prof. dr. rené algesheimer marketing analytics iia5ac08ee-f115-450b-9a9... · data analysis. this...

Chair of Marketing and Market ResearchDepartment of Business AdministrationUniversität Zürich, Switzerland © Zürich, 2016/2017. All rights reserved.

SyllabusEach Fall SemesterLast edit: 05.09.2016

Marketing analytics iiProf. Dr. René Algesheimer

Marketing Analytics II - Syllabus - 1

Preamble

Welcome to my “Marketing Analytics” syllabus!

“I learnt very early on the difference between knowing the name of something and knowing something.”

Richard Feynman

This course aims to deepen student’s knowledge about actual research problems in marketing and consumer research and to support student’s development into a well-informed practitioner of state-of-the-art market research. A practitioner is capable to formulate and structure market research problems, to collect and to analyze quantita-tive market research data, and finally to infer effective marketing decisions based on data analysis. This interactive course enables students to design and conduct market research analyses using a variety of confirmatory multivariate tools like regression, variance analysis, discriminance analysis, structural equations modelling, but also exploratory multivariate tools like factor analysis, cluster analysis, or multi-dimensional scaling. Depending on the time and the interest of the students this course will also cover topics like conjoint analysis, social network analysis or data mining methods. These skills are also very useful for students that plan to go into a consulting or marketing career.

This course will always take place in the fall semesters and it is the follow-up of the course Marketing Analytics I. Nevertheless, students can also start with this course if they have a profound statistical knowledge. You’ll find all the necessary information concerning the course within this syllabus. From time to time, updates will be posted on our website, at the “Marketing” blackboard at Andreasstrasse 15, 4th floor, and on our eLearning platform.

www.market-research.uzh.ch.

I am pleased to welcome you to my course.

Enjoy this introduction.

All the best,

René Algesheimer


Quick Overview:

Instructor:

Prof. Dr. René Algesheimer,

Office: Andreasstrasse 14, CH-8050 Zürich, Switzerland

Phone: +41 44 634 2918

E-mail: [email protected]

Office hours are by appointment.

Web: www.market-research.uzh.ch

Teaching Assistants:

Raluca Gui, Xin-Yu Zou

Target Audience:

This course is reckonable for MA and is assigned to the „Wahlpflichtbereich” BWL 4.

Frequency:

Each fall semester

AP (ECTS):

6

Work load statement:

Part Workload Total Time ECTS

Course attendance 15 lectures à 90min, 2 weeks 22.5h

Exercise attendance 15 exercises à 90min, 2 weeks 22.5h

L&E preparation 14h per week, 2 weeks 28h

Literature study Preparation before class 71h

Group assignments Workshop plus preparation 36h

Total 180h 6

Maximum Amount of Students:

limited only by room size

Content:

Practical introduction into understanding, applying, interpreting and documenting quantitative market research methods by using R 3.3.

Language:

English


Basic Literature:

Field, Andy (AF): Discovering Statistics Using R, 1st ed., London et al.: Sage, 2012.

Hair, Joseph F. Jr.; Black, William C.; Babin, Barry J. & Anderson, Rolph E. (HBBA): Multivariate Data Analysis. A Global Perspective, 7th ed., Upper Saddle River et al.: Pearson, 2010.

James, Gareth (GJ) et al.: An Introduction to Statistical Learning with Applications in R, Springer, 2013.

Additional literature will be given in-class.

Prerequisite:

Recommended: Marketing Analytics I, Statistics, Empirical Research Methods.

Access:

Join our courses and make up your mind if you want to participate. In the positive case, register yourself and sign up for the courses, you want to participate, at our chair. Then officially register yourself using the booking tools at the University of Zurich.

Grading:

Participation, exercises, group work & presentations, code-writing..

Dates:

Block course, September 5th -September 16th, 9.30-17.30h each day and homework.

Location:

Lectures: AND-5-29/31

Further information:

º www.market-research.uzh.ch

º blackboard Marketing, Andreasstreasse 15, 4th floor

Note:

This information in the syllabus supports the official information in the electronic university calendar (VVZ – Vorlesungsverzeichnis). In cases of doubt, the official information at the VVZ is valid.


1. INTRODUCTION AND OBJECTIVE

“I checked it very thoroughly,“ said the computer, “and that quite definitely is the answer.

I think the problem, to be quite honest with you, is that you’ve never actually know what the question is.“

D. Adams, The Hitchhiker’s Guide to the Galaxy

Course Purpose & Objectives

At the heart of superior marketing practice there is always a decision. One, for example, has to decide how to price a product, what kind of distribution channels one wants to use, or how to advertise a specific product. In order to reduce complexity and support one alternative from a multitude, quantitative marketing methods are essential in organizations. Thus, gaining a thorough understanding of instruments that can be implemented and applied to a diversity of marketing settings is the purpose of this course and of the market research course trilogy. The objective of this course is to become accustomed with, understand and apply quantitative marketing methods that are typically used in marketing management.

The reasons for this are twofold: First, many students are afraid of maths and statis-tics. Nevertheless, mathematics is the ultimate language (besides music) and I believe it is an essential building block in the development of many career paths. Students can fear these subjects because no one has explained them in a language that may be understood and enjoyed, in a way that allows the student to really fall in love with numbers and apply them to different settings. When I was a student, statistics was something about axioms, definitions, propositions, evidence, and occasionally it was also about software programming. I was often interested in WHY I should apply this statistical test and HOW it works, what the outputs mean and what we can conclude from the results. For me, statistics has always been a way to support higher order real life situations or managerial decisions with the advantage of offering practical insights and recommendations. Therefore, our market research courses should help students to solidify and ground the knowledge gained in basic marketing courses and also help them to develop effective marketing thinking on their own. This is done by exposing students to a variety of old, new, and sometimes unusual, instruments of quantitative marketing methods. Furthermore, the course will motivate and encourage students to practice these concepts in practical exercises as well as exposing them to the neces-sary software packages while simultaneously developing a spirit of problem solving. I understand my courses as “hands-on” and so should you.

Second, I believe that quantitative thinking is supportive to logical and structured business thinking, a capacity that is very often necessary in organizations. Therefore, I hope that the classes will not only enhance your quantitative knowledge, but also your ability to think in business terms.

This course should (a) sensitize students to typical quantitative marketing problems; (b) introduce students to quantitative marketing methods that are typically used in marketing management; (c) develop students’ abilities to identify, apply and evaluate these methods; (d) develop students’ skills in gathering information, drawing conclu-sions from it, and presenting the material, and (e) develop student’s hands-on compe-tence in quantitative marketing research methods.


Course Contribution towards Marketing Management

The course includes a comprehensive presentation of the main quantitative methods that are typically used in marketing management. These elements are discussed in class and supported by relevant examples, taken either from specialized academic and professional literature, or from the personal experience of the teaching staff. The approach adopted encourages students to critically evaluate given marketing situa-tions, and methods to question and discuss their applicability as well as to solve given marketing decision problems.

Course Contribution towards Analytical Competence

The course presents the main quantitative marketing instruments that are applied in the professional world, and which help marketing managers to analyze marketing situations, to formulate marketing strategies and plans, and to evaluate their impact. The student’s understanding of these analytical instruments, taught to them from basics, is realized through theoretical discussions, examples, exercises, and practical assignments. While many books separate different methods and tests, the approach in this course is to build a unique perspective that draws similarities across several statistical methods and tests.

Course Contribution towards Correctly Understanding and Applying Marketing Instruments

One course objective is to show how analytical marketing instruments can support marketing decisions. The quantitative methods presented and discussed in class will be instruments providing students with an image of the complexity and pitfalls of typi-cal marketing problems. These instruments have to be correctly applied by students in order to successfully solve their assignments and to answer the questions included in the final exam.

Course Contribution towards Critical Thinking, and Problem Solving Skills

As all instruments are directly applied to realistic marketing situations, students need to formulate the related marketing problem and marketing questions to these given situations. Problem solving skills are developed as a consequence of applying quanti-tative methods, and alternatives are also discussed in class. The results of quantitative marketing methods are interpreted and critically analyzed in order to foster critical thinking.

Course Contribution towards Ethical and Social Responsibility

The cases that are presented in class integrate ethical questions in order to develop a sense of ethical and social responsibility and to actively generate an understanding of different cultural perspectives. An open minded, tolerant, and respectful atmosphere within class is necessary to maintain this. The pedagogical approach adopted in this course encourages students to participate contributing their opinions, experience, and comments to the discussions developed around the presented marketing meth-ods and to seriously consider and discuss other’s opinions.


