NOTE: YOU CAN EARN 5 POINTS FOR EACH QUESTION, TOGETHER IN TOTAL MAXIMALLY 20
POINTS
QUESTION 1
A marketing agency explores how emotions affect the way people view commercials. For “Dove care
for men”, they created a commercial in which a young man uses the shampoo of his girlfriend. Due to
this shampoo effect, he grows shiny, long hair. The commercial ends with the advice to men to never
use their girlfriend's shampoo but to buy Dove shampoo for men. The agency wants to test whether
the attitude towards the commercial translates in higher purchase intention for men compared to
women.
1a. Formulate the main hypothesis in this research.



1b. Present this model graphically.





1c. What is the mathematical model to test this model?


QUESTION 2
2a. A telecom provider wants to know which Android smartphone sells best. To learn more about the
needs and interest of their customers, they design a survey. (i) Would you use comparative or non-
comparative survey questions? (ii) Give one advantage and one disadvantage of your choice.








2b. Name (i) two types of comparative scales and (ii) two types of non-comparative scales that are
available to a marketer.



2c. The telecom provider has a customer base of 15.000 people in the Netherlands. They want to draw
a representative sample. (i) What sampling method would you advise them, and why? (ii) Describe the
sampling process.



2d. One of the main motives for the survey is to find out how much customers are willing to pay for a
specific Android phone Galaxy S6 Edge. The price in the population is 250$, which you take as the true
population value. You would like to get an estimate that lies within a 10$ of the true population value
with 95% confidence. The standard deviation in the population is 80$. Calculate the sample size (see
formula sheet at the end of the exam).









QUESTION 3
Oxyme was approached by Philips to learn more about the customer's perception of their product
Respironics. Respironics is a mask that you can place on your face which is connected to an oxygen
machine. Oxyme collected all webtexts in which Respironics was mentioned. They tagged the texts in
which a mask is named by 1, and others by 0. Additionally, Oxyme coded all texts on the reported
sentiment which is either negative (-1), neutral (0) or positive (1). Below you find the cross table.
Mask * Sentiment Crosstabulation
Sentiment Total
-1,00 ,00 1,00
Mask
0
Count 130 197 127 454
Expected Count 141,7 172,1 140,2 454,0
% within Mask 28,6% 43,4% 28,0% 100,0%
1
Count 52 24 53 129
Expected Count 40,3 48,9 39,8 129,0
% within Mask 40,3% 18,6% 41,1% 100,0%
Total
Count 182 221 180 583
Expected Count 182,0 221,0 180,0 583,0
% within Mask 31,2% 37,9% 30,9% 100,0%

3a. The chisquare value is 26,263 with 2 degrees of freedom. Which hypothesis is tested by the
chisquare test on this table?



3b. Use the table provided to assess whether this hypothesis should be rejected or not at 95%
confidence interval.





3c. In the online texts, is the mask more often discussed as a) positive, b) negative, c)neutral, or a
combination of a,b, and c?






QUESTION 4
A luxurious clothing company wants to know whether people buy the new expensive clothes to
compensate negative feelings in their live, to uphold social status, or to derive hedonic pleasure. They
designed an online survey to be held among their customers. The customer list includes age, gender
and how long the customer is subscribed to their newsletter.
4a. (i) Would you choose proportionate stratified or systematic sampling? (ii) Describe one advantage
and (iii) one disadvantage for each sampling approach in this case.














4b. The survey included the compulsive buying scale (1 to 6), and the hedonic consumption scale (7 to
12). Interpret the output below. Discuss discriminant and convergent validity.
Pattern Matrix
a

Factor
1 2 3
1. If I have any money left at the end of the pay period, I just
have to spend it.
,824 ,261
2. I felt others would be horrified if they knew of my spending
habits.
,786 ,149
3. I bought things even though I couldn’t afford them. ,353 ,512
4. I wrote a check when I know I didn’t have enough money in the
bank to cover it.
,733
5. I felt anxious or nervous on days I didn’t go shopping. ,795
6. I feel driven to shop and spend, even when I don’t have the
time or the money.
,555 ,322
7. I like to shop for the novelty of it. ,742
8. Shopping satisfies my sense of curiosity. ,642
9. I feel like I’m exploring new worlds when I shop. ,858
10. Shopping offers new experience. ,604
11. I go shopping to watch other people. ,721
12. I go shopping to be entertained. ,431








Factor Correlation Matrix
Factor 1 2 3
1 1,000 -,262 ,215
2 -,262 1,000 -,182
3 ,215 -,182 1,000






4c. The company compares customers who score high on compulsive buying (labeled 1) and those who
score low (labeled 2). (i) Calculate the t-value and (ii) interpret the group differences in the table
below. Use the formula sheet provided at the end of the exam.
Group Statistics

Hedonic
Shopping
N Mean Std. Deviation Std. Error Mean
Purchase intention
1 103 1,46 ,872 ,086
2 101 1,18 ,518 ,052


QUESTION 5

Company GreenBuzz has a sustainable product they want to market in both Japan and the US. They
collect survey data among the customers in these countries about the motivations for environmental
behavior. In the data analysis, they regress environmental behavior (EB) on willingness to pay (WTP)
and environmental concern (EC). Below you can find their results. Note that the variable country is a
nominal variable, 0 identifies the US, and 1 Japan. EC stands for environmental concern and EB for
environmental behavior, both measured at a 5-point likert scale.
Model Summary
Model R R Square
Adjusted R
Square
Std. Error of the
Estimate
1 ,372
a
,139 ,137 ,80082586
ANOVA
a

