
33. The scatter chart below displays the residuals verses the dependent variable, x. Which of the following conclusions
can be drawn based upon this scatter chart?
34. The _____________ is an indication of how frequently interval estimates based on samples of the same size taken
from the same population using identical sampling techniques will contain the true value of the parameter we are
estimating.
a. residual b. tolerance factor
c. confidence level d. accuracy level
35. ______________ refers to the degree of correlation among independent variables in a regression model.
a. Multicollinearity b. Tolerance
c. Rank d. Confidence level
36. The degree of correlation among independent variables in a regression model is called
a. multicollinearity. b. interaction.
c. the coefficient of determination. d. the sum of squared errors
(SSE).
37. ________________ is used to test the hypothesis that the values of the regression parameters B0, B1, B2, ... Bq are all
zero.
a. An F test b. A t test
c. The least squares d. Extrapolation
Copyright Cengage Learning. Powered by Cognero. Page 6
CH 7 - Linear Regression
method
38. A variable used to model the effect of categorical independent variables in a regression model which generally takes
only the value zero or one is called
a. a residual. b. the coefficient of determination.
c. a dummy variable. d. interaction.
39. A variable used to model the effect of categorical independent variables in a regression model is known as a
a. dependent variable. b. response.
c. dummy variable. d. predictor variable.
40. Which of the following regression models is used to model a nonlinear relationship between the independent and
dependent variables by including the independent variable and the square of the independent variable in the model?
a. a multiple regression b. quadratic regression model
model
c. a simple regression model d. a least squares regression
model
41. The prespecified value of the independent variable at which its relationship with the dependent variable changes in a
piecewise linear regression model is referred to as the
a. milestone. b. knot.
c. tipping point. d. watchpoint.
42. _____________ refers to the scenario in which the relationship between the dependent variable and one independent
variable is different at different values of a second independent variable.
a. Interaction b. Multicollinearity
c. Autocorrelatio d. Covariance
n
43. Fitting a model too closely to sample data, resulting in a model that does not accurately reflect the population is
termed as
a. approximation b. hypothesizing.
.
c. overfitting. d. postulating.
44. Assessing the regression model on data other than the sample data that was used to generate the model is known as
a. approximation. b. cross-validation.
c. graphical d. postulation.
validation.
45. __________ is the data set used to build the candidate models.
a. Range b. Codomain
c. Validation set d. Training set
Copyright Cengage Learning. Powered by Cognero. Page 7
CH 7 - Linear Regression
46. _____________ refers to the data set used to compare model forecasts and ultimately pick a model for predicting
values of the dependent variable.
a. Codomain b. Training set
c. Validation set d. Range
a.Developascatt
erc hartf
ort heabov
edata.Wha tdoesthi
schar
tindi
cat
eabouttherel
at
ionshipbet
wee nmark
et
capi
tal
izati
onandpr ofit
?
b.Usetheda t
atodev elopane sti
mate
dregr
e s
sionequati
onthatc
ouldbeusedtoesti
mateafirm’spr
ofitba
sedoni
ts
ma r
ketcapit
ali
zat
ion. Whati sthees
tima
tedregress
ionmode l
?
c.Wh atist
hepredictedprofitfort
hemarketcapi
tali
zat
ionof70,
721.3(mil
li
on)?
48. Ares
earc
hc en
teri
sin
terest
edininv
esti
gat
ingabou
thei
ghta
nda g
eofchi
ldr
enwhoa
rebe
twe
en5t
o9y
ear
sol
d.
Inordertodothi
s,asa
mpleo f15chi
ldreni
sselec
tedandt
hedat
aisgiv
enbe
low.
Ag e(
inyears) He i
ght( i
nches
)
7 47.3
8 48.8
5 41.3
8 50.4
8 51
7 47.1
7 46.9
7 48
9 51.2
8 51.2
5 40.3
8 48.9
6 45.2
Copyright Cengage Learning. Powered by Cognero. Page 8
CH 7 - Linear Regression
5 41
.9
8 49
.6
a.Deve
lopascatte
rc ha
rtwithageastheinde
pe ndentvar
iable
.Wh a
tdoesthesca
tt
ercharti
ndi
cat
eaboutther
ela
ti
onshi
p
bet
weentheheightandageo fchi
ldr
en?
b.Usethedat
atode vel
opa nest
imat
edregres
sione qua
ti
ont hatcouldbeusedt
oe s
ti
ma t
ethehei
ghtbase
dontheage.
Whatisthees
ti
ma tedregr
e s
sionmodel?
c.Howmuc hofthev ar
iat
ioninthesampleval
ue sofhei
ghtdoe sthemodelest
imate
dinp ar
t(b
)expla
in?
a.Deve
lopas cat
terchartf
ort
heabovedat
a.Wha tdoest
hischar
tindi
cat
eabou
ttherel
at
ionshi
pbet
weent
he
Nati
onalIncomeandt heCompany'
ssal
esint
hep eri
od2000to2011?
b.Usetheda t
atodevelopane
sti
matedre
gres
sionequati
onthatc
ouldbeusedt
oesti
matethecompa
ny’
ssa
les
base
dont henati
onalincome.
Wha ti
stheest
imatedregr
essi
onmodel?
CH 7 - Linear Regression
Tes
twhethere
achofther
egr
ess
ionpar
amete
rsβ0andβ1i
sequa
ltoz
eroa
ta0
.05l
eve
lofs
igni
fic
anc
e.Wha
tar
ethe
cor
rec
tin
t e
rpr
eta
ti
onsoft
heest
imat
edregr
ess
ionpar
amet
ers
?
51. Theda t
alist
edbel
owi sth
ea ver
ageper
sona
linc
omea
ndp
ers
ona
lcons
ump
tione
xpe
ndi
tur
esb
ase
dont
hes
urv
ey
conduc te
dint heye
ar1995to2009i nU.
S.
Pe rsonalincome Personalconsumption
(
$) expenditur
es($)
23,
310 18,714
24,
444 19,569
25,
657 20,414
27,
260 21,434
28,
336 22,738
30,
317 24,227
31,
162 25,074
31,
448 25,865
32,
282 26,848
33,
872 28,228
35,
423 29,818
37,
723 31,210
39,
418 32,551
40,
156 33,273
39,
113 32,853
a.Devel
opas c
atte
rchar
tfortheaboveda t
a .Wha tdoest
hischarti
nd icatea boutthere l
at
ions hipbe twe ena ver
age
per
sonalincomea ndper
sonalconsump t
ione xpe ndit
ure?
b.Devel
opa nes
timat
edregr
e s
sionequati
ons howi nghowpe r
sonalcons umpt ionexpe nditur
ei sr el
a t
e dpe rs
onal
inc
ome.
c.Whatp ropor
tionofva
riat
ioninthes a
mpl ev a l
uesofproport
ionofp ersona lc
on s
ump t
ione xpe nditur edoe sthi
s
modelexplain?
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE SIMPLE LINEAR REGRESSION MODEL, LEAST SQUARES METHOD,
ASSESSING THE FIT OF THE SIMPLE LINEAR REGRESSION MODEL, Pages 304 and
308
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
52. Theda t
alist
edbel
owi sth
ea ver
ageper
sona
linc
omea
ndp
ers
ona
lcons
ump
tione
xpe
ndi
tur
esb
ase
dont
hes
urv
ey
conduc te
dint heye
ar1995to2009i nU.