Course Contribution towards the Development of Good Teamwork and Communication Skills (depending on group size)

The capability to effectively work in teams and to communicate during the work-ing process is an essential skill for modern marketing managers. The pedagogical approach adopted in this course encourages students to participate forwarding their opinions, experience, and comments to the discussions developed around the presented marketing methods. The in class exercises are also conducted in groups on our workshop days so that the course encourages students to develop interpersonal communication skills, as well as to debate and negotiate ideas and decisions during their group work. Finally, students are obliged to use both verbal and written commu-nication during their course work and evaluation, which reinforces these skills.

Course Description

The course presents popular quantitative marketing methods with practical exercises to familiarize students both with the theoretical and practical aspects of marketing methods.

2. COURSE MATERIAL

Students have access to our web-based e-learning platform on OLAT to download the slides presented in class, participate in self-learning modules, find relevant mate-rial, datasets and literature, discuss with your classmates the latest topic in class and much more thus benefiting from complementary information available online and in the library.

A system of different learning abilities has been developed. The following procedure is strongly recommended as preparation for the classes.

Overview of classes

On the webpage an overview of all classes given by our team can be found. Develop an idea of the classes and how they best fit into your personal agenda. Keep in mind that quantitative marketing research classes are only offered once a year. It is also necessary to have successfully completed the prior course to proceed with the fol-lowing.

Hands-on guides

Several files have been prepared that should provide you with background knowledge of my expectations in the classroom and some tips concerning “How to give presenta-tions in class”, “How to write in an academic style”… If you read them prior to class, then you’ll obtain a good understanding of what is expected from you.

Syllabus

For each course, a detailed syllabus exists with all details concerning that specific course. This is your guideline for the class and a MUST read. You’ll find everything in here concerning the grading of the course, the agenda, the planned topics, the work-load, readings and much more…


The main materials used in this course are:

The Slides

The slides presented and discussed in class are available in a digital format, on the e-learning platform. You can download the slides to each class. The slides don’t com-pletely cover the entire syllabus, therefore it is necessary to participate in the class. All slides will be uploaded after each module and contain lecture notes as well.

All our slides follow our detailed standardized slide format. All presentations in the classroom also have to follow this format.

The Reading List

The reading list is split into three categories depending on your time and involvement in the class. REQUIRED readings are necessary readings before each class and pre-pare you for the actual content. RECOMMENDED readings are articles that go into more details and widen your knowledge. FOLLOW-UP readings will help you to draw together your newly acquired knowledge of the content or solve some troubles if you are in the middle of your own practical work. EXEMPLARY articles apply the learned knowledge within different marketing areas and allow you to establish utilization of the learned methods.

The Exercises and workshops

For each class you’ll find about 10 multiple choice questions on the eLearning plat-form that you can solve on your own. Furthermore, you’ll receive a bunch of practical lessons combined with exercises that you will solve with my TAs during practical ses-sions. On the workshop days, group-work is given to you to practise your skills!

Additional Materials

The academic and professional papers published online or in marketing journals can also be used by students to obtain additional information about marketing concepts, theories and methods. The following journals are reputable and are therefore strongly recommended to the students:

Marketing journals:

º Marketing Science

º Journal of Marketing Research

º Journal of Marketing

º Journal of Consumer Research

º Quantitative Marketing and Economics

º International Journal of Research in Marketing

º Journal of the Academy of Marketing Science

º Journal of Interactive Marketing

º Journal of Service Research

º Journal of Product and Innovation Management

º Harvard Business Review


º Sloan Management Review

º McKinsey Quarterly

3. COURSE CONTENTS

The course will potentially cover the following topics: introduction into quantitative market research, definitions of basic concepts, introduction into R 3.11 and Gephi 0.8.2, exploring data with graphs, key assumptions in quantitative research, correla-tion and causation, descriptive statistics. You will find a detailed schedule below.

Required readings in class:

Field, Andy (2012): Discovering Statistics Using R, 1st ed., London et al.: Sage. [AF]

Hair, Joseph F. Jr.; Black, William C.; Babin, Barry J. & Anderson, Rolph E. (2010): Multivariate Data Analysis. A Global Perspective, 7th ed., Upper Saddle River et al.: Pearson.[HBBA]

Recommended readings in class:

Hanssens, Dominique M., Parsosns, Leonard J. & Schultz, Randall L. (2003): Market Response Models: Econometric and Time Series Analysis, 2nd ed., Boston et al.: Kluwer Academic Publishers.

Iacobucci, Dawn (2012): Marketing Models, Cengage Learning. [I]

Lilien, Gary L. & Rangaswamy, Arvind (2006): Marketing Engineering, revised 2nd ed., Victoria: Trafford Publishing.

Leeflang, Peter S.H., Wittink, Dick R., Wedel, Michel & Naert, Philippe A. (2000): Building Models for Marketing Decisions, Boston et al.: Kluwer Academic Publish-ers.

Weiss, Neil A. (2007): Introductory Statistics, 8th ed., Boston et al.: Addison Wesley.

Wierenga, Berend (2008): Handbook of Marketing Decision Models, New York et al.; Springer.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 9

Plan

ned

cour

se s

ched

ule[1]

:

Day

Topi

cTe

xt C

hapt

ers

Read

ings

1 (9

h30

- 11h

00)

Intr

oduc

tion

to M

ulti

vari

ate

Met

hods

(mul

tivar

iate

re

gres

sion

, an

ova,

di

scrim

inan

t an

alys

is,

expl

orat

ory

and

confi

rm-a

tory

fac

tor

anal

ysis

, lo

gist

ic

regr

essi

on,

conj

oint

mea

sure

men

t, st

ruct

ural

equ

atio

n m

odel

ing,

clu

ster

ana

lysi

s, s

ocia

l ne

twor

k an

alys

is a

nd

mac

hine

lear

ning

).

Exer

cise

: ¹

Intr

oduc

tion

to th

e st

atis

tical

pro

gram

R 3

.3

Dat

aset

s:

¹H

BAT

.csv

Requ

ired

read

ing

¹A

F, p

p. 1-

61.

¹H

BB

A, p

p. 1-

30.

Reco

mm

ende

d re

adin

g ¹

AF,

pp.

63-

115

(esp

ecia

lly fo

r tho

se, w

ho d

id n

ot v

isit

Qua

ntita

tive

Mar

ket R

esea

rch

I). ¹

Page

, Sco

tt (2

011)

, “Re

gres

sing

you

r inn

er h

edge

hog,

” (ac

cess

ed A

ugus

t 12t

h, 2

014

),

[ava

il-ab

le a

t htt

p://

ww

w.c

scs.

umic

h.ed

u/~s

page

/tea

chin

g_fil

es/m

odel

ing_

lect

ures

/MO

DEL

5/Le

c1.p

df].

Follo

w-u

p re

adin

g ¹

Coh

en, J

. (19

90),

“Thi

ngs

I hav

e le

arne

d (s

o fa

r),”

Am

eric

an P

sych

olog

ist,

45(12

), 13

04–

1312

. ¹

Coh

en, J

. (19

94),

“The

ear

th is

roun

d (p

< .0

5),”

Am

eric

an P

sych

olog

ist,

49 (1

2), 9

97-1

003

. ¹

Coh

en,

J. (

200

9),

“Mis

sion

Im

prob

able

: A

Con

cise

and

Pre

cise

Defi

nitio

n of

p-v

alue

,” Sc

ienc

eNow

Dai

ly, (

acce

ssed

Oct

ober

30

th, 2

00

9), [

avai

labl

e at

htt

p://

scie

ncen

ow.s

ci-

ence

mag

.org

/cgi

/con

tent

/ful

l/20

09/

1030

/1].

¹N

uzzo

, R. (

2014

), “S

tatis

tical

err

ors,

” Nat

ure,

vol

. 50

6 (F

ebru

ary

13th

), 15

0-1

52.

¹Ke

inin

gham

, T. L

., B

. Coo

il, T

.W. A

ndre

asse

n, a

nd L

. Aks

oy (2

00

7), “

A L

ongi

tudi

nal E

xam

ina-

tion

of N

et P

rom

oter

and

Firm

Rev

enue

Gro

wth

,”, J

ourn

al o

f Mar

ketin

g, 7

1 (Ju

ly),

39-5

1. Ex

empl

ary

read

ing

¹Re

ichh

eld,

Fre

deric

k F.

(20

03),

“The

onl

y nu

mbe

r you

nee

d to

gro

w,”

Har

vard

Bus

ines

s Re

-vi

ew, D

ecem

ber,

Repr

int:

R03

12C

.

