ASSIGNMENT 3
Comparing means and binary logistic regression
Introduction
This assignment uses the data collected for a manufacturer of branded chocolate and confectionery
products. The firm hired a market research company mainly to get some insight into the effects that
ambient aromas may have on shopping responses. In particular, they want to know how spraying
chocolate aroma in a supermarket would influence shopping behavior in general, and the purchasing
of their chocolate paste CHOC specifically.
The research company first conducted an experiment in which 125 shoppers were invited to their
lab, which was staged as a basic supermarket. They all received an experimental budget of €25 which
they could freely spend.
Scent was manipulated between subjects: one group of participants completed their shopping in a
lab that was filled with the aroma of chocolate (SCENT=1); the others did their shopping without
being exposed to this ambient chocolate scent (SCENT=0).
Dataset
The key dependent variables of interest were:
‒ SPENDING: total amount (in euros) participants were spending in the simulated supermarket,
‒ PURCHASE: a variable which indicates whether the focal product (CHOC) was bought
(purchase=1) or not (purchase=0).
Additional measures:
‒ GENDER: 1(man) versus 2(woman)
‒ COLD: whether one has a cold (1=yes; 2=no)
‒ HUNGERENTER/ HUNGERLEAVING: One’s self-reported hunger ratings on a 100-point scale
when entering or leaving the lab.
‒ HEALTH: refers to the following item: “I care about my health a lot” (ranging from 1=totally
disagree to 7=totally agree)
‒ CONTROL: refers to the following item: “I am generally very well able to control my
impulses” (ranging from 1=totally disagree to 7= totally agree)
‒ FAMILIAR: refers to the following item: “I often bought the product CHOC in the past”
(ranging from 1=totally disagree to 7= totally agree)
‒ LIST: whether one is using a shopping list or not (0=no; 1=yes)
Questions
1
This study source was downloaded by 100000867004677 from CourseHero.com on 05-10-2023 14:39:51 GMT -05:00
https://www.coursehero.com/file/167757469/assignment-3-KEYpdf/
, 1. Select only those participants who do not have a COLD for further analysis.
“data”, “select”, if, cold=2
2. Check whether GENDER and the use of a shopping LIST are related. State Ho, and conduct the
correct statistical test. Also interpret the “standardized residuals” in the CROSSTABS table, and
visualize your results in a bar chart. Write down your conclusion and provide the test results.
H0: The use of a list and gender are not related.
TEST: Chi-squared test because the two variables are categorical (nominal). Basically, you could
also use a binary logistic regression to test this.
CONCLUSION: The use of a shopping list and gender are not related (X2 = 0.013, p=0.911). The
standardized residuals give us some extra info. For example: In the male category, there are not
significantly more than expected people using a shopping list (z=-0.1 which is lower than 1.96)
RESULTS:
2
This study source was downloaded by 100000867004677 from CourseHero.com on 05-10-2023 14:39:51 GMT -05:00
https://www.coursehero.com/file/167757469/assignment-3-KEYpdf/
Comparing means and binary logistic regression
Introduction
This assignment uses the data collected for a manufacturer of branded chocolate and confectionery
products. The firm hired a market research company mainly to get some insight into the effects that
ambient aromas may have on shopping responses. In particular, they want to know how spraying
chocolate aroma in a supermarket would influence shopping behavior in general, and the purchasing
of their chocolate paste CHOC specifically.
The research company first conducted an experiment in which 125 shoppers were invited to their
lab, which was staged as a basic supermarket. They all received an experimental budget of €25 which
they could freely spend.
Scent was manipulated between subjects: one group of participants completed their shopping in a
lab that was filled with the aroma of chocolate (SCENT=1); the others did their shopping without
being exposed to this ambient chocolate scent (SCENT=0).
Dataset
The key dependent variables of interest were:
‒ SPENDING: total amount (in euros) participants were spending in the simulated supermarket,
‒ PURCHASE: a variable which indicates whether the focal product (CHOC) was bought
(purchase=1) or not (purchase=0).
Additional measures:
‒ GENDER: 1(man) versus 2(woman)
‒ COLD: whether one has a cold (1=yes; 2=no)
‒ HUNGERENTER/ HUNGERLEAVING: One’s self-reported hunger ratings on a 100-point scale
when entering or leaving the lab.
‒ HEALTH: refers to the following item: “I care about my health a lot” (ranging from 1=totally
disagree to 7=totally agree)
‒ CONTROL: refers to the following item: “I am generally very well able to control my
impulses” (ranging from 1=totally disagree to 7= totally agree)
‒ FAMILIAR: refers to the following item: “I often bought the product CHOC in the past”
(ranging from 1=totally disagree to 7= totally agree)
‒ LIST: whether one is using a shopping list or not (0=no; 1=yes)
Questions
1
This study source was downloaded by 100000867004677 from CourseHero.com on 05-10-2023 14:39:51 GMT -05:00
https://www.coursehero.com/file/167757469/assignment-3-KEYpdf/
, 1. Select only those participants who do not have a COLD for further analysis.
“data”, “select”, if, cold=2
2. Check whether GENDER and the use of a shopping LIST are related. State Ho, and conduct the
correct statistical test. Also interpret the “standardized residuals” in the CROSSTABS table, and
visualize your results in a bar chart. Write down your conclusion and provide the test results.
H0: The use of a list and gender are not related.
TEST: Chi-squared test because the two variables are categorical (nominal). Basically, you could
also use a binary logistic regression to test this.
CONCLUSION: The use of a shopping list and gender are not related (X2 = 0.013, p=0.911). The
standardized residuals give us some extra info. For example: In the male category, there are not
significantly more than expected people using a shopping list (z=-0.1 which is lower than 1.96)
RESULTS:
2
This study source was downloaded by 100000867004677 from CourseHero.com on 05-10-2023 14:39:51 GMT -05:00
https://www.coursehero.com/file/167757469/assignment-3-KEYpdf/