S.
Pe rsonalincome Personalconsumption
(
$) expenditur
es($)
23,
310 18,714
24,
444 19,569
25,
657 20,414
27,
260 21,434
28,
336 22,738
30,
317 24,227
31,
162 25,074
Copyright Cengage Learning. Powered by Cognero. Page 10
CH 7 - Linear Regression
31,
448 25,
865
32,
282 26,
848
33,
872 28,
228
35,
423 29,
818
37,
723 31,
210
39,
418 32,
551
40,
156 33,
273
39,
113 32,
853
a
.Wha
tist
he95p
erc
entc
onfide
ncei
nte
rva
lfort
her
egr
ess
ionpa
rame
terβ1?Ba
sedont
hisi
nte
rva
l,wha
t
c
onc
lus
ionc
any
ouma
kea
bou
ttheh
ypo
the
sest
hatt
her
egr
ess
ionpa
rame
terβ1i
seq
ualt
oze
ro?
d
.Wh
ati
sthe95pe
rce
ntc
onfide
ncei
nte
rva
lfo
rther
egr
ess
ionpa
rame
terβ0?Ba
sedont
hisi
nte
rva
l,wha
t
c
onc
lus
ionc
any
ouma
kea
bou
ttheh
ypo
the
sest
hatt
her
egr
ess
ionpa
rame
terβ0i
seq
ualt
oze
ro?
53. Asurve
yisc
onduct
edtode
ter
minewhe
the
rth
eag
eofc
ari
nflue
nce
sthea
nnua
lma
int
ena
ncec
ost
.As
ampl
eof10
carsisse
lect
eda
ndthedat
aiss
hownbel
ow.
AnnualMaint
enanc
e
Ag
eofc
ar(
mont
hs)x
Cost(
$)y
3 120
5 115
6 135
7 290
9 275
10 300
11 350
13 475
14 500
15 550
a.Devel
opas ca
tte
rchartf
orthe
sedatawithageofcar
sa stheindependentvari
able
.Wha tdoest
hes c
att
erchar
t
i
ndicat
eabouttherel
ati
ons
hipbetweenageofacarandtheannua lma i
ntenancecos
t?
b.Usethedatat
ode v
e l
opanest
imate
dr e
gress
ionequa
tionthatcouldbeus edtopredi
ctt
hea nnualmai
ntena
nce
costgi
ventheageoftheca
r.Whatistheest
imate
dregre
s s
ionmode l?
ANSWER:
c
.Thescatt
erchar
twithageofc ar
sast h
ei ndepende
ntvari
ablefol
lows.
CH 7 - Linear Regression
Thissca
tt
erchar
tindi
cat
esther
eisapos
it
iv
eli
nea
rre
lat
ion
shi
pbe
twe
ena
geofac
ar
andtheannua
lma i
nte
nancecost
.
d.Thefol
lowi
ngExc
eloutputpr
ovidesthee
sti
mat
edre
gre
ssi
onequa
ti
ont
hatc
oul
dbe
usedt
oe s
ti
matet
heannua
lma in
tenancebas
edonthea
geofaca
r.
Thee
sti
mat
eds
imp
lel
ine
arr
egr
ess
ione
qua
ti
oni
s .
The estimated simple linear regression equation can also be found by adding a trendline to the
scatter chart as shown on page 304 of the text book.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE SIMPLE LINEAR REGRESSION MODEL, LEAST SQUARES METHOD, Page 304
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
54. As
urv
eyi
sconduc
tedt
ode
ter
minewhe
the
rth
eag
eofc
ari
nflue
nce
sthea
nnua
lma
int
ena
ncec
ost
.As
ampl
eof10c
ars
CH 7 - Linear Regression
i
sse
lect
edandthedat
aiss
hownbelo
w.
AnnualMaint
enanc
e
Ageofcar(months
)x
Cost(
$)y
3 120
5 115
6 135
7 290
9 275
10 300
11 350
13 475
14 500
15 550
a.Tes
twhe t
here
achoft
heregr
e s
sionpara
me t
e rsβ0a ndβ1isequalt
oz eroata0.05lev
elofs
igni
fic
anc
e.
b.Int
erpr
ett
heest
imat
edre
gressi
onpa r
ame t
ers?Ar etheseint
erpr
eta
tionsre
a s
onable
?
ANSWER:
a.Firs
twec heckthec ondit
ionsneces
saryforvali
dinfer
encei
nregr
essi
on.TheEx
cel
plotoftheres
idualsanda geo fca
rsfol
lows.
Be c
ausewea r
eworkingwi
thonly10obse
rvati
ons,assessi
ngthecondi
ti
onsnecessa
ry
forinf
ere
ncetobevali
dinregr
ess
ioni
sextr
e mel
ydifficul
t.However
,th
isscat
terchar
t
doesnotpro
videstr
ongevi
denceofavi
ola
tionofthec ondit
ions
,sowewillproce
e d
withourinf
erenc
e.
Ex
celout
put
:
CH 7 - Linear Regression
The p-value associated with the estimated regression parameter b1 is 4.14E-06, which is
0.00000414 or essentially zero. Because this p-value is less than the 0.05 level of significance,
we reject the hypothesis that β1 = 0. We conclude that there is a linear relationship between
age of a car and the annual maintenance cost. Based upon the regression model, we estimate
is that as age of car increases by one year the maintenance cost increases by $38.34, on
average. The maintenance cost of a car is expected to increase as the age increases. So this
result is consistent with what is expected.
The p-value associated with the estimated regression parameter b0 is 0.229, which is greater
than the 0.05 level of significance, we cannot reject the hypothesis that β0 = 0. The estimated
regression parameter b0 suggests that when the age of car is 0, the maintenance coat is –
$45.60. This result is obviously not realistic and the parameter estimate and the test of the
hypothesis that β0 = 0 are meaningless because the y-intercept has been estimated through
extrapolation (there is no car in the sample data with age zero).
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: INFERENCE AND REGRESSION, Page 322
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
55. Aresearcht
eaminatt
heGonzagaUniv
ers
it
yisi
nte
res
tedi
npred
ict
ingast
udent
'
sove
ralluni
ver
si
tyGPAifhi
s/
her
hi
ghs c
hoo lGPAiskno
wn.Assumethatar
andomsa
mpleof20st
udent
sisse
lec
tedf
romthedatal
is
tedbe
low.
Hi ghSchoolGPA Uni
ver
sityGPA
3.65 3.
72
3.08 3.
21
3.12 3
2.87 2.
67
3.34 3.
57
2.7 2.
97
3.25 2.
83
3.01 2.
89
3.2 3.8
2.62 2.
43
Copyright Cengage Learning. Powered by Cognero. Page 14
CH 7 - Linear Regression
3
.71 3
.82
2
.26 2
.66
3
.02 3
.62
3
.58 3
.83
3
.67 3
.74
2
.92 2
.64
3
.71 3
.53
3
.22 2
.99
2
.27 2
.37
2.8 2
.69
a.Deve
lopas ca
tt
e rchartf
orthesedatawit
hHi ghSchoolGP Aastheinde
pe nde
ntv ari
able.