1(14

h00

-15h3

0)M

ulti

vari

ate

regr

essi

on (O

LS)

(ord

inar

y le

ast s

quar

es (o

ls),

tota

l sum

of s

quar

es, m

odel

su

m o

f sq

uare

s, r

esid

ual s

um o

f sq

uare

s, m

ean

sum

of

squa

res,

roo

t m

ean

squa

re e

rror

, r2,

adj

uste

d r2

, f-t

est,

t-te

st, l

inea

rity,

hom

osce

dast

icity

, aut

ocor

rela

tion,

mul

ti-co

lline

arity

, Coo

k’s

dist

ance

, lev

erag

e, m

ahal

anob

is d

2,

dffit,

dfbet

a)

Exer

cise

: ¹

R ex

ampl

e of

reg

ress

ion

diag

nost

ics

in p

redi

ctin

g sa

les

of m

usic

alb

ums.

Dat

aset

s:

albu

m_s

ales

_dat

aset

.csv

Requ

ired

read

ing

¹A

F, p

p. 2

45-2

66.

Reco

mm

ende

d re

adin

g ¹

Gar

son,

G. D

avid

(20

14),

“Mul

tiple

reg

ress

ion”

, Sta

tistic

al A

ssoc

iate

s Pu

blis

hing

, [av

aila

ble

at h

ttp:

//w

ww

.sta

tistic

alas

soci

ates

.com

/boo

klis

t.htm

], la

st a

cces

sed:

Aug

ust 1

2th,

20

14.

¹H

eckm

an,

J. (

1979

), “S

ampl

e Se

lect

ion

Bia

s as

a S

peci

ficat

ion

Erro

r,” E

cono

met

rica,

47,

15

3–61

. ¹

Oze

r-B

alli,

H. a

nd B

.E. S

oren

sen

(20

10),

“Int

erac

tion

effec

ts in

Eco

nom

etric

s,” W

orki

ng P

a-pe

r, U

nive

rsity

of M

asse

y an

d U

nive

rsity

of H

oust

on.

Follo

w-u

p re

adin

g ¹

Alg

eshe

imer

, R. S

. Bor

le, U

.M. D

hola

kia,

and

S. S

iddh

arth

(20

10),

“The

Impa

ct o

f Cus

tom

er

Com

mun

ity P

artic

ipat

ion

on C

usto

mer

Beh

avio

rs: A

n Em

piric

al In

vest

igat

ion,

” Mar

ketin

g Sc

ienc

e, 2

9 (4

), 75

6-76

9.

1. A

rtic

les

are

liste

d in

the

orde

r of p

ublic

atio

n ye

ar, f

ollo

wed

by

book

s. T

he o

nly

exce

ptio

ns a

re re

adin

gs in

the

AF

or H

BB

A b

ooks

that

are

alw

ays

on to

p.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 10

Day

Topi

cTe

xt C

hapt

ers

Read

ings

2(9

h30

- 11h

00)

OLS

ass

umpt

ions

(line

arity

, ho

mos

ceda

stic

ity,

auto

corr

elat

ion,

mul

ticol

-lin

earit

y).

Exer

cise

:

¹Te

stin

g as

umpt

ions

.

Dat

aset

s:

¹al

bum

_sal

es_d

atas

et.c

sv

Requ

ired

read

ing

¹A

F, p

p. 2

66-3

11.

Reco

mm

ende

d re

adin

g ¹

Gar

son,

G. D

avid

(20

14),

“Mul

tiple

reg

ress

ion”

, Sta

tistic

al A

ssoc

iate

s Pu

blis

hing

, [av

aila

ble

at h

ttp:

//w

ww

.sta

tistic

alas

soci

ates

.com

/boo

klis

t.htm

], la

st a

cces

sed:

Aug

ust 1

2th,

20

14.

¹So

yer,

E.,

and

Hog

arth

, R. M

. (20

12),

“The

illu

sion

of p

redi

ctab

ility

: How

regr

essi

on s

tatis

tics

mis

lead

exp

erts

,” In

tern

atio

nal J

ourn

al o

f For

ecas

ting,

Jul

y-Se

ptem

ber,

28 (3

), 69

5-71

1. ¹

Ritt

er, D

. (20

14),

“Whe

n to

act

on

corr

elat

ion,

and

whe

n no

t,” H

arva

rd B

usin

ess

Revi

ew,

19.3

.20

14, r

ecei

ved

from

: ww

w.b

logs

.hbr

.com

, las

t acc

esse

d: A

ugus

t 12t

h, 2

014

.

Follo

w-u

p re

adin

g ¹

Aks

oy, L

erza

n et

al.

(20

08)

, “Th

e Lo

ng T

erm

Sto

ck M

arke

t Val

uatio

n of

Cus

tom

er S

atis

fac-

tion,

” Jou

rnal

of M

arke

ting,

72,

105-

122.

¹B

radl

ey E

. (19

79),

“Boo

tstr

ap M

etho

ds: A

noth

er L

ook

at t

he J

ackk

nife

,” Th

e A

nnal

s of

Sta

-tis

tics,

7, 1,

1979

, 1–2

6. ¹

Litt

le, R

. J. A

. and

D.B

. Rub

in (2

00

2), S

tatis

tical

Ana

lysi

s w

ith M

issi

ng D

ata,

2nd

ed.

, Hob

o-ke

n: J

ohn

Wile

y &

Son

s. ¹

Wilc

ox, R

. R. (

200

5), I

ntro

duct

ion

to R

obus

t Es

timat

ion

and

Hyp

othe

sis

Test

ing,

2nd

ed.

, B

urlin

gton

, MA

: Els

evie

r.

Exem

plar

y re

adin

g ¹

The

Am

eric

an C

usto

mer

Sat

isfa

ctio

n In

dex,

htt

p://

ww

w.th

eacs

i.org

/.

2(14

h00

-15h3

0)O

LS c

halle

nges

(mul

ticol

linea

rity,

het

eros

keda

stic

ity, o

utlie

rs, e

rror

s in

va

riabl

es,

endo

gene

ity,

self-

sele

ctio

n, g

raph

ical

pro

fil-

ing,

uni

varia

te p

rofil

ing

with

his

togr

ams,

biv

aria

te p

rofil

-in

g w

ith b

oxpl

ots

and

scatt

erpl

ots)

Exer

cise

:

¹O

utlie

r and

mul

ticol

linea

rity

dete

ctio

n, a

sses

sing

im-

pact

of o

mitt

ed v

aria

ble

bias

.

Dat

aset

s:

¹al

bum

_sal

es_d

atas

et.c

sv

Requ

ired

read

ing

¹A

F, p

p. 2

66-3

11.

¹W

oold

ridge

J.M

. Int

rodu

ctor

y Ec

onom

etric

s, (2

00

9),

Sout

h-W

est-

ernC

enga

ge L

earn

ing,

4th

edi

tion,

C

h.3,

8, 1

5, (h

ttp:

//nc

bae.

yola

site

.co

m/r

esou

rces

/Int

rodu

ctor

yEco

no-

met

rics_

AM

oder

nApp

roac

h_Fo

ur-

thEd

ition

_Jeff

rey_

Woo

ldrid

ge.p

df)

Reco

mm

ende

d re

adin

g ¹

Ang

rist,

J. a

nd P

isch

ke, J

-S, (

200

8), M

ostly

Har

mle

ss E

cono

met

rics:

An

Empi

ricis

tís C

om-

pani

on, h

ttp:

//w

ww

.dev

elop

men

t.wne

.uw

.edu

.pl/

uplo

ads/

Mai

n/re

crut

_eco

nom

etric

s.pd

f.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 11

Day

Topi

cTe

xt C

hapt

ers

Read

ings

3(9

h30

- 11h

00)

Ana

lysi

s of

Var

ianc

e (A

NO

VA)

(gen

eral

lin

ear

mod

els

(glm

), t-

test

s, i

ndep

en-

dent

t-t

est,

depe

nden

t t-

test

, an

ova,

Hot

ellin

g t2

, f-t

est,

Bro

wn-

Fors

ythe

f-te

st, W

elch

s f-

test

, pl

anne

d co

mpa

rison

, pl

anne

d co

ntra

st,

post

-ho

c te

sts,

tre

nd a

naly

sis,

gra

nd m

ean,

gro

up

mea

n, w

ithin

-gro

up v

aria

nce,

bet

wee

n-gr

oups

va

rianc

e, e

ta s

quar

ed, o

meg

a sq

uare

d).

Exer

cise

: ¹

Lear

n ho

w to

con

duct

GLM

in R

, und

erst

and

GLM

with

sev

eral

pre

dict

ors,

und

erst

and

how

to a

sses

s th

e fit

of a

GLM

mod

el a

nd

inte

rpre

t the

resu

lts.

Dat

aset

s:

¹Ic

eCre

am.c

sv, P

urch

ase.

csv

Requ

ired

read

ing

¹A

F, p

p. 3

59-6

52.

¹A

F, p

p. 6

96-7

48.

¹H

BB

A, p

p. 4

39-4

47.

