SUMBUL MUGHAL

MBA CASE STUDY 2019

QUANTITATIVE METHODS AND REASONING
SUBMITTED TO: Dr. Niaz Hussain Ghumro
Tuesday, January 14, 2020

MULTIPLE REGRESSION
1. . The regression model was Y = total sales (thousands of dollars), X1 = display
floor space (square meters), X2 = competitors’ advertising expenditures
(thousands Observations are taken on sales of a certain mountain bike in 30
sporting goods stores of dollars), X3 = advertised price (dollars per unit).
a) write the fitted regression equation
b) interpret each coefficient
c) would the intercept have meaning in this regression?
d) Make a prediction for sales when floor space=80 competing ads=100 and
price=1200

Predictor coefficient
Intercept 1225.44
Floor space 11.52
Competing ads -6.935
Price -0.1496

Solution

A. Sales= 1225.44 +11.52floorspace -6.935 competing ads -0.1496price

B. One additional square meter of floor space will increase sales by
11.52(thousands of dollars). One additional thousands of dollar spend by
competitors on ads will reduce our sales by 6.935(thousands of dollars) and
one additional dollar on per unit of advertisement will reduce the sales by
0.1496(thousands of dollars)
C. Intercept do not have meaning in this regression as there is no floor space
with ‘0’ square meter.
D. Sales = 1225.44 +11.52(80) -6.935(100) -0.1496(1200)
=$1274.02(thousands of dollars)
2. Refer to the ANOVA table below.
a) State the degrees of freedom for the F test for overall significance.
b) Use appendix F to look up the critical value of F for alpha 0.05
c) Calculate the F statistic. Is the regression significant overall?
d) Calculate R2 and R2adj showing your formulas clearly

Source d.f SS MS
Regression 3 1,196,410 398,803
Error 26 379,332 14,590
Total 29 1,575,742
 Solution:
a) Degree of freedom corresponding to the regression is 3 and degree of
freedom corresponding to the error term is 26.
b) F critical = 2.98
Yes, regression is significant because Fcal > Fcritical
c) Fstatistic = MSR/MSE

=398,803
=27.33

d) R2 = SSR/SST
=1,196,410/1,575,742
=0.76
2
R adj =SSE/n-k-1÷SST/n-1
=379,332/26÷1,575,742/29
=0.27

CH13 2ERQ

3. Observations are taken on net revenue from sales of a certain LCD TV at 50

retail outlets. The regression Model was Y= net revenue (thousands of
dollars), X1 = shipping cost (dollarsper unit), X2 = expenditures on print
advertising (thousands of dollars), .X3= expenditure on electronic media ads
(thousands), X4= rebate rate (percent of retail price).
(a) Calculate the t statistic for each coefficient to test for β = 0.
(b) Look up the critical value of Student’s t in Appendix D for a two-tailed test
at α = .01 Which coefficients differ significantly from zero?

Predictor Coefficient SE
4.310 70.82
Intercept
–0.0820 4.678
Ship Cost
2.265 1.050
Print Ads
2.498 0.8457
Web Ads
16.697 3.570
Rebate%
 Solution
a) ship cost: tcalc =-0.0820/4.678
=-0.018
Print ads:tcalc =2.265/1.050
=2.16
web ads: tcalc =2.498/0.8457
=2.95
rebate:tcalc = 16.697/3.570
=4.68

b) t0.01 = 1.684
We cannot determine which coefficients are significant as p-values are not given

CH1 15SE
:

4. A regression model to predict the price of a condominium for a weekend

getaway in a resort community included the following predictor variables:
number of nights needed, number of bed-rooms, whether the condominium
complex had a swimming pool or not, and whether or not a parking garage was
available. (a) Identify the quantitative predictor variable(s). (b) How many
binary variables would be included in the model? (c) Write the proposed model
form for predicting condominium price.