Whatdoesthescatterchar
tindic
ateaboutt
h erel
at
ions
hipbe t
weenhighschoolGP Asando ver
alluni
ver
sit
yGPA?
b.Deve
lopa nest
ima te
dr e
gress
ionequati
ons howi
ngho whighschoolGPAi sre
latedtooveral
lunive
rsi
tyGPA.
Whatistheest
ima te
dr egr
essi
onmode l?
c.Whatisthepredict
edo ve
ral
lunive
rsit
yGP AofSophi a,ast
udentwhoha sbeena dmitt
edtoGo nza
ga
Univ
ersi
ty,wit
h3 .40highschoolGPA?
ANSWER:
a.Thescatt
erchartwit
hhighschoolGPAa st
hei ndependentvar
iablef
oll
ows.
Thi
ssc
a t
tercharti
ndic
ate
sthe
remaybeapo
sit
iv
eli
nea
rre
lat
ion
shi
pbe
twe
enhi
ghs
chool
GPAsando v
e r
alluni
ver
si
tyGPA.
b
.Thef
ollo
wingExcelout
putpr
ovi
destheest
ima
tedr
egre
ssi
one
q ua
ti
onthatc
ou l
dbeusedt
o
e
sti
mat
etheov
eral
lunive
rsi
tyGPA(y)base
donstude
nt’
sGPAscore
dinhighschool(
x).
CH 7 - Linear Regression
56. Ar
ese
arche
rwa
nte
dtostud
yeffe
ctoft
wofa
ctor
s,x
1andx
2,ony
iel
d()
y.Theob
ser
vat
ionsa
regi
venbe
low.
Obs e
rvati
ons x1 x2 y
1 42.
5 30.
1 260.
4
2 43.
5 29.
3 261.
7
3 43.
9 31.
1 273.
6
4 44.
8 29.
6 278.
6
5 46.
8 29.
7 281.
5
6 47.
5 29.
9 294.
6
7 50.
1 30.
1 301.
2
8 51.
9 30.
4 314.
6
9 54.
7 30.
5 320.
5
10 54.
8 31 324.
7
11 57.
1 31.
8 356.
7
12 57.
8 31.
4 370.
3
13 62.
3 31.
5 378
14 66.
7 32.
1 384.
8
15 71.
2 32.
5 396.
9
a
.De
vel
opa
nes
ti
mat
edl
ine
arr
egr
ess
ione
qua
ti
onwi
tht
hef
act
orx
1ast
hei
nde
pend
entv
ari
abl
e.
Te
stf
oras
ign
ific
antr
ela
ti
ons
hipb
etwe
enf
act
orx
1andy
iel
datt
he0
.05l
eve
lofs
ign
ific
anc
e.
CH 7 - Linear Regression
b.Ho
wmuc hoft
hev
ari
at
ioninthesampl
evaluesofyiel
ddoesth
emodelinpar
t(a)e
xpl
ain?
ANSWER: a.Thefol
lowi
ngEx celout
putpr
ovide
sthee
stima
tedli
nea
rre
gress
ione
qua
ti
ont
hatc
an
beusedtopre
dic
ty i
eld(y)gi
vent
hefact
orx.
1
Theest
imate
dline
arre
gres
sione
qua
ti
onis .
Bef
oretes
ti
ngthehypot
hes
isβ1=0f
ort
hisr
egr
ess
ionmode
l,wec
hec
kthec
ondi
ti
ons
nece
ssa
ryf
orv
ali
din
fer
enc
einr
egr
ess
ion
.Th
eEx
celp
lotoft
her
esi
dua
lsa
ndf
act
orx
1
fo
llo
ws.
Be c ausewea rewor ki
ngwi thonl y15ob servat i
ons ,ass essingthecondi ti
onsne cess
ary
fori nferenc etobev al
idinr egres
sioni sex treme l
ydi ffic ul
t.Ho wever,thesescattercha
rts
dono tpr ovides tron gevi
d enc eofav iolationoft hec ondi ti
ons .
The p-value for the test of the hypothesis that β1 = 0 is 9.64E-10 = 0.000000000964, which
is essentially zero.
Be c auset hisp- valueislesstha nthe0. 05l evelofs igni ficance,wer ej
e ctt
heh ypothesi
sthatβ1
=0 ,a ndc onc lude
tha ttherei sar e
lationshi
pbe t
we enyielda ndf a ctorx1a tthe0. 05levelofsignific
ance.
b
.Thec
oeffic
ien
tofde
ter
mina
tionR2i
s0.
9483,s
othere
gres
sionmodele
sti
mat
edi
npa
rt
(
a)e
xpl
ai
nsa ppr
oxi
matel
y94.
8%o ft
heva
ria
ti
onintheyi
el
dinthesampl
e.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE SIMPLE LINEAR REGRESSION MODEL, LEAST SQUARES METHOD,
ASSESSING THE FIT OF THE SIMPLE LINEAR REGRESSION MODEL, Pages 308, 312
Copyright Cengage Learning. Powered by Cognero. Page 17
CH 7 - Linear Regression
and 316
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
57. Ar
ese
arc
herwa
nte
dtos
tud
yeffe
ctoft
wof
act
ors
,x1a
ndx
2,ony
iel
d()
y.Theob
ser
vat
ionsa
regi
venbe
low.
Obs
erv
ati
ons x1 x2 y
1 42.
5 30.
1 260.
4
2 43.
5 29.
3 261.
7
3 43.
9 31.
1 273.
6
4 44.
8 29.
6 278.
6
5 46.
8 29.
7 281.
5
6 47.
5 29.
9 294.
6
7 50.
1 30.
1 301.
2
8 51.
9 30.
4 314.
6
9 54.
7 30.
5 320.
5
10 54.
8 31 324.
7
11 57.
1 31.
8 356.
7
12 57.
8 31.
4 370.
3
13 62.
3 31.
5 378
14 66.
7 32.
1 384.
8
15 71.
2 32.
5 396.
9
a.Devel
opa nesti
ma t
edr egressi
one q uati
onwithbothf ac
torsx1andx 2astheindepe ndentvari
abl
e s
.Istheove
ral
l
regr
essi
onstat
ist
icall
ysignific
anta tthe0.05levelofsigni
fic
a nce
?I fso,
thent e
stwhe thereachoftheregre
ssi
on
paramet
ersβ0,β1,andβ2ise qua
lt oz eroata0.
01l evelofsi
gnific
a nc
e.Wha tarethec orre
ctint
erpret
ati
onsofthe
esti
matedre
gressi
onpa rame ter
s?
b.Ho wmuchoft hev a
riat
ioni nthes a mpleval
uesofydoe st hemode li
npa rt(a)explain?
ANSWER:
Thef oll
o wingExceloutputprovidestheest
imatedmu ltipleli
nearre
gressi
one q
uati
ontha
t
couldb eus edtopr
edictthey i
el
d( y)giventhetwof ac
tor sx1andx2.
CH 7 - Linear Regression
Thee
sti
mat
edmu
lti
plel
ine
arr
egr
ess
ione
qua
ti
oni
s .