Reco

mm

ende

d re

adin

g ¹

Gar

son,

G. D

avid

(20

14),

“Gen

eral

Lin

ear M

odel

s”, S

tatis

tical

Ass

ocia

tes P

ublis

hing

, [av

aila

ble

at

http:

//w

ww

.sta

tistic

alas

soci

ates

.com

/boo

klis

t.htm

], la

st a

cces

sed:

Aug

ust 1

2th,

20

14.

¹H

BB

A, p

p. 4

39-4

76.

Follo

w-u

p re

adin

g ¹

AF,

pp.

457

-50

5 (R

epea

ted

Mea

sure

s).

¹A

F, p

p. 5

06-

538

(Mix

ed D

esig

ns).

¹Ke

nny,

D. A

. and

C.M

. Jud

d (19

86),

“Con

sequ

ence

s of v

iola

ting

the

inde

pend

ence

ass

umpt

ion

in

anal

ysis

of v

aria

nce,

” Psy

chol

ogic

al B

ulle

tin, 9

9(3)

, 422

-431

. ¹

Oze

r-B

alli,

H. a

nd B

.E. S

oren

sen

(20

10),

“Int

erac

tion

effec

ts in

Eco

nom

etric

s,”

Wor

king

Pap

er,

Uni

vers

ity o

f Mas

sey

and

Uni

vers

ity o

f Hou

ston

.Ex

empl

ary

read

ing

¹U

rban

y, J

. E.;

Bea

rden

, W. O

.; W

eilb

aker

, D. C

. (19

88):

The

Effec

t of P

laus

ible

and

Exa

gger

ated

Re

fere

nce

Pric

es o

n C

onsu

mer

Per

cept

ions

and

Pric

e Se

arch

, Jou

rnal

of C

onsu

mer

Res

earc

h,

15(1)

, 95–

110

.

3(14

h00

-15h3

0)Ex

plor

ator

y Fa

ctor

Ana

lysi

s (E

FA)

(exp

lora

tory

fa

ctor

an

alys

is,

alph

a fa

ctor

ing,

A

nder

son-

Rubi

n m

etho

d,

com

mon

va

rianc

e,

com

mun

ality

, Cro

nbac

h’s

Alp

ha, d

irect

obl

imin

, fa

ctor

load

ing,

fact

or m

atrix

, fac

tor s

core

s, in

tra-

clas

s co

rrel

atio

n co

effici

ent

(icc)

, Kai

ser-

Mey

er-

Olk

in (k

mo)

mea

sure

of s

ampl

ing

adeq

uacy

, Kai

-se

r cr

iterio

n, l

aten

t va

riabl

e, o

bliq

ue r

otat

ion,

pr

inci

pal

com

pone

nt

anal

ysis

(p

ca),

rand

om

varia

nce,

rota

tion,

scr

ee p

lot,

sing

ular

ity, u

niqu

e va

rianc

e, V

arim

ax).

Exer

cise

: ¹

Und

erst

and

EFA

mod

els,

how

to d

o EF

A in

R

, int

erpr

eta

EFA

mod

els.

Dat

aset

s:

¹EF

A_W

illia

ms.

csv

Requ

ired

read

ing

¹A

F, p

p. 7

49-8

11.

¹H

BB

A, 2

010

, pp.

50

5-56

4.

Reco

mm

ende

d re

adin

g ¹

Gar

son,

G. D

avid

(20

14).

“Fac

tor A

naly

sis”

, Sta

tistic

al A

ssoc

iate

s Pu

blis

hing

, [av

aila

ble

at h

ttp:

//w

ww

.sta

tistic

alas

soci

ates

.com

/boo

klis

t.htm

], la

st a

cces

sed:

Aug

ust 1

2th,

20

14.

¹H

BB

A, p

p. 9

1-15

2.

Follo

w-u

p re

adin

g ¹

Chu

rchi

ll, G

. A. (

1979

): A

Par

adig

m f

or D

evel

opin

g B

etter

Mea

sure

s of

Mar

ketin

g C

onst

ruct

s,

Jour

nal o

f Mar

ketin

g Re

sear

ch, 1

6 (F

ebru

ary)

, 64-

73.

¹C

ortin

a, J

. M. (

1993

), “W

hat i

s C

oeffi

cien

t Alp

ha?

An

Exam

inat

ion

of T

heor

y an

d A

pplic

atio

ns,”

Jour

nal o

f App

lied

Psyc

holo

gy, 7

8, 9

8-10

4. ¹

Ger

bing

, D.W

. and

J.C

. And

erso

n (19

88),

“An

Upd

ated

Par

adig

m fo

r Sc

ale

Dev

elop

men

t Inc

or-

pora

ting

Uni

dim

ensi

onal

ity a

nd I

ts A

sses

smen

t,” J

ourn

al o

f M

arke

ting

Rese

arch

, 25

(May

), 18

6-19

2. ¹

Gor

such

, R. L

. (19

90),

“Com

mon

Fac

tor

Ana

lysi

s Ve

rsus

Com

pone

nt A

naly

sis:

Som

e W

ell a

nd

Litt

le K

now

n Fa

cts,

” Mul

tivar

iate

Beh

avio

ral R

esea

rch,

25,

33-

39.

¹St

ewar

t, D

. W. (

1981

), “T

he A

pplic

atio

n an

d M

isap

plic

atio

n of

Fac

tor

Ana

lysi

s in

Mar

ketin

g Re

-se

arch

,” Jo

urna

l of M

arke

ting

Rese

arch

, 18

(Feb

ruar

y), 5

1-62

.

Exem

plar

y re

adin

g ¹

Cav

usgi

l, S.

T. a

nd S

. Zou

(199

4), “

Mar

ketin

g St

rate

gy-P

efor

man

ce R

elat

ions

hip:

An

Inve

stig

atio

n of

the

Empi

rical

Lin

k in

Exp

ort M

arke

t Ven

ture

s,” J

ourn

al o

f Mar

ketin

g, 5

8 (J

anua

ry),

1-21

. ¹

Dol

l, W

. J.,

W. X

ia, a

nd G

. Tor

kzad

eh (1

994)

, “A

Con

firm

ator

y Fa

ctor

Ana

lysi

s of

the

End

-Use

r C

ompu

ting

Satis

fact

ion

Inst

rum

ent,

MIS

Qua

rter

ly, 1

(2),

453-

461.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 12

Day

Topi

cTe

xt C

hapt

ers

Read

ings

4(9

h30

-11h3

0)C

onfir

mat

ory

Fact

or A

naly

sis

(CFA

)(la

tent

var

iabl

e, o

bser

ved

varia

ble,

fac

tor

load

ings

, pa

th c

oeffi

cien

t, fa

ctor

cor

rela

tions

, fixe

dpa

ram

eter

s, c

onst

rain

ed p

aram

eter

s, fr

eepa

ram

eter

s, r

ando

m m

easu

rem

ent

erro

r, no

nrad

om

mea

sure

men

t err

or, f

orm

ativ

e in

dica

tors

,re

flect

ive

indi

cato

rs,

estim

ator

s, m

odel

fit

indi

ces,

pa

ram

eter

crit

eria

(co

nstr

uct

valid

ity,

conv

erge

nt

valid

ity, d

iscr

imin

ant

valid

ity, n

omol

ogic

al a

nd f

ace

valid

ity, s

tand

ardi

zed

resi

dual

s, r

esid

ual

varia

nces

, m

odifi

catio

n in

dice

s).

Exer

cise

: ¹

R ex

ampl

e of

ope

ratio

naliz

ing

Prod

uct

¹Im

age

and

Prod

uct L

oyal

ityD

atas

ets:

¹

HB

AT_S

EM.c

sv

Requ

ired

read

ing

¹G

arso

n,

G.

Dav

id

(20

14).

“Con

firm

ator

y Fa

ctor

A

naly

sis”

, St

atis

tical

A

ssoc

iate

s Pu

blis

hing

, [a

vaila

ble

at h

ttp:

//w

ww

.sta

-tis

tical

asso

ciat

es.c

om/b

ookl

ist.h

tm],

last

ac

cess

ed: A

ugus

t 12t

h, 2

014

.

Reco

mm

ende

d re

adin

g ¹

Mile

s, J

. N. V

. (20

05)

, “C

onfir

mat

ory

fact

or a

naly

sis

usin

g M

icro

soft

Exce

l,” B

ehav

ior

rese

arch

m

etho

ds, 3

7(4)

, 672

-676

. ¹

Bro

wn,

Tim

othy

A.

(20

06)

, C

onfir

mat

ory

Fact

orA

naly

sis

for

App

lied

Rese

arch

. Lo

ndon

: Th

eGui

lford

Pre

ss, c

hapt

er 3

and

4.