Solution
a) Two quantitative predictors: number of nights and number of
bedrooms
b) Two binary variables would be included: swimming pool(1,0) and
parking garage(0,1)
c) Price= B0+B1 number of nights+ B2 number of bedrooms+B3swimming
pool+B4 parking garage
 5. . Literacy rate is a reflection of the educational facilities
and quality of education available in a country, and mass
communication plays a large part in the educational
process. In an effort to relate the literacy rate of a country
to various mass communication outlets, a demographer
has proposed to relate literacy rate to the following
variables: number of daily newspaper copies (per 1000
population), number of radios (per 1000 population), and
number of TV sets (per 1000 population). Here are the data
for a sample of 10 countries:
TV
literacy
sets
country rate newspapers radios
228
Australia 0.98 280 266
201
Italy 0.93 142 230
2
China 0.25 10 114
227
Norway 0.99 391 313
82
USA 0.79 86 329
11
Russia 0.72 17 42
16
Brazil 0.32 21 49
472
Canada 0.99 314 1695
185
Nigeria 0.99 333 430
89
Indonesia 0.82 91 182
 Solution

y X1 X2 X3
literacy
country rate newspapers radios TV sets Predicted SSE SSR SST
Australia 0.98 280 266 228 1.029621 0.099242 0.063313 0.002462
Italy 0.93 142 230 201 0.407585 -1.04483 0.166126 0.272917
China 0.25 10 114 2 0.396818 0.293635 0.157464 0.021555
Norway 0.99 391 313 227 217.721 433.4621 47402.45 46972.34
USA 0.79 86 329 82 291.1305 580.681 84756.98 84297.61
Russia 0.72 17 42 11 35.08765 68.73529 1231.143 1181.135
Brazil 0.32 21 49 16 12.207 23.774 149.0108 141.3008
Canada 0.99 314 1695 472 0.548 -0.884 0.300304 0.195364
Nigeria 0.99 333 430 185 0.914894 -0.15021 0.837032 0.005641
Indonesia 0.82 91 182 89 0.680526 -0.27895 0.463116 0.019453
average 0.778 1104.687 133541.6 132592.9

INTERPRETATION

 SSE>SST that shows that this model is a good fit

Multiple Regression Analysis

Predictor coefficient SE coefficient t-test p-value

Constant 0.5186 0.09368 5.5 0.002
Newspaper 0.0005421 0.0008653 0.63 0.554
Radios -0.0003535 0.0003285 -1.08 0.323
Television 0.001988 0.00155 1.28 0.247

S 0.1886455
R-Sq 69.90%
R-Sq(adj) 54.80%
INTERPRETATION

 This regression analysis tells that for each one unit increase in newspaper will increase
literacy rate by 0.054%. For each additional unit of radio per 1000 population, literacy rate
will decrease by 0.035% and for each additional unit of television (per1000 population)
literacy rate increases by 0.0198%
 According to the values of coefficients of newspaper, radios and television none of the
predictors differ significantly from zero because of their respective p-values. Only
“television” is somehow near to significantly differ from zero as it has p-value slightly near
to zero.
 According to the t-test, television is the strongest predictor among other predictors as it
has greatest t-value as compared to others.
 Comparison of R2 and R2adj tells that there is somehow large difference between their
values, which indicates that this model contains useless predictors.

Analysis of variance

Source DF sum of squares Mean square f p

regression 3 0.48397 0.16132 4.64 0.053
residual error 6 0.20859 0.03477
Total 9 0.69256

INTERPRETATION

 Degree of freedom corresponding to the regression is3 and degree of freedom

corresponding to the error term is 6.
 48% variation in response data is explained by sum of squares each term in the model
corresponding to the regression
 20% variation in response data is explained by sum of squares each term in the model
corresponding to the error term.
 69% total variation is explained in the model.
 Model is significant because of the f-value that is 4.64 and p-value 0.053

“This question is also solved in excel sheet that is attached with the mail”