Bef
orete
sti
ngforasigni
ficanto
veral
lre
gre
ssionrelat
ionshi
p(tha
tis
,te
sti
ngtheh
ypothes
is
t
hat
β
1=
β
2=0),wechec
kt heconditi
onsneces
sar
yforv a
lidinf
e r
enceinre
gre
ssi
on.TheExce
lp l
otso
f
t
heresi
dual
sande ac
hoft hetwoindepe
ndentvaria
blesfoll
ow.
Becausewear
eworkin
gwi t
honly15observa
ti
ons ,ass
ess
ingthecondi
ti
onsnece
ssa
ryf
or
i
nfer
e ncet
obeva
li
dinr e
gres
sioni
sextr
emelydiffic
ult
.Thiss
c a
tt
erchar
tofther
esi
dua
ls
ve
rsusfact
orx
1doesnotpro
videst
ron
ge vi
denceofav iol
at
ionofthecondi
ti
ons.
Similarly, this scatter chart of the residuals versus factor x2 does not provide strong evidence
of a violation of the conditions, so we will proceed with our inference.
The p-value associated with the F test for an overall regression relationship is 9.87E-09 =
0.00000000987, which is essentially zero.
Because this p-value is less than the 0.01 level of significance, we reject the hypothesis that
β1 = β2 = 0. We conclude that there is an overall regression relationship at the 0.01 level of
significance
The p-value associated with the estimated regression parameter b1 is 2.41E-05 = 0.0000241,
which is essentially zero. Because this p-value is less than the 0.01 level of significance, we
reject the hypothesis that β1 = 0. We conclude that there is a relationship between yield and
Copyright Cengage Learning. Powered by Cognero. Page 19
CH 7 - Linear Regression
factor x1 at the 0.01 level of significance. The best estimate is that if we hold factor
x2 constant, a 1 unit increase in factor x1 corresponds to an average increase of 4.49 unit
change in yield.
The p-value associated with the estimated regression parameter b2 is 0.2607. Because this p-
value is greater than the 0.01 level of significance, we do not reject the hypothesis that β2 = 0.
We fail to conclude that there is a relationship between factor x2 and yield at the 0.01 level of
significance.
The estimated regression parameter b0 suggests that when factors x1 and x2 are both zero, the
yield is –138.97 units. This result is obviously not realistic as yield cannot be negative.
b.Thec oeffic
ientofde te
rmi natonR2i
i s0.9537 ,sot hisr egr essionmode le xplains
appr ox i
ma tely95% oft hev ariat
ioni nthes ampl ev a l
ue sofyi eld .
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: INFERENCE AND REGRESSION, Pages 315-332
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
58. Ast
ude
ntisi
nte
res
tedinst
udyi
ngtheimpactofn
umberofboo
ksr
efe
rre
dbyst
udentsonasta
ti
sti
cscour
seandthe
numberofle
ctur
est
heyatt
endedonth
efinalgr
adeonth
ecours
e.Asa
mpleof25s
tudent
sisse
lect
edandtheda
taisgi
ven
below.
a
.De vel
opanest
imate
dr e
gress
ionequat
ionusi
ngn umbe
rofbooksref
err
edandthenumberofl
ect
ure
sat
tende
dtopr
edi
ct
t
hefinalgr
adeonthecour
se.
b
.Josephref
err
ed4booksa ndatt
ended19lec
tur
es.Whatish
ispredi
ct
edfina
lsc
oreonthecour
se?
CH 7 - Linear Regression
ANSWER:
a.Thefol
lowingEx ce
loutputpr
ovidesthee
sti
matedmult
ipl
eli
nea
rre
gre
ssionequa
tionwit
h
numberofbooksr e
ferredbyst
udentsonasta
ti
sti
cscour
seandt
henumberofle
ctur
esthe
y
at
tende
da stheindependen
tvari
ables.
Thee
sti
mat
edmu
lti
plel
ine
arr
egr
ess
ione
qua
ti
oni
s .
b
.IfJ
ose
phr
efe
rs4boo
ksa
nda
tt
ends19l
ect
ure
s,hi
spr
edi
ct
edfina
lsc
orewi
llbe
ora
ppr
oxi
mat
el
y75.
POINTS: 1
DIFFICULTY: Easy
REFERENCES: THE MULTIPLE REGRESSION MODEL, Pages 310-315
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
59. Ast
ude
ntisi
nte
res
tedinst
udyi
ngtheimpactofn
umberofboo
ksr
efe
rre
dbyst
udentsonasta
ti
sti
cscour
seandthe
numberofle
ctur
est
heyatt
endedonth
efinalgr
adeonth
ecours
e.Asa
mpleof25s
tudent
sisse
lect
edandtheda
taisgi
ven
below.
CH 7 - Linear Regression
4 21 82
5 14 61
3 18 54
3 13 46
1 8 64
5 22 90
a.UsetheFtes
ttodete
rmi netheoverallsi
gnifica
nc eoft herelat
ionsh i
p.Wha tisyourconcl
usion
att
he0.05le
velofsi
gnifican
ce?Uset hettesttod et
e r
mi nethes igni
ficanceofe ac
hindependent
var
iabl
e?Whatareyourc oncl
usi
onsa tthe0.05l ev
e lofsignific
an ce
?
b.Howmuc hofthevar
ia t
ioninthefinalgr
adedoe sthemode linpa rt(a)expla
in?
ANSWER: a.Totes
ta nyhypo theses,wec hec kthec onditi
onsneces
saryforvali
dinfe
renceinr
egr
ess
ion
.
TheEx c
e lplot
soft here si
dua l
sa nde achoft hetwoindepe
nde ntvar
iabl
esfol
low.
Be c
auseweareworki
ngwithonly20obse
rvati
ons,assessi
ngthecondi
ti
onsnec
essa
ry
forinf
ere
ncetobeval
idi
nr e
gres
sioni
sextr
e mel
ydifficul
t.However
,th
issca
tt
erchar
t
doesnotpro
videst
ron
gevidenc
eofav i
ola
tionofthec ondit
ions
.
Simil
arl
y,t
hiss
cat
terchar
tdoesnotpr
ovidest
ron
gevi
den
ceo
favi
ola
ti
onoft
he
condi
ti
ons,s
owewi llpr
oceedwit
hourinfer
ence
.
Ex
celOut
put
:
CH 7 - Linear Regression
Thep-va
lueassoci
ate
dwiththeFtestforanoveral
lregress
ionrel
at
ionshipi
s0.0151 .
Beca
usethisp-v
a l
ueisl
esst
hanthe0.05levelofsi
gnific
ance,werej
ectthehypot
hes i
s
t
hatB1 =B2=0 .
Weconcludethatt
herei
sano v
eral
lregres
sionrel
at
ionshipatthe0.
05levelofsi
gnific
anc
e.
Thep-val
ueassocia
tedwi ththee s
ti
matedregres
sionpara
mete
rb1is0.1945.Because
t
h i
sp-val
ueisgreat
erthant he0.05lev
elofsigni
fica
nce,wedonotrej
ectthehypot
hesi
sthat
B1=0.
Wec annotc
oncludethatthereisarel
ati
onshipbetweennumberofbooksrefe
rredandthefina
l
score
atthe0.
05leve
lofs i
gnific
an ce
.
Thep- val
uea ssoci
at
edwit
htheest
imatedre
gress
ionpar
ame t
erb2is0.