¹K

line,

Rex

B. (

2010

), Pr

inci

ples

and

Pra

ctic

e of

Stru

ctur

al E

quat

ion

Mod

elin

g. N

ew Y

ork:

Th

eGui

lford

Pre

ss, c

hapt

er 5

and

9.

¹M

uthé

n, L

inda

K. a

nd B

engt

O. M

uthé

n (2

010

), M

plus

Use

r’s G

uide

(6

ed.).

Los

Ang

eles

: M

uthé

n &

Mut

hén,

cha

pter

5, 1

5, 16

, 17

and

18.

Follo

w-u

p re

adin

g ¹

Hu,

Li-t

ze a

nd P

eter

M. B

entle

r (19

98),

“Fit

Indi

ces

in C

ovar

ianc

e St

ruct

ure

Mod

elin

g: S

en-

sitiv

ity t

o U

nder

para

met

eriz

ed M

odel

Mis

spec

ifica

tion,

” Ps

ycho

logi

cal M

etho

ds, 3

(4)

, 42

4-53

. ¹

Hu,

Li-t

ze a

nd P

eter

M. B

entle

r (19

99),

“Cut

off C

riter

ia fo

r Fit

Inde

xes

in C

ovar

ianc

e St

ruc-

ture

Ana

lysi

s: C

onve

ntio

nal C

riter

ia V

ersu

s N

ew A

ltern

ativ

es,”

Stru

ctur

al E

quat

ion

Mod

-el

ing,

6 (1

), 1-

55.

¹M

arsh

, Her

bert

W.,

Kit-

Tai H

au, J

ohn

R. B

alla

, and

Dav

id G

rays

on (1

998)

, “Is

Mor

e Ev

er T

oo

Muc

h? T

he N

umbe

r of I

ndic

ator

s Pe

r Fac

tor i

n C

onfir

mat

ory

Fact

or A

naly

sis,

” Mul

tivar

i-at

e Be

havi

oral

Res

earc

h, 3

3 (2

), 18

1 - 2

20.

¹Re

illy,

Ter

ence

(199

5), “

A N

eces

sary

and

Suffi

cien

t Con

ditio

n fo

r Ide

ntifi

catio

n of

Con

firm

a-to

ry F

acto

r Ana

lysi

s Mod

els o

f Fac

tor C

ompl

exity

One

,” So

ciol

ogic

al M

etho

ds R

esea

rch,

23

(4),

421-

41.

Exem

plar

y re

adin

g ¹

Dol

l, W

. J.,

W. X

ia, G

. Tor

kzad

eh (1

994)

, “A

Con

firm

ator

y Fa

ctor

Ana

lysi

s of

the

End

-Use

r C

ompu

ting

Satis

fact

ion

Inst

rum

ent,”

MIS

Qua

rter

ly, 1

(2),

453-

461.

¹Lo

o, R

. and

P. L

oew

en (2

00

4), “

Con

firm

ator

y Fa

ctor

Ana

lyse

s of

Sco

res

From

Ful

l and

Sho

rt

Vers

ions

of

the

Mar

low

e–C

row

ne S

ocia

l Des

irabi

lity

Scal

e,”

Jour

nal o

f A

pplie

d So

cial

Ps

ycho

logy

, 34(

11), 2

343-

2352

. ¹

Yoo,

Mah

n H

ee a

nd J

aebe

om S

uh (2

003

), “O

rgan

izat

iona

l citi

zens

hip

beha

vior

s and

ser

vice

qu

ality

as

exte

rnal

effe

ctiv

enes

s of

con

tact

em

ploy

ees,

” Jo

urna

l of B

usin

ess

Rese

arch

, 56

(8),

p. 5

97-6

11.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 13

Day

Topi

cTe

xt C

hapt

ers

Read

ings

4(14

h00

- 15h

30)

Stru

ctur

al E

quat

ion

Mod

elin

g (S

EM)

(Lat

ent v

aria

ble,

man

ifest

var

iabl

e, e

ndog

enou

sva

riabl

e, e

xoge

nous

var

iabl

e, m

easu

rem

ent m

od-

el,

stru

ctur

al m

odel

, m

easu

rem

ent

erro

rs,

anal

ysis

of co

varia

nce

stru

ctur

es,

path

an

alys

is,

nest

ed

mod

els,

non

-nes

ted

mod

els,

mod

el s

peci

ficat

ion,

m

odel

iden

tifica

tion,

par

amet

er e

stim

atio

n, g

oodn

ess-

offit

asse

ssm

ent,

mod

el m

odifi

catio

n).

Exer

cise

: ¹

R ex

ampl

e of

kee

ping

loya

l cus

tom

ers.

Dat

aset

s:

¹H

BAT

_SEM

.csv

Requ

ired

read

ing

¹G

arso

n, G

. Dav

id (

2014

). “S

truc

tura

l Equ

a-tio

ns M

odel

ing”

, St

atis

tical

Ass

ocia

tes

Publ

ishi

ng, [

avai

labl

e at

htt

p://

ww

w.s

ta-

tistic

alas

soci

ates

.com

/boo

klis

t.htm

], la

st

acce

ssed

: Aug

ust 1

2th,

20

14.

¹M

uthé

n, L

inda

K.

and

Ben

gt O

. M

uthé

n (2

010

): M

plus

Use

r’s G

uide

(6

ed.).

Los

A

ngel

es: M

uthé

n &

Mut

hén,

cha

pter

5,

15, 1

6, 17

and

18.

Reco

mm

ende

d re

adin

g ¹

Bau

mga

rtne

r, H

ans

and

Chr

istia

n H

ombu

rg (1

996)

, “A

pplic

atio

ns o

f Str

uctu

ral E

quat

ion

Mod

el-

ing

in M

arke

ting

and

Con

sum

er R

esea

rch:

A R

evie

w,”

Inte

rnat

iona

l Jou

rnal

of

Rese

arch

in

Mar

ketin

g, 13

(2),

139-

61.

¹Ia

cobu

cci,

Daw

n (2

00

9), “

Ever

ythi

ng y

ou a

lway

s w

ante

d to

kno

w a

bout

SEM

(str

uctu

ral e

qua-

tions

mod

elin

g) b

ut w

ere

afra

id to

ask

,” Jo

urna

l of C

onsu

mer

Psy

chol

ogy,

19(4

), 67

3-68

0.

¹Ia

cobu

cci,

Daw

n (2

010

), “S

truc

tura

l equ

atio

ns m

odel

ing:

Fit

Indi

ces,

sam

ple

size

, and

adv

ance

d to

pics

,“ Jo

urna

l of C

onsu

mer

Psy

chol

ogy,

20(1)

, 90

-98.

¹

Klin

e, R

ex B

. (20

10),

Prin

cipl

es a

nd P

ract

ice

of S

truc

tura

l Equ

atio

n M

odel

ing.

New

Yor

k: T

he

Gui

lford

Pre

ss, c

hapt

er 2

, 4, 5

and

10.

¹St

eenk

amp,

Jan

-Ben

edic

t E.M

. and

Han

s B

aum

gart

ner (

200

0), “

On

the

Use

of S

truc

tura

l Equ

a-tio

n M

odel

s fo

r Mar

ketin

g M

odel

ing,

” Int

erna

tiona

l Jou

rnal

of R

esea

rch

in M

arke

ting,

17 (2

-3),

195-

202.

Follo

w-u

p re

adin

g ¹

Asp

arou

hov,

T. a

nd B

. Mut

hén

(20

09)

, “Ex

plor

ator

y St

ruct

ural

Equ

atio

n M

odel

ing,

” St

ruct

ural

Eq

uatio

n M

odel

ing:

A M

ultid

isci

plin

ary

Jour

nal,

16(3

), 39

7-43

8.

¹B

olle

n, K

enne

th A

. (19

89),

Stru

ctur

al E

quat

ions

with

Lat

ent V

aria

bles

. New

Yor

k: J

ohn

Will

ey &

So

ns, c

hapt

er 1.

¹

Bol

len,

Ken

neth

A. a

nd W

. R. D

avis

(20

09)

, “Tw

o Ru

les

of Id

entifi

catio

n fo

r St

ruct

ural

Equ

atio

n M

odel

s,” S

truc

tura

l Equ

atio

n M

odel

ing:

A M

ultid

isci

plin

ary

Jour

nal,

16(3

), 52

3 - 5

36.

¹Ja

rvis

, Che

ryl B

., Sc

ott B

. Mac

Kenz

ie, a

nd P

hilip

M. P

odsa

koff

(20

03) “

A C

ritic

al R

evie

w o

f Con

-st

ruct

Indi

cato

rs a

nd M

easu

rem

ent

Mod

el M

issp

ecifi

catio

n in

Mar

ketin

g an

d C

onsu

mer

Re-

sear

ch,”

Jour

nal o

f Con

sum

er R

esea

rch,

30

(2),

199-

218.

