This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Econ 321 MidTerm February 25‘“, 2009, 2:303:50 Carrie" Q X Student ID number Instructions:
*Write your name on top of the next pages.
*You can use the other side of the page to complete your answers. *Problems should be answered as completely as possible. Partial credit on problems is only
possible if I can locate errors in your calculations *Neatness and organization of your answers is essential Question 1 (35 pts)
. Question 2 (40 pts) Question 3 (25 pts) qoopl Luck! Question 1 (35 pts) Our firm produces lubricants for oil well drills and we are trying vario ,gawnasmmprgyjag
We have tried additional advertising in trade journals; we have also tried investing more in training our
sales force and created a sophisticated incentive scheme, we have also been known to provide the
purchasing managers for certain large users with signiﬁcant amounts of cash in small unmarked bills as
an inducement to use our product. We have 130 sales ofﬁces around the world and, after a one year
trial of our new approach to sales, we plan to estimate a regression of the form. Salesi = [30 + Bl*Advertisingi + [32*Trainingi + [33*Bribesi + 8i a) (5pts) Is this a population or a sample regression. What is the difference between a population and a
sample regression? M Cb) ‘ , M ._ A ._ "7 l, .
. Bl : fa bf), 1 M 5+ (11 (Mb; 7/53 511% laud or». (3033”be ”mu 4.) ll/M ‘i We go ahead and estimate the following regression on our 130 observations. The
estimate of the regression line is (all variables measured in $1,0005): {ales = 1000 + 20*Advertisingi+ 30*Trainingi + 50*Bribesi
‘ (500.2) (5.1) (10.3) ‘ (34.6)
R2 = .35 R2__adjusted =32
Where standard errors of estimates are in parenthesis. b) (8pts) interpret each ofthe estimated parameters. L O CLJJCI.‘
“‘  v whip
t: make/l Sade a: «ﬁlm M $0 «a. .9”,
E)“ X: Flam j {36 {WAR it; _ 0321) 5291? QC; HALO“ M l/\ 1" ﬁg rﬁVL A“ $1 0‘33 l g
\ exalt/£337 ,7“ kill?» Lab 39, (93015 A (,
// t/ .s ml
z ,r /\ r ' "Edi L Student Name c) (2pts) What is the prediction of sales for a location with $1M in Advertising, $100, 000' In Training and
$750,000 In Bribes? (6.0% ts QDXU (mkw. > I— 30 ((03,999)
+ ‘90 (‘m/ we) _ ‘ 57:3” 03 0:20
{3013 ’20 090 31;} 000,033+ / 6195) (DC / ) (2. Spts) What' Is the definition of R2? Why do we prefer the adjusted R2? Q1: mm ﬂ SCH/wﬁ UCMACXUJM ‘n \/\‘ {Kg who 5) b} ‘WW LN A
[ ﬂmt/(x WadiLu»; \\/x)((‘ > (L1 83}: ,4" EEK “ , , ,._.. gg‘T SST
e) (2.5 pts) Why don’t we have 8: in this estimated equation? Q Lg LU M bf 0 mi MY E(E/x>:©m f) (Spts) Based on the regression results above, test at the 5% level whether [31 is significant (show the
null and alternative hypothesis, the wide of the test statistic and conclude). ' it \2
M $5790 ° ﬁnal lie at» i Swiwb t 0 ~ g) (Spts) Based on previous experience, we believe that Training should increase sales by $35,000. Do a
hypothesis test for this statement. Student Name h) (5 pts) Why is the variable Training added to the regression model? a) Rig/Jr 19L Candi}, [ﬁe/l ‘VUin
M \i’€« (M) if? Q .5) DM‘IW7 w 'an>.
own foamiu Va“ L‘iCLLs . Question 2 (40 pts) We are estimating a regression equation in which the value lost in property crimes is explained by
expenditures on public safety (SAFEXP) and the population of the government unit (village, town, city,
county, parish), with i=1,..,45 (in other words, the sample size is 45). The latter variable is POPULATION.
The value lost in crimes and eXpenditures on public safety are measured in thousands of dollars. a) (20 pts) For each of the 4 functional forms below, explain, as clearly as possible, the relationship
between a one unit increase in every explanatory variables and the change in the value lost in property ~ \XJ crimes. 1) VALUELOST,= 5,000 — 25*ln(SAFEXP,~) + 500*POPULATION,
R2: 0.10 Rzadjusted: 0.10 . 2) ln(VALUELOST,)= 50 — 1.25*$AFEXP,+ O.13*POPULATION,
R2=0.16 Rzadjusted=015 3) n(VALU'EL05T,)= 436 — 3.63*ln(SAFEXP,)+ 0.05*n(POPULATION,)
2 _ 2 _ '
R — .18 R adjusted.018 4) ln(VALUELO$T,)= 50 + 4.05*$AFEXP,—O.0051*$AFEXPi2+ 0.10*POPULATION,
2_ 2 ‘ _ »
R 020 R adjusted—0.18 K » a , o , 4 ‘ u, j) ‘ aka [03}
l) D‘j 2 F400. A Dx. . l /o (Mask ”' ‘ 1:3 g0g3 6M mm 00400033 f who [M lgjgmlgwfﬁ 49> 74M)“ 95‘) lb )0» “L03 0010 low? maria in ”7W 300% [ml Ly lama W mm. {AM 7* 0.00 lmr Lab, 63%, lO/o gig; in 36%ﬁbf £33 vr’iw»
‘ )O/‘a dd? \\r\ PW 7‘ Liam/k L5); 1:79. OtQS'Sy’L" Li “/vb‘ll v x xvx>xx
V M mm \NN ”Aid ’27 (Aw Student Name b) (5 pts) Can you draw conclusions on which of the four speciﬁcations you prefer? Explain. NUKVQYWC) W E. M (if ha ﬁx yioﬁf: 74 NM QWWB‘L 52)?)ﬁ Ll) W c) (5 pts) Using specification 4), what is the effect of a change in expenditures on public safety on the
dependent variable, the log of value lost? AV) ’ [email protected] («.03 _. ll '©’x><x§‘>$?¥“’<9 ”Bf7v.” You choose the following population regression model:
lanALUELOST,)= [30 + [31*SAFEXP, + [32*SAFEXP,2+ 83*POPULATION,+ 5, Suppose you want to test the null hypothesis that, holding other factors fixed, expenditure on public safetyhas no effect on’the log of value lost.