2081.Be c
ausethep-
value
i
sgr eate
rthant he0.
05l
evelofsi
gni
fic
ance,wedonotrej
ectthehypot
hes
isthatB2=0 .We
cannot
concludethattherei
sare
lat
ions
hipbetwee
nt h
enumb e
ro fl
ectur
esatt
ende
da ndthefinal
scoreofstudents.
POINTS: 1
DIFFICULTY: Challenging
REFERENCES: INFERENCE AND REGRESSION, Pages 315-331
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
60. Asurveyconducte
db yares
earcht
eam wa
stoi
nves
ti
gat
ehowthee ducati
onle
vel
,te
nurei
ncurr
ente
mpl
oyme
nt,
and
age,ar
er el
ate
dtoa nnuali
ncome.As a
mple20empl
oyee
sisse
lec
teda ndt hedat
aisgi
venbel
ow.
Lengthoftenurein
Educa t
ion(No .of curr
ent
Age(No
.ofyear
s)Annuali ncome($)
years) employment(No.
o
fyears)
17 8 40 124, 000
12 12 41 30 ,
000
20 9 44 193, 000
Copyright Cengage Learning. Powered by Cognero. Page 23
CH 7 - Linear Regression
14 4 42 88
,000
12 1 19 27
,000
14 9 28 43
,000
12 8 43 96
,000
18 10 37 110,000
16 12 36 88
,000
11 7 39 36
,000
16 14 36 81
,000
12 4 22 38
,000
16 17 45 140,000
13 7 42 11
,000
11 6 18 21
,000
20 4 40 151,000
19 7 35 124,000
16 12 38 48
,000
12 2 19 26
,000
10 6 44 124,000
a.Determi nethee s
ti
ma t
edmu lt
ipleli
nearr egres s
ioneq uat
ionthatcanbeusedt opredi
ct
t
hea nnua lincomegivennumb erofy e
a r
ss choo lc ompleted(Educat
ion)
,lengthoftenure
i
nc urr
e nte mployment,andage.
b.Uset heFt esttodet
e r
minetheo ver
allsignifican c
eo ftheregr
essi
onrela
tionship.
Wha tisthec onclus
ionatthe0.05levelofs i
gni fic
a nc
e?
ANSWER:
a.Thef ol
lowin gEx celoutputprovi
destheest
ima t
edmu lti
plel
ine
arr
e gre
ssionequa
ti
onwi
th
educati
on( x)
1, l
e ngthoft enureincurre
ntemployme nt(x2)
,age(3)
x astheindepe
ndent
varia
bles.
Thee
sti
mat
edmu
lti
plel
ine
arr
egr
ess
ione
qua
ti
oni
s
.
b
.Beforeper
for
minganyhypot
hes
ist
est
sonther
esu
lts
,wechec
kthecondi
ti
onsnec
e s
sar
yfor
v
ali
dinfer
encei
nr e
gre
ssi
on.TheExc
elpl
otso
fther
esi
dual
sandeduc
ati
on,l
engt
ho ft
enure
i
ncurre
ntemployment
,agef
oll
ow.
CH 7 - Linear Regression
Becaus
ewea reworki
ngwi t
honly20ob s
e r
vat
ions,assess
ingthecondi
ti
onsneces
sary
f
orinfe
rencetobeval
idinregr
ess
ionisextr
eme l
ydiffic
ult
.Ho wever
,noneoft
hesescat
ter
cha
rtsprovi
destr
ongevide
nceofaviola
tionoftheconditionsnec
essar
yforva
lidinf
ere
nce
i
nregres
sion,sowewillpr
oceedwit
hou rinf
erenc
e.
Thep-val
uea ssoc
iate
dwitht heFtes
tforanovera
llre
gress
ionrel
ati
onshipis0.00039.
Becaus
ethisp-valueisl
essthanthe0.
05l e
velofsi
gni
fic
ance,werej
ecttheh ypot
hesi
s
t
hatB1=B2=B3=0 .Wec oncludet
hatther
eisanovera
llregr
ess
ionrela
tionshi
pa tt
he0.
05
l
evelofsi
gnific
ance.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: INFERENCE AND REGRESSION, Pages 315-331
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
Copyright Cengage Learning. Powered by Cognero. Page 25
CH 7 - Linear Regression
KEYWORDS: Bloom’s: Application
61. Asur
veyconducte
db yares
ear
cht
eam wa
stoi
nve
sti
gat
ehowtheeduc
ati
onle
vel
,te
nurei
ncurr
ente
mpl
oyme
nt,
and
age,ar
erel
ate
dtoa nnuali
ncome.Asa
mple20emp
loy
eesisse
lec
teda
ndthedat
aisgi
venbel
ow.
Le
ngt
hoftenurei
n
Educ
ati
on(No
.ofy
ear
s)c
urre
ntempl
oyment(No
. Ag
e(No
.ofy
ear
s) Annuali
nco
me(
$)
ofye
ars)
17 8 40 124
,000
12 12 41 30,
000
20 9 44 193
,000
14 4 42 88,
000
12 1 19 27,
000
14 9 28 43,
000
12 8 43 96,
000
18 10 37 110
,000
16 12 36 88,
000
11 7 39 36,
000
16 14 36 81,
000
12 4 22 38,
000
16 17 45 140
,000
13 7 42 11,
000
11 6 18 21,
000
20 4 40 151
,000
19 7 35 124
,000
16 12 38 48,
000
12 2 19 26,
000
10 6 44 124
,000
a.Che ckiftheFt e
s tleadstoconcl
udet hatanoverallre
gress
ionr elati
onshipexi
sts
.I fyes,usethettes
tto
dete
rmi nethesigni
fic a
nc eofeachindependentvari
ab l
e.Wha tisthec onclusi
onfore ac
ht estatt
he0.05leve
lof
si
gnific
a n
ce?
b.Remo vea lli
ndepe ndentvar
iable
st hatarenotsi
gn i
fic
antatthe0 .05l e
velofsi
gnifica
ncef romtheesti
mated
regr
essioneq ua
tion.Wh ati
syoure st
ima t
edregres
sione qu
ationint hiscase?
ANSWER:
a.Usingthefollowin
gEx ce
loutput,wec oncl
udethatthereisano ver
allr
egres
sion
re
lati
onsh i
patthe0 .
05le v
elofsignific
a nce.
CH 7 - Linear Regression
Thep-
val
uea
sso
cia
tedwi
tht
hee
sti
mat
edr
egr
ess
ionpa
rame
terb1i
s0.
0013
.
Bec a
us ethisp-val
ueislesst
hant he0.05levelofs
ignific
anc e,werej
e c
tthehypo t
hesi
st tB1
ha
=0 .
Wec onc ludethatt
hereisarelat
ionshi
pbe tweentheannua lincomea ndtheeducati
onleve
lat
the0.05l evelofsi
gni
fic
a nce
.
Ourbe ste st
imatei
sthatifwehol dthelengthofte
nurei nc urre
ntemployme ntandtheage
constant,a
oney earincreasei
nthee ducat
ionlevelcorr
espondstoa nincreaseof$10,011.92inannua
l
income ,ona ver
age.
Thep- v
alueassocia
tedwi ththeesti
ma t
edre gres
sionparame t
erb2is0.