¹

Mar

sh, H

. W.,

B. M

uthé

n, e

t al

. (20

09)

, “Ex

plor

ator

y St

ruct

ural

Equ

atio

n M

odel

ing,

Inte

grat

ing

CFA

and

EFA

: App

licat

ion

to S

tude

nts’

Eval

uatio

ns o

f Uni

vers

ity T

each

ing,

” St

ruct

ural

Equ

a-tio

n M

odel

ing:

A M

ultid

isci

plin

ary

Jour

nal,

16(3

), 43

9-47

6.

¹Sc

hrei

ber,

Jam

es B

. et a

l. (2

00

6), “

Repo

rtin

g St

ruct

ural

Equ

atio

n M

odel

ing.

The

Jou

rnal

of E

du-

catio

nal R

esea

rch,

99(

6), 3

23-3

37. C

onfir

mat

ory

Fact

or A

naly

sis

Resu

lts: A

Rev

iew

,” Th

e Jo

ur-

nal o

f Edu

catio

nal R

esea

rch,

99(

6), 3

23-3

37.

Exem

plar

y re

adin

g ¹

Alg

eshe

imer

, R.,

U. D

hola

kia

and

A. H

errm

ann

(20

05)

: The

Soc

ial I

nflue

nce

of B

rand

Com

mun

ity:

Evid

ence

from

Eur

opea

n C

ar C

lubs

, Jou

rnal

of M

arke

ting,

69(

7), p

. 19-

34.

5(9

h30

- 17h

30)

WO

RKSH

OP

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 14

Day

Topi

cTe

xt C

hapt

ers

Read

ings

6(9

h30

-11h0

0)Lo

gist

ic R

egre

ssio

n

(logi

stic

regr

essi

on, m

axim

um li

kelih

ood

estim

a-tio

n, lo

g-lik

elih

ood,

bas

elin

e m

odel

, r, H

osm

er

and

Lem

esho

w’s

r2l,

r-st

atis

tic, C

ox a

nd S

nell’

s r2

cs, N

agel

kerk

e’s

r2n,

Wal

d st

atis

tic, o

dds

ratio

(e

xp(b

)), in

com

plet

e in

form

atio

n, c

ompl

ete

sepa

-ra

tion,

ove

rdis

pers

ion)

.

Exer

cise

:

¹R

exam

ple

of lo

gist

ic re

gres

sion

. ¹

Test

ing

logi

stic

regr

essi

on a

ssum

ptio

ns.

Dat

aset

s:

¹H

BAT

.csv

Requ

ired

read

ing

¹A

F, p

p.31

2-35

8. ¹

HB

BA

, 20

10, p

p. 2

61-3

30.

Reco

mm

ende

d re

adin

g

¹G

arso

n, G

. Dav

id (2

00

9), “

Logi

stic

Reg

ress

ion,

” fr

om S

tatn

otes

: Top

ics

in M

ultiv

aria

te A

naly

sis,

(ac-

cess

ed F

ebru

ary

19th

, 20

10),

[ava

ilabl

e at

htt

p://

facu

lty.c

hass

.ncs

u.ed

u/ga

rson

/pa7

65/s

tatn

ote.

htm

l]. ¹

Mal

hotr

a, N

ares

h K

. (19

84),

“The

Use

of

Line

ar L

ogit

Mod

els

in M

arke

ting

Rese

arch

,” Jo

urna

l of

Mar

ketin

g Re

sear

ch, 2

1 (1/

Feb.

), 20

-31.

¹Pe

ng, C

.-Y. J

., K

.L. L

ee, a

nd G

.M. I

nger

soll

(20

02)

, “A

n In

trod

uctio

n to

Log

istic

Reg

ress

ion

Ana

lysi

s an

d Re

port

ing,

” The

Jou

rnal

of E

duca

tiona

l Res

earc

h, 9

6 (1

), 3-

14.

Follo

w-u

p re

adin

g ¹

And

rew

s, R

ick

L., A

ndre

w A

insl

ie, a

nd Im

ran

S. C

urri

m (

200

2), “

An

Empi

rica

l Com

pari

son

of L

ogit

Cho

ice

Mod

els

with

Dis

cret

e V

ersu

s C

ontin

uous

Rep

rese

ntat

ions

of

Het

erog

enei

ty,”

Jour

nal o

f M

arke

ting

Rese

arch

, 39

(4),

479-

487.

¹Sw

ait,

Joffr

e an

d Jo

rdan

Lou

vier

e (19

93),

“The

Rol

e of

the

Sca

le P

aram

eter

in t

he E

stim

atio

n an

d C

ompa

riso

n of

Mul

tinom

ial L

ogit

Mod

els,

” Jou

rnal

of M

arke

ting

Rese

arch

, 30

(3/A

ug.),

30

5-31

4.

Exem

plar

y re

adin

g ¹

Reic

hhel

d, F

. (20

03),

“The

Onl

y N

umbe

r Yo

u N

eed

to G

row

,” H

arva

rd B

usin

ess

Revi

ew, D

ecem

ber,

46-5

4. ¹

Win

er, R

usse

ll S.

(198

6), “

A R

efer

ence

Pri

ce M

odel

of B

rand

Cho

ice

for F

requ

ently

Pur

chas

ed P

rod-

ucts

,” Jo

urna

l of C

onsu

mer

Res

earc

h, 13

(2/S

ep.),

250

-256

.

6(14

h00

-15h3

0)C

onjo

int M

easu

rem

ent (

CO

NJO

INT)

(late

nt A

dapt

ive

conj

oint

ana

lysi

s (a

ca),

choi

ce

set,

choi

ce s

imul

ator

, cho

ice-

base

d co

njoi

nt a

nal-

ysis

(cbc

), co

mpo

sitio

n ru

le, c

ompo

sitio

nal m

odel

, co

njoi

nt

anal

ysis

, de

com

posi

tiona

l m

odel

, fu

ll pr

ofile

met

hod,

full

fact

oria

l exp

erim

enta

l des

ign,

ho

ldou

t pro

files

, mai

n eff

ects

, mar

ket s

imul

atio

ns,

orth

ogon

al d

esig

n, p

lanc

ards

, rev

ersa

ls, t

rade

-off

met

hod,

(par

t-w

orth

) util

ity).

Exer

cise

: ¹

R ex

ampl

e of

dev

elop

ing

effec

tive

prod

uct

desi

gn w

ith C

ON

JOIN

T an

alys

isD

atas

ets:

¹

RB

CRe

spon

ses.

csv.

Requ

ired

read

ing

¹H

BB

A, 2

010

, pp.

261

-330

.Re

com

men

ded

read

ing

¹G

reen

Pau

l E. a

nd V

. Sri

niva

san

(1990

), “C

onjo

int

Ana

lysi

s in

Mar

ketin

g: N

ew D

evel

opm

ents

With

Im

plic

atio

ns fo

r Res

earc

h an

d Pr

actic

e,” J

ourn

al o

f Mar

ketin

g, 5

4(4)

, 3-1

9.

¹G

reen

Pau

l E.,

A.M

. Kri

eger

, and

Y. W

ind

(20

01),

“Th

irty

Yea

rs o

f Con

join

t Ana

lysi

s: R

eflec

tions

and

Pr

ospe

cts,

” Int

erfa

ces,

31(

3), 5

6-73

. ¹

Hau

ser,

J. R

. and

V.R

. Rao

(20

03),

“Con

join

t Ana

lysi

s, R

elat

ed M

odel

ing,

and

App

licat

ions

,” in

Mar

ket

Rese

arch

and

Mod

elin

g: P

rogr

ess

and

Pros

pect

s: A

Tri

bute

to P

aul E

. Gre

en (I

nter

natio

nal S

erie

s in

Qua

ntita

tive

Mar

ketin

g), Y

oram

(Jer

ry) W

ind

and

Paul

E. G

reen

(eds

.), IS

QM

, Int

erna

tiona

l Se-

ries i

n Q

uant

itativ

e M

arke

ting,

Klu

wer

Aca

dem

ic P

ublis

hers

.

Follo

w-u

p re

adin

g ¹

Hub

er, J

. (20

04)

, Con

join

t Ana

lysi

s: H

ow w

e go

t her

e an

d w

here

we

are

(an

upda

te),

Saw

toot

h So

ft-w

are

Rese

arch

Pap

er S

erie

s, 1-

15.