Model Sum of Squares df Mean Square F Sig.
1 Regression 208,106 3 69,369 108,165 ,000
b

Residual 1292,905 2016 ,641
Total 1501,011 2019
Coefficients
a

Model
Unstandardized Coefficients
Standardized
Coefficients
t Sig. B Std. Error Beta
1 (Constant) -,256 ,024 -10,466 ,000
EC ,318 ,025 ,351 12,603 ,000
Country -,511 ,037 -,292 -13,836 ,000
EC*Country ,142 ,038 ,101 3,693 ,000
Dependent Variable: EB

5a. Which hypothesis is tested by the F values in the output?










5b. GreenBuzz expects that environmental concern (EC) more strongly translates into behavior (EB) in
Japan than in the US as Japan scores higher on collectivism. In collectivist countries, people are more
willing to do something for the greater good than in individualist countries. Are they correct, based on
the output above? Give argumentation based on your interpretation of the coefficients.








QUESTION 6
In the following analysis, the intention to go eating out in restaurants is explained by past behavior of
eating out in restaurants, optimistic feelings on the financial situation of their household, gender (male
are reference group), and a moderation effect of gender and past eating out behavior (eat_gender).
Both intention and past behavior are measured on a 7 point Likert scale ranging from ‘much less so’(1)
to ‘much more so’(7). Optimism is measured on a 7 point scale ranging from ‘not at all’(1) to
‘exceptionally’(7). For the question formulation see table 'descriptive statistics' below.



6a. Interpret the parameters, the fit of the model, and significance levels.






6b. Regression analysis can also be used to predict behavior, attitudes or purchase intentions. What
purchase intention scores does this model predict for a female person who is exceptionally optimistic
(score 7 on optimism) and is used to eating out in restaurants in the past six months (score 5 on past
behavior)?



6c. Below you can find the scatter plot of the standardized residuals and the predicted values. Based
on this scatter plot, do you think some assumptions of linear regression have been violated? Which
one(s)?









QUESTION 7
A travel agency wants to know the psychological benefits of holidays. Based on a survey held among
customers, they expect people who pursue happiness are more likely to go on holidays which will, in
turn, lead to lower levels of loneliness. The output of this analysis is shown below.
shortPI = intention to purchase short holidays
purhapp = pursuit of happiness
lonely = loneliness

**************************************************************************
Model
Y = lonely
X = purhapp
M = shortPI

Sample size
2458
**************************************************************************
Outcome: shortPI

Model Summary
R R-sq MSE F df1 df2 p
,1556 ,0242 ,9802 60,9152 1,0000 2456,0000 ,0000

Model
coeff se t p LLCI ULCI
constant ,0302 ,0200 1,5087 ,1315 -,0091 ,0694
purhapp -,1723 ,0221 -7,8048 ,0000 -,2156 -,1290

**************************************************************************
Outcome: lonely

Model Summary
R R-sq MSE F df1 df2 p
,2428 ,0589 ,7436 76,8930 2,0000 2455,0000 ,0000

Model
coeff se t p LLCI ULCI
constant -,0158 ,0174 -,9037 ,3662 -,0500 ,0184
shortPI ,1228 ,0176 6,9848 ,0000 ,0883 ,1572
purhapp -,1759 ,0195 -9,0355 ,0000 -,2140 -,1377

******************** DIRECT AND INDIRECT EFFECTS *************************

Direct effect of X on Y
Effect SE t p LLCI ULCI
-,1759 ,0195 -9,0355 ,0000 -,2140 -,1377

Indirect effect of X on Y
Effect Boot SE BootLLCI BootULCI
shortPI -,0212 ,0042 -,0309 -,0140

**************************************************************************


7a. Draw the conceptual model that is estimated below. Place the coefficients next to the arrows in
the model.





7b. Calculate the indirect effect that is noted in the output above, and discuss whether this is
significant based on a 95% confidence interval.









Formula sheet
2 2
2
z
n
D

 ,
/
X
t
s n

 ,
1
c
nN
n
N n

 
, 1 2
2 2
1 1 2 2/ /
X X
t
s n s n



,
2
2
(1 )z
n
D
 
 ,
(1 )
1
p
N n
n N
 

 


,
X
n

  ,
(1 )
p
p p
S
n

 ,
2 2
2
C z
n
R
 , D R ,
X
X
z



 ,
1X
N n
Nn





, D R , 1
/
D
n
D
D
t
s n



 , 1 2 1 2
2 2
1 1 2 2
( ) ( )
/ /
X X
t
n n
 
 
  


, 1 2 1 2
2 2
1 1 2 2
( ) ( )
/ /
X X
t
s n s n
   



2n
2
i 1
(obs exp )
exp
i i
i
X


 , 2 ( 1) ( 1)Xdf c r    ,
/
/
regression regression
residual residual
SSE df
F ratio
SSE df
  ,
2 explained
total
SSE
R
SSE

critical z-value 95% confidence interval (two-tailed) is 1.96
critical z-value 99% confidence interval (two-tailed) is 2.58
critical z-value 95% confidence interval (one-tailed) is 1.65
critical z-value 99% confidence interval (one-tailed) is 2.33