3246.
Becausethisp-val
ueisgre at
erthanthe0.05l eve
lofsignific
ance,wedonotrej
ectt
he
hypothes
isthatB2=0.
Wec annotconcludethatthereisarela
ti
on s
hi pbetweenthel en
gthoftenur
eincurr
ent
employme ntand
t
hea nnualincomea tthe0.05levelofsi
gnificancewhenc ontrol
li
ngforeduca
ti
onleve
land
age.
Thep-
val
uea
sso
cia
tedwi
tht
hee
sti
mat
edr
egr
ess
ionpa
rame
terb3i
s0.
0149
.Be
caus
ethi
sp-
valueislesst hanthe0 .
05leve
lofsi
gnifica
n c
e,werej
ectthehypot
hesist
hatB3=0 .We
concludetha tthereisarel
ati
onshi
pbetwe enthea
g ea
ndthea nnuali
ncomeatthe0.05le
velo
f
si
gnifican
ce .Ourbe s
test
imateist
hatifweho ldt
heeducati
onlevelandle
ngt
ho fte
nurein
currentempl oyme ntconst
ant
,anincre
a seinagebyayearcorre
spondstoanincr
easeof
$2,689.24ina nnualinc
ome .
b.Thef ol
lowingExceloutputi
sobtaine
db yre movi
ngtheinde pe
ndentvari
abl
e,l
engt
hof
tenurein
currentemplo
yme nt(
x 2)
,whichisnotsi
gnific
a nta
t0.05si
gnifica
ncelev
el.Hence
,thi
sout
put
provides
theestimat
edmulti
p l
el i
nearre
gres
sionequa t
ionwitheduc
a t
ion( 1)
x ,
age(x3)ast
he
independentva
ria
bles.
Thee
sti
mat
edmu
lti
plel
ine
arr
egr
ess
ione
qua
ti
oni
s
.
Copyright Cengage Learning. Powered by Cognero. Page 27
CH 7 - Linear Regression
POINTS: 1
DIFFICULTY: Challenging
REFERENCES: INFERENCE AND REGRESSION, Pages 315-331
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
62. Cons
idert
hef
oll
owi
ngda
tawi
tht
hede
pende
ntv
ari
abley
,inde
pende
ntv
ari
abl
ex,
andt
hedu
mmyv
ari
abl
ed.
Y d x
100.0 0 9
.3
100.0 0 6
.5
50
.0 0 4
.2
75
.0 1 7
.4
65
.0 0 6
.0
90
.0 0 7
.6
85
.0 1 9
.6
65
.0 1 6
.5
50
.0 1 6
.0
75
.0 1 9
.9
75
.0 1 8
.6
65
.0 0 4
.9
95
.0 0 7
.2
95
.0 0 9
.9
90
.0 1 7
.8
80
.0 1 7
.2
75
.0 0 7
.0
95
.0 0 8
.7
100.0 1 9
.9
90
.0 1 9
.3
70
.0 1 9
.1
100.0 0 8
.9
40
.0 0 5
.4
90
.0 0 8
.8
75
.0 1 8
.8
40
.0 1 5
.0
45
.0 1 6
.5
55
.0 0 5
.2
100.0 0 10.
0
90
.0 0 8
.4
80
.0 1 6
.4
80
.0 0 6
.7
90
.0 1 8
.5
60
.0 0 5
.9
95
.0 1 8
.9
45
.0 0 6
.3
65
.0 0 7
.5
85
.0 1 6
.8
60
.0 1 5
.7
CH 7 - Linear Regression
75
.0 1 4
.5
a.Deve
lopthees
ti
ma t
edregres
sionequati
onusin
ga l
lo ft
heindepende ntvar
iablesincl
udedinthedat
a.
b.Te
stfora
no v
eral
lregr
essi
onre l
ati
onshi
patthe0.
05l eve
lofsignific
ance.Isthereasigni
fic
antre
gres
sionr
ela
ti
onsh
ip?
ANSWER:
a.Thefollo
wingEx c
eloutputpro
videsthee st
imatedmu l
ti
plel
inearr
egre
ssi
onequati
ont
hat
coul
d
beus e
dtopr e
dic
tygiventhedummyv ar i
able
,d, andx.
Thee
sti
mat
edmu
lti
plel
ine
arr
egr
ess
ione
qua
ti
oni
s .
b
.Befor
etes
ti
nganyh
ypothes
esforthi
sregr
ess
ionmodel
,wechec
kt hecond
iti
onsnece
ssa
ry
f
or
v
ali
dinf
ere
nceinr
egr
ess
ion.Exc
e lp
lotsoft
heresi
dua
lsande
achindepe
ndentvar
iabl
e
f
oll
ow.
CH 7 - Linear Regression
Theres
idua
lsappeart
oha v
eameanofzeroa
nddonota
ppeart
obebadl
yske
weda
tan
yva
lue
ofan
yindependen
tv a
ria
ble.Wet
her
efor
ewil
lpr
ocee
dwithouri
nfe
renc
e.
The p-value associated with the F test for an overall regression relationship is 7.813E-07 =
0.0000007813, which is essentially zero.
Be ca
uset hisp-valuei slesst hant he0. 05l e v
e lofsigni fic
an ce,wer ejecttheh ypo t
hesisthatB1
=B2=0
andc
onc
ludet
hatt
her
eisa
nov
era
llr
egr
ess
ionr
ela
ti
ons
hipa
tthe0
.05l
eve
lofs
igni
fic
anc
e.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: CATEGORICAL INDEPENDENT VARIABLES, Pages 332-337
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
63. Cons
iderthef
oll
owi
ngda
tawi
ththede
pende
ntv
ari
abl
ey,inde
pende
ntv
ari
abl
ex,
andt
hedu
mmyv
ari
abl
ed.
Y d x
100.
0 0 9.
3
100.
0 0 6.
5
50.0 0 4.
2
75.0 1 7.
4
65.0 0 6.
0
90.0 0 7.
6
85.0 1 9.
6
65.0 1 6.
5
50.0 1 6.
0
75.0 1 9.
9
75.0 1 8.
6
65.0 0 4.
9
95.0 0 7.
2
95.0 0 9.
9
90.0 1 7.
8
80.0 1 7.
2
75.0 0 7.
0
95.0 0 8.
7
100.
0 1 9.
9
90.0 1 9.
3
Copyright Cengage Learning. Powered by Cognero. Page 30
CH 7 - Linear Regression
70.
0 1 9
.1
100
.0 0 8
.9
40.
0 0 5
.4
90.
0 0 8
.8
75.
0 1 8
.8
40.
0 1 5
.0
45.
0 1 6
.5
55.
0 0 5
.2
100
.0 0 10.
0
90.
0 0 8
.4
80.
0 1 6
.4
80.
0 0 6
.7
90.
0 1 8
.5
60.
0 0 5
.9
95.
0 1 8
.9
45.
0 0 6
.3
65.
0 0 7
.5
85.
0 1 6
.8
60.
0 1 5
.7
75.