Exem

plar

y re

adin

g ¹

Qua

ltrix

.com

(20

11): C

onjo

int A

naly

sis:

Exp

lain

ing

Full

Profi

le a

nd S

elf E

xplic

ated

App

roac

hes,

[ava

il-ab

le a

t: htt

p://

ww

w.q

ualt

rics

.com

/doc

s/C

onjo

intO

verv

iew

.pdf

]. ¹

Witt

ink,

D. a

nd P

. Catt

in (1

989)

, “C

omm

erci

al U

se o

f Con

join

t Ana

lysi

s: A

n U

pdat

e,”

Jour

nal o

f Mar

-ke

ting,

53

(Jul

y), 9

1–96

.

Con

join

t Pac

kage

s ¹

Aiz

aki,

H. a

nd N

ishi

mur

a, K

. (20

08)

, “D

esig

n an

d A

naly

sis

of C

hoic

e Ex

peri

men

ts U

sing

R: A

Bri

ef

Intr

oduc

tion“

, Agr

icul

tura

l Inf

omat

ion

Rese

arch

, 17(

2), 8

6-94

. ¹

Cro

issa

nt, Y

. (20

09)

, “m

logi

t: a

R pa

ckag

e fo

r the

est

imat

ion

of th

e m

ultid

imen

sion

al lo

git“

, Pre

sent

a-tio

n at

LET

Uni

vers

ity L

yon

II, J

uly

9, 2

00

9. ¹

Imai

, K. a

nd v

an D

.A.D

yk (2

00

5), “

MN

P: R

Pac

kage

for F

ittin

g th

e M

ultin

omia

l Pro

bit M

odel

“, Jo

urna

l of

Sta

tistic

al S

oftw

are,

14(3

), 1-3

2.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 15

Day

Topi

cTe

xt C

hapt

ers

Read

ings

7(9

h30

- 11h

00)

Soci

al N

etw

ork

Ana

lysi

s (S

NA

)(s

ocia

l net

wor

k an

alys

is, d

yadi

c da

ta, n

ode,

tie,

full-

netw

ork

met

hods

, sno

wba

ll m

etho

ds, e

go-c

entr

icne

twor

ks w

ith a

lters

, ego

-cen

tric

net

wor

ks, e

goon

ly, c

entr

ality

, bet

wee

nnes

s, b

ridge

, clo

sene

ss,

clus

terin

g-co

effici

ent,

cohe

sion

, deg

ree,

den

sity

,pa

th le

ngth

, str

uctu

ral h

ole)

.

Exer

cise

:

Exam

ple

of S

ocia

l Net

wor

k A

naly

sis

Dat

aset

s:

¹tb

a

Requ

ired

read

ing

¹B

orga

tti,

S. P

. and

P. F

oste

r (2

003

), “T

he

New

Par

adig

m i

n O

rgan

izat

iona

l Re

-se

arch

: A R

evie

w a

nd T

ypol

ogy,

” Jou

r-na

l of M

anag

emen

t, 29

(6),

991-

1013

. ¹

Iaco

bucc

i, D

. and

S. W

asse

rman

(19

88),

“A G

ener

al F

ram

ewor

k fo

r th

e St

a-tis

tical

Ana

lysi

s of

Seq

uent

ial D

yadi

c In

tera

ctio

n D

ata,

” Ps

ycho

logi

cal

Bul

-le

tin, 1

03 (3

), 37

9-39

0.

Reco

mm

ende

d re

adin

g ¹

Scott

, J. (

200

0), S

ocia

l Net

wor

k A

naly

sis,

New

bury

Par

k C

A: S

age.

¹W

asse

rman

, S. a

nd K

. Fau

st (

1994

), So

cial

Net

wor

k A

naly

sis:

Met

hods

and

App

licat

ions

, Cam

-br

idge

Uni

vers

ity P

ress

.

Follo

w-u

p re

adin

g ¹

Adl

er, P

. and

S. K

won

(20

02)

, “So

cial

cap

ital:

Pros

pect

s fo

r a n

ew c

once

pt,”

Aca

dem

y of

Man

age-

men

t Rev

iew

, 27(

1), 17

-40

. ¹

Car

ringt

on, S

. and

S. W

asse

rman

(Eds

) (20

05)

, Mod

els

and

Met

hods

in S

ocia

l Net

wor

k A

naly

sis,

C

ambr

idge

Uni

vers

ity P

ress

. ¹

Han

nem

an, R

. A. a

nd M

. Rid

dle

(20

05)

, Int

rodu

ctio

n to

Soc

ial N

etw

ork

Met

hods

. Riv

ersi

de, C

A:

Uni

vers

ity o

f Cal

iforn

ia, R

iver

side

, [av

aila

ble

at: h

ttp:

//fa

culty

.ucr

.edu

/~ha

nnem

an/]

. ¹

Kild

uff, M

. and

W. T

sai (

2003

), So

cial

Net

wor

ks a

nd O

rgan

izat

ions

. Lon

don:

Sag

e.

Exem

plar

y re

adin

g ¹

Dod

ds, P

., R

. Muh

amad

, and

D. W

atts

(20

03),

“An

Expe

rimen

tal S

tudy

of S

earc

h in

Glo

bal S

ocia

l N

etw

orks

,” Sc

ienc

e, 3

01 (

5634

), 82

7-82

9. ¹

Gol

denb

erg,

J.,

B. L

ibai

, E. M

ulle

r, an

d S.

Str

emer

ch (2

010

), “D

atab

ase

Subm

issi

on -

The

Evol

ving

So

cial

Net

wor

k of

Mar

ketin

g Sc

hola

rs,”

Mar

ketin

g Sc

ienc

e, 2

9(3)

, 561

-567

. ¹

Goy

al, S

. (20

07)

, Con

nect

ions

: An

Intr

oduc

tion

to th

e Ec

onom

ics

of N

etw

orks

, Prin

ceto

n U

nive

r-si

ty P

ress

. ¹

Gra

nove

tter

, M. (

1973

), “T

he S

tren

gth

of W

eak

Ties

,” A

mer

ican

Jou

rnal

of S

ocio

logy

, 78

(6),

1360

-13

80.

¹H

anak

i, N

., A

. Pet

erha

nsl,

P. D

odds

, and

D.J

. Watt

s (2

00

7), “

Coo

pera

tion

in E

volv

ing

Soci

al N

et-

wor

ks,”

Man

agem

ent S

cien

ce, 2

00

7, 5

3 (7

), 10

36-1

050

. ¹

New

man

, M.,

A. B

arab

asi,

and

D.J

. Watt

s (2

00

6), T

he S

truc

ture

and

Dyn

amic

s of

Net

wor

ks, P

rinc-

eton

Uni

vers

ity P

ress

. ¹

Van

den

Bul

te, C

. and

Y. J

oshi

(20

07)

, “N

ew P

rodu

ct D

iffus

ion

with

Infl

uent

ials

and

Im

itato

rs,”

Mar

ketin

g Sc

ienc

e, 2

6(3)

, 40

0-4

21.

¹W

atts,

D. J

., an

d S.

H. S

trog

atz

(1998

): “C

olle

ctiv

e D

ynam

ics

of ‘S

mal

l-Wor

ld’ N

etw

orks

,” N

atur

e,

393

(668

4), 4

09–

10.

¹W

atts,

D.J

. and

J. P

erett

i (20

07)

, “V

iral M

arke

ting

in t

he R

eal W

orld

,” H

arva

rd B

usin

ess

Revi

ew,

200

7, M

ay, 2

2-23

. Po

pula

r rea

ding

¹B

arab

asi,

A.-L

. (20

02)

, Lin

ked:

The

New

Sci

ence

of N

etw

orks

, Cam

brid

ge, M

A: P

erse

us.

¹B

ucha

nan,

M. (

200

2), S

mal

l Wor

ld: U

ncov

erin

g N

atur

e’s

Hid

den

Net

wor

ks, L

ondo

n: W

iede

nfel

d an

d N

icol

son.

¹

Watt

s, D

unca

n J.

(199

9), S

mal

l Wor

lds:

The

Dyn

amic

s of

Net

wor

ks b

etw

een

Ord

er a

nd R

ando

m-

ness

. Prin

ceto

n, N

J: P

rince

ton

Uni

vers

ity P

ress

.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 16

Day

Topi

cTe

xt C

hapt

ers

Read

ings

Key

Soft

war

e ¹

Gep

hi -

free

war

e - g

ener

al s

oftw

are,

htt

p://

ww

w.g

ephi

.org

. ¹

Keyp

laye

r – fr

eew

are

– ide

ntify

impo

rtan

t nod

es w

hose

elim

inat

ion

may

dis

rupt

net

wor

ks, h

ttp:

//w

ww

.an

alyt

icte

ch.c

om/p

rodu

cts.

htm

. ¹

Net

draw

– fr

eew

are

– vis

ual a

naly

sis (

pack

aged

with

UC

INET

), htt

p://

ww

w.a

naly

tict

ech.

com

/net

draw

/ne

tdra

w.h

tm.