d) (5pts) Which specification would you need to e) (5 pts) Describe in details the test (hypothesis + define the statistic for the test + distribution of the
statistic + how to conclude) that you would perform to ansWer part d) (in other words to test the null
hypothesis that, holding other factors fixed, expenditure on public safety has no effect on the log of
value lost). '. Student Name
Question3(2\5’pt:l (\CALJ/w «W lﬂbhi {R/urvf'ﬁ'x 4. Q',;
> A justification forjob training programs is that they improve worker productivity. Suppose you are asked
to evaluate whether more job training makes workers more productive. To answer this question, you
are not given data on individual workers but data on manufacturing firms in Ontario. In particular, for
each firm, you are given information on hours ofjob training per worker (Train__hrs) and number of non
defective items produced per worker hour (output). ‘ a) (5 pts) Carefully state the ceteris paribus thought experiment underlying this question. b) (5 pts) What problems do you foresee given the data you have to use for addressing the question?
What factors would you consider that are likely to affect the analysis of the impact of training .
(measurable and possibly unmeasurable by the econometrician)? Stilt \ML/b M Li m rosters; 0dr Qme H We l Part / lbw r/l {“VVKFS ‘
 ‘r Vial/WW (him. ‘of' Ur: haul/ray, W
fin Imam vi; he f mgr whims , c) (5 pts) Suppose you are given information on each firm’s main characteristics (industry, number'of
employees, location) andon worker characteristics for each firm in terms of an average over the whole
workforce (for example ”average number of xxx for the workers at firm i", where xxx could be years of ’
education or age...), write down the population regression you would run to estimate the effect of training on oUtput, consistent with your arguments in part a) and b). GWNV 2 039* oil lme/rs t all Flam dwwl‘
H
\WoMbA—d
Styx ‘ d . ya (it/walLKM What
A? “BK/big: 93% *h 2 ’ kwwm Mal. d) (5 pts) Using your regression equation defined in c), how would a positive relationship between
output and training translate in mathematical terms (in terms of the parameter(s) of the regression equation)? 4‘30. Student Name e) (5 pts) if you do find a positive relationship between output and training, do you have convincingly
established in this analysis that job training makes workers more productive? Explain: {\l O . 05' WW _ lam/NM l3 Wilma) CW’ZWW
lem / (claim Qua it W)
(ﬁlm eye/(CUM . ‘9quin Come/(1d;
LLW‘UK hwulﬂ/ 0AM M )W l) cw mull Mm :3 My 11 T3 E L E {:3 . 1 iiunwhath{"21 3n2n315 Llf'lsii'iff‘l” the 31.31 IM'TLTI l'vifl r3‘uizlfill’l“l‘lé5t.[ 1:3“iSHi[ZzLllili‘ﬂ‘l illﬂlfll'l}? ., 3—} x ' if 1‘, "Q; .' tilliiulnliﬁl? 'iiiiJLll i'ff iii/:3) . , " .7 1 I liHI’Il'F f. _
”A” .‘ .1' I' v'] ’1'."
H}.vl__;o Student Name U “C! I. U w
U ,63‘414 I} 'iLf‘xHLL—l ‘x'um.hj'i'l.l, :1. than lii‘x‘If ..
'. This table mm 5 ruled using the. SL313 ' fun»: ti on n u::v,rm'p.rx.:nlfa. T625 i’
if 1?. 1:: 4 LE (12 firming] '\,".;;1ue:; .~:..i: ‘ ‘ [4‘ _J I" .. .; 3 I'J ."« le f‘JI'LJILJIJlJ w 1143 i' 1..) l' [J in, y.) _~‘ [ ,. ailue 1131‘ a cum—tuned Im't. with 2:1 igif'is 3; :5::‘
I355.
as“ ‘mruz'tivrum il]‘~"lt£i.il. : ’]'_hr'.. Ix; I...) Ina) lw‘ L: In.» l_'.«' J Ix) lJ [2,: L; In; [:2 z «a ._ [J _ I —2 I9: I") _} In) I...» L) L.) IV.) [1 I A I‘JJ Ia.) !.. thgxlg "J L.) In.» I; \ 'riitiml value fora. t1.L'::l:.ilvz‘;l Student Namé TRELE C .‘t . . A
_.____.__.._————————————————————————~—————
" Critiml "$3.31!! a ciiuf ‘t‘ha: = ‘ ‘ Nmmi: ‘Ei 111:" i Z = 4 an" '1 1:111"
1’ :I'um: tit: 11 11.; 15 V trawl? 1331‘.i'lllmﬁﬁl'lfvi‘ ." “
. This". [il.[‘!li’ W11: LLHIFTEIJIZ‘U lan._ the:Emmifunzrtmn.irr."F=t;'1:il. ...
View
Full Document
 Fall '09
 Louis

Click to edit the document details