0 1 4
.5
a.Chec kiftheoveral
lr e
gress
ionr ela
ti
onshi
pe xist
sfortheabo vedata
.Ify e
s,tes
tthere l
ati
onshipbetweenea c
h
i
nde pendentvari
ablea ndthedepe ndentvar
iabl
ea tt
h e0.05levelofsi
gnific
ance,andin t
erpr
ettherelat
ions
hipbet
wee
n
eachoft heinde
pe ndentvari
ablesa ndthedependentvari
a bl
e.
b.Ho wmuc hoft heva ri
ati
oninthes amplev al
uesofdelaydoe sthi
se s
ti
ma tedregres
sione q
uati
one xplai
n?
ANSWER:
a.Us ingthefol
lowingEx ce
lout put,weconcludethatther
ei sanoveral
lregres
sion
rela
tionshipatt
he0 .05leve
lofs igni
fic
ance.
Wea l
sohavechec
kedthenec
essa
rycondi
ti
onsf
o rva
li
din
fer
encei
nr e
gre
ssi
onintheprevi
ous
ques
ti
on.
Hence,t
hep-v
aluefo
rthete
stoftheh
ypot
hesi
sthatB1=0is0.
1540.Bec
auset
hisp-v
alueis
grea
tert
hanthe0.
05leve
lofs i
gni
fic
ance,wedonotr
eje
ctt
hehypot
hes
isthatB1=0, a
ndwe
concl
udetha
tthe
reisnodiffe
renc
einthevari
abl
esyanddcont
rol
li
ngfortheva
ria
blex.
CH 7 - Linear Regression
Thep-
val
uef
ort
het
estoft
heh
ypo
the
sist
hatB2=is 1.52E-07 = 0.000000152.
Bec
aus
ethi
s
p-val
ueisle
sst ha
nthe0.05leve
lo fs
igni
ficance,werej
ectt
hehypothesi
sthatB2=0, and
concl
udethattherei
sadiffe
rencebet
weenv a r
iab
lesyandthei
ndepe ndentvar
iablexatt
he
0.05le
velofsigni
fic
ance
. Wee s
timat
ethat,holdi
n gt
hevari
abl
edc onstant,1uniti
ncr
easei
n
xincre
ases8.010unit
sofy ,
onaverag
e.
b
.Thecoe
ffic
ientofde
termi
nat
i sR2=0.
oni 5324,s
other
egr
essi
onmode
lexpl
ai
ns
a
ppr
oxi
mately53% oftheva
ria
ti
onint
heyv
alue
sinthes
ample
.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: CATEGORICAL INDEPENDENT VARIABLES, Pages 332-337
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
64. Aproduc
ti
oncompa
nyisst
udyi
ngthere
lat
ions
hipbe
tweent
heav
era
gec
ost
/uni
tandnumb
ero
fun
itspr
oduc
edi
na
batch.Asampleo
f10bat
chesi
ssel
ect
edandthedat
aisgi
venbe
low.
CH 7 - Linear Regression
As i
mplelinearr
egres
sionmodeldoesno ta
ppeartobea pp
ropria
te;t
hereappe
arstobea
nonl
ine
a rr
e l
at
ionshi
pbe t
weencost
/unitandnumbe rofuni
tsproducedinabat
ch.
b.Thefoll
owingEx ce
loutputpr
ovi
de stheest
imate
ds i
mplelinearre
gress
ionequa
tiontha
t
coul
dbeus edtopredi
ctcost/
uni
t(y)giventhenumberofunit
spr oduce
dinabatch(x).
Thee s
ti
ma te
dmu l
ti
pleli
nea
rregres
sionequat
ioni
s ,andthe
2
coeffic
ientofde
ter
minati
onfort
hismode li
sR =0.9177,s
otheregr
ess
ionmodele
xplai
ns
appro xi
ma t
el
y92% ofthevar
iat
ioninthesamplev
alue
sofcos
t/uni
t.
POINTS: 1
DIFFICULTY: Easy
REFERENCES: THE SIMPLE LINEAR REGRESSION MODEL, LEAST SQUARES METHOD Pages 297
and 304
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
65. Aproduc
ti
oncompa
nyisst
udyi
ngthere
lat
ions
hipbe
tweent
heav
era
gec
ost
/uni
tandnumb
ero
fun
itspr
oduc
edi
na
batch.Asampleo
f10bat
chesi
ssel
ect
edandthedat
aisgi
venbe
low.
CH 7 - Linear Regression
produced
20 37.7158
35 35.0158
50 30.8158
65 25.8158
80 20.0364
95 16.9064
110 16.3766
120 13.8564
135 13.9696
150 13.847
a.De ve
lopa ne s
ti
ma te
dq uadrat
icregressione quationforthedata.Howmuc hva r
iat
ioninthesamplevaluesofcost/
uni
t
doesthisr
e gre
s s
ionmode lexplai
n?
b.Istheove r
allregr
essi
onr el
ati
onships ignific
anta ta0.05levelofsi
gnific
ance?Ifso,thente
stt
h ere
lat
ionshipbetween
theindependentva r
iabl
ea ndthedepen de ntvaria
bl eata0.05levelofsi
gn i
fic
ance.
ANSWER:
a.Thef ol
lo win gEx celoutputpro
v i
desthees t
imatedsecondorderq
ua dr
ati
cr e
gress
ion
equati
ont hatc ouldb eusedt opredic
tcost
/un i
t(y)give
nt henumberofunit
spr oduced(x)
.
Thee
sti
mat
eds
econdor
derq
uadr
ati
cre
gre
ssi
one
qua
ti
oni
s
,andthecoeffici
entofdet
ermi
nati
onf orth i
smode lisR2=
0.9824,sothequadra
ticregre
ssionmodelexpl
ai
nsa ppro
x i
mate
ly98% oft hev ar
iat
ionint he
samplev a
luesofcost
/uni
t.
b.Be f
oretes
tin
ga nyhypothesesaboutt
hisre
gre
s sionmode l
,weagai
nc hec ktheconditions
necessar
yforvali
dinfere
nc einregr
essi
on.TheEx ce
lplotoft
heres
idualsa ndvehic
les peed
foll
ows.
CH 7 - Linear Regression
The s
es c
a tt
erchartsdono tpro
vides t
rongevide
nceofav iola
ti
onoft hec onditi
on s,sowewi l
l
proceedwi thourinferenc
e .
Thep- val
uea ssociat
edwi t
htheFt estforanoveral
lregre
ssi
onr e
lati
onshipi s7.24E-07 =
0.000000724. Beca us
ethisp-val
ueisl e
ssthanthe0.05leve
lofsigni
fic
a nce, wer e
jectthe
hypothesisthatB1=B2=0a ndc oncludet
h a
ttherei
sa noveral
lregr
essionr ela
tionshipatt
he
0.05levelofsignific
ance.
Thep- val
uef orthetestofthehypothesi
sthatB1=0i s4.
32E-05.Becaus ethisp-valueisles
s
t
hant
he0.
05l
eve
lofs
igni
fic
anc
e,wea
gai
nre
jec
ttheh
ypo
the
sist
hatB1=0.
Simi
la
rly
,thep-
v
aluef
ort
het
estoft
heh
ypo
the
sist
hatB2=0i
s0.
0014.
Bec
aus
eth
isp-
val
uei
sle
sst
hant
he
0.05le
velo
fsi
gnific
ance,werej
ectth
eh ypo
thesi
sthatB2=0.Wet
heref
oreconc
ludetha
t
ther
eisanonl
ine
arr e
lat
ions
hipbetweencost
/un
itandnumbe
rofun
itsproduc
edinthebat
ch.
POINTS: 1
DIFFICULTY: Easy
REFERENCES: MODELING NONLINEAR RELATIONSHIPS, Pages 337-342
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
66. Consi
derthebelowda tawhi chisbas
e donac ompany
’ss
ale
sint
hepe
riod2000t
o2011a
lon
gwi
tht
hen
ati
ona
l
incomeoft hecountr
y,whe r
et hebusi
nessiss e
tup.
NationalInc omeCompany '
ssal
es
Year (
in mi ll
ion sof (i
n thou s
andsof
dollars)x dollars
)y
2000 305 470
2001 316 485
Copyright Cengage Learning. Powered by Cognero. Page 35
CH 7 - Linear Regression
2002 358 499
2003 350 515
2004 375 532
2005 392 532
2006 400 556
2007 398 576
2008 430 583
2009 456 587
2010 578 601
2011 498 605
a.Devel
opas cat
terchartf
orthe sedat
a,tr
eati
ngthenati
onalincomea stheindepende ntv
ari
able
.Doesasimplel
ine
ar
regr
ess
ionmode lappeartobea ppropr
iat
e?
b.Devel
opa nappropri
ateest
ima tedre
gress
ionequat
iontopr e
dictthecompa ny'
ssales,gi
venthena
ti
onalinc
ome.How
muchv ar
iat
ioninthesamplev aluesofcompany’
ssalesi
se xplai
nedb ythisr
e gr
essionmodel
ANSWER:
a
. Thes ca
tt
erchartwit
hna t
ionalincomea stheindependentvar
iabl
efol
lows
.
As i
mpleline
arregress
ionmode ldoesnotappea
rt obeappr
opri
ate
.Th e
reappe
arst
obea
non-l
inea
rrel
ati
ons h
ipbetweenna t
ionali
ncomea ndcompany’
ssale
s.
b.Basedonthescatt
erchart,weesti
matequadra
ticregr
ess
ionequat
iontha
tcoul
dbeus
edto
nat
ionali
ncome(y )gi
venc ompany’ssa
les(x)
.
CH 7 - Linear Regression
Thee s
ti
ma t
e dmu lt
ipl
eli
nearregre
ssionequat
ionis ,
2
andthecoeffici
entofdet
erminat
ionforthismodeli
sr=0 .
9380,s
other
egre
ssi
onmode
l
expl
ainsappr o
ximatel
y94% o ft
hev ari
at
ioninthesa
mplev
alueso
fcompan
y’ss
ale
s.
POINTS: 1
DIFFICULTY: Easy
REFERENCES: MODELING NONLINEAR RELATIONSHIPS, Pages 337-342
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
67. Gi
veanest
imat
edsi
mp l
el i
nearr
egres
sionequat
ionof =46 .
2+589.
2xwi
thac
oeffic
ientofd
ete
rmi
nat
ionR^
2of
0.7523,i
nte
rpr
ett
hecoe
ffic
ientofdet
erminat
ionforthi
seq
uat
ion.
ANSWER: The regression model estimate explains approximately 75.2% of the variation the y-values
RATIONALE: The coefficient of determination is a measure of the goodness of fit of the estimated
regression equation.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: ASSESSING THE FIT OF THE SIMPLE LINEAR REGRESSION MODEL, Page 308
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
68. The multiple regression model represents pricing for residential housing in a certain market. Predicted Price =
19,856.56 + 6985.25 bedrooms + 87.53 square foot. A house in the local market has 4 bedrooms and 2500 square feet of
living area. Use the multiple regression model to determine the predicted price.
ANSWER: $266,
623
RATIONALE: Predicted Price ̂ = 19,856.56 + 6985.25 bedrooms + 87.53 square foot = 19,856.56 +
6985.25(4) + 87.53(2500) = 266,622.56.
POINTS: 1
DIFFICULTY: Moderate
CH 7 - Linear Regression
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 310
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
69. The multiple regression model represents pricing for residential housing in a certain market. Predicted Price ̂ =
19,856.56 + 6,985.25 bedrooms + 87.53 square foot. A house in the local market has 5 bedrooms and 3,500 square feet of
living area. Use the multiple regression model to determine the predicted price and the residual if the house sells for
$360,200.
ANSWER: Predicted Price = $361,138, Residual = –$937.81
RATIONALE: Predicted Price ̂ = 19,856.56 + 6985.25 bedrooms + 87.53 square foot = 19,856.56 +
6985.25(5) + 87.53(3,500) = 361,137.81. Residual = actual - predicted = 360,200 –
361,137.81 = –937.81.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 315
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
70. A multiple regression model has the form = 10 + 2 x1 + 12 x2. As x1 increases by 1 unit (holding x2 constant), is
expected to increase by how many units?
ANSWER: 2
RATIONALE: As x1 increases by 1 unit (holding x2 constant), is expected to increase by 2 units. This is
the case because the value x2 has no effect on if it is held constant and 10 is a constant, so as
x1 changes by 1, the value of will change by 2…the coefficient of x1.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 310
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
71. A multiple regression model has for predicted heart rate is as follows: heart rate = 10 – 0.4 run speed + 12 body
weight. As the run speed increases by 1 unit (holding body weight constant), heart weight is expected to increase by how
much?
ANSWER: 0.4
RATIONALE: As the run speed increases by 1 unit (holding body weight constant), heart weight is expected
to decrease by 0.4 units. This is the case because the coefficient on the only term that is
changing is 0.4. All other terms remain constant.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 310
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
72. Given the partial Excel output from a multiple regression, formulate the regression model.
CH 7 - Linear Regression
73. The multiple regression model represents pricing for residential housing in a certain market. Predicted Price ̂ =
19,856.56 + 6,985.25 bedrooms + 87.53 square foot. A house in the local market has 4 bedrooms and 2,200 square feet of
living area. Use the multiple regression model to determine the predicted price.
ANSWER: $240, 364
RATIONALE: Predicted Price ̂ = 19,856.56 + 6,985.25 bedrooms + 87.53 square foot = 19,856.56 +
6,985.25(4) + 87.53(2200) = 240,363.56
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 315
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application
74. The multiple regression model represents pricing for residential housing in a certain market. Predicted Price ̂ =
19,856.56 + 6,985.25 bedrooms + 87.53 square foot. A house in the local market has 5 bedrooms and 3,200 square feet of
living area. Use the multiple regression model to determine the price and the residual if the house sells for $352,200.
ANSWER: predicted price = $334,879, error = – $17,321
RATIONALE: Predicted Price ̂ = 19,856.56 + 6,985.25 bedrooms + 87.53 square foot = 19,856.56 +
6,985.25(5) + 87.53(3200) = 334,878.81. Residual error = actual – predicted = 334,879 –
352,200 = –17,321.
POINTS: 1
DIFFICULTY: Moderate
REFERENCES: THE MULTIPLE REGRESSION MODEL, Page 315
NATIONAL STANDARDS: United States - BUSPROG: Analytic skills - and DISC: Descriptive Statistics
KEYWORDS: Bloom’s: Application