¹PA

JEK

– fr

eew

are-

sui

tabl

e fo

r la

rge

netw

orks

and

clu

ster

ing

algo

rithm

s, h

ttp:

//vl

ado.

fmf.u

ni-lj

.si/

pub/

netw

orks

/paj

ek/d

efau

lt.h

tm.

¹U

CIN

ET 6

– fr

ee fo

r 60

days

- exc

elle

nt g

ener

al p

acka

ge, h

ttp:

//w

ww

.ana

lyti

ctec

h.co

m/u

cine

t/.

Soci

al N

etw

ork

Bib

liogr

aphy

¹htt

p://

ww

w.s

ocia

lnet

wor

ks.o

rg/.

Jour

nals

on

Net

wor

ks a

nd S

ocia

l Str

uctu

re

¹So

cial

Net

wor

ks: h

ttp:

//ee

s.el

sevi

er.c

om/s

on/.

¹C

onne

ctio

ns: h

ttp:

//w

ww

.ana

lyti

ctec

h.co

m/c

onne

ctio

ns/.

¹Jo

urna

l of S

ocia

l Str

uctu

re: h

ttp:

//w

ww

.cm

u.ed

u/jo

ss/.

¹N

etw

ork

Scie

nce:

htt

p://

jour

nals

.cam

brid

ge.o

rg/N

WS.

7(14

h00

- 15h

30)

Clu

ster

Ana

lysi

s (C

A)

(clu

ster

ana

lysi

s, c

lust

er s

olut

ion,

den

drog

ram

, het

-er

ogen

eity

, ho

mog

enei

ty,

hier

arch

ical

m

etho

ds,

nonh

iera

rchi

cal

met

hods

, in

tero

bjec

t si

mila

rity,

di

stan

ce m

easu

res,

agg

lom

erat

ive

met

hods

, di

vi-

sive

met

hods

, abs

olut

e Eu

clid

ean

dist

ance

, ave

rage

lin

kage

, cen

troi

d m

etho

d, c

ity-b

lock

dis

tanc

e, c

lus-

ter

cent

roid

, sin

gle

linka

ge, k

-mea

ns, m

ahal

anob

is

dist

ance

(d2)

, War

d’s

met

hod)

.

Exer

cise

: ¹

Rexa

mpl

e of

con

sum

er s

egm

enta

tion

with

CA

.D

atas

ets:

¹

HB

AT.c

sv

Requ

ired

read

ing

¹A

F, p

p. 7

49-8

11.

¹H

BB

A, 2

010

, pp.

50

5-56

4.

Reco

mm

ende

d re

adin

g ¹

Abo

nyi,

J. a

nd B

. Fei

l (20

07)

, Clu

ster

Ana

lysi

s fo

r Dat

a M

inin

g an

d Sy

stem

Iden

tifica

tion.

Bos

ton

and

Bas

el, S

witz

erla

nd: B

irkhä

user

Bas

el.

¹G

arso

n, G

. D

avid

(20

09)

, “C

lust

er A

naly

sis,

” St

atis

tical

Ass

ocia

tes

Publ

ishi

ng,

[ava

ilabl

e at

htt

p://

ww

w.s

tatis

tical

asso

ciat

es.c

om/b

ookl

ist.h

tm],

last

acc

esse

d: A

ugus

t 12t

h, 2

014

.Fo

llow

-up

read

ing

¹Ed

elbr

ock,

C. (

1979

), “C

ompa

ring

the

Acc

urac

y of

Hie

rarc

hica

l Clu

ster

ing

Alg

orith

ms:

The

Pro

b-le

m o

f Cla

ssify

ing

Ever

ybod

y,” M

ultiv

aria

te B

ehav

iora

l Res

earc

h, 14

, 367

-384

. ¹

Mill

igan

, G. W

. and

M.C

. Coo

per

(1985

), “A

n Ex

amin

atio

n of

Pro

cedu

res

for

Det

erm

inin

g th

e N

umbe

r of C

lust

ers

in a

Dat

a Se

t,” P

sych

omet

rika,

50

(29)

; 159

-179.

Exem

plar

y re

adin

g ¹

Ketc

hen,

D.J

. and

C.L

. Sho

ok (1

996)

, “Th

e A

pplic

atio

n of

Clu

ster

Ana

lysi

s in

Str

ateg

ic M

anag

e-m

ent R

esea

rch:

An

Ana

lysi

s an

d C

ritiq

ue,”

Stra

tegi

c M

anag

emen

t Jou

rnal

, 17,

441

-458

. ¹

Punj

, G. a

nd D

. Ste

war

t (19

83),

“Clu

ster

Ana

lysi

s in

Mar

ketin

g Re

sear

ch: R

evie

w a

nd S

ugge

stio

ns

for A

pplic

atio

n,” J

ourn

al o

f Mar

ketin

g Re

sear

ch, 2

0 (M

ay),

134-

148.

Mar

ketin

g A

naly

tics

II - S

ylla

bus

- 17

Day

Topi

cTe

xt C

hapt

ers

Read

ings

8(9

h30

- 11h

00)

Mac

hine

Lea

rnin

g (D

ecis

ion

Tree

s)(m

achi

ne le

arni

ng, c

lass

ifica

tion

tree

s, b

aggi

ng,

rand

om fo

rest

).

Exer

cise

: ¹

tba

Dat

aset

s:

¹tb

a

Requ

ired

read

ing

¹Ja

mes

, G.,

Witt

en, D

., H

astie

, T. a

nd T

ibsh

irani

, R.

(20

15),

AnI

intr

oduc

tion

to S

tatis

tical

Lea

rn-

ing

with

App

licat

ions

in R

, Cha

pter

8. (

http:

//w

ww

-bcf

.usc

.edu

/~ga

reth

/ISL

/ISL

R%

20Si

xth%

20Pr

intin

g.pd

f) ¹

Pan

g-N

ing

Tan,

Mic

hael

Ste

inba

ch, V

ipin

Ku-

mar

,(20

14):

Intr

oduc

tion

to D

ata

Min

ing,

Cha

p-te

r 4 (h

ttps

://w

ww

-use

rs.c

s.um

n.ed

u/~k

umar

/dm

book

/ch4

.pdf

).

Reco

mm

ende

d re

adin

g

¹Fr

iedm

an, J

erom

e, T

revo

r Has

tie a

nd R

ober

t Tib

shira

ni, (

2001

): Th

e el

emen

ts o

f sta

tistic

al le

arni

ng,

Cha

pter

9 (h

ttps

://w

eb.st

anfo

rd.e

du/~

hast

ie/l

ocal

.ftp/

Sprin

ger/

OLD

/ESL

II_pr

int4

.pdf

). ¹

The

rnea

u, T

erry

, and

Eliz

abet

h A

tkin

son,

(20

15):

An

intr

oduc

tion

to r

ecur

sive

par

titio

ning

us-

ing

the

RPA

RT r

outin

es, (

https

://cr

an.r-

proj

ect.o

rg/w

eb/p

acka

ges/

rpar

t/vi

gnett

es/l

ongi

ntro

.pd

f).

¹Li

aw, A

ndy

and

Matt

hew

Wie

ner,

(200

2): C

lass

ifica

tion

and

regr

essi

on b

y ra

ndom

Fore

st,

(htt

p://

ww

w.b

ios.u

nc.e

du/~

dzen

g/BI

OS7

40/r

ando

mfo

rest

.pdf

).

8(14

h00

- 15h

30)

Mac

hine

Lea

rnin

g (S

VM

)(m

achi

ne le

arni

ng, s

uppo

rt v

ecto

rs, s

epar

atin

g hy

perp

lane

, lin

ear

kern

el,

poly

nom

ial

kern

el,

RB

F ke

rnel

).Ex

erci

se:

¹tr

aini

ng a

sup

port

vec

tor m

achi

neD

atas

ets:

¹

sim

ulat

ed d

atas

et

Requ

ired

read

ing

¹G

unn,

S.R

.,(19

98):

Supp

ort

Vect

or M

achi

nes

for

Cla

ssifi

catio

n an

d Re

gres

sion

, Te

chni

cal

Re-

port

, (htt

p://

ww

w.e

cs.s

oton

.ac.

uk/s

rg/p

ublic

a-tio

ns/p

df/S

VM

.pdf

). ¹

Jam

es G

., W

itten

, G

., H

astie

, T.

and

Tib

shira

ni,

R.,(

2015

): A

n in

trod

uctio

n to

Sta

tistic

al L

earn

-in

g w

ith A

pplic

atio

ns i

n R

,Chp

ater

3,

8 an

d 15

, Sp

ringe

r Sc

ienc

e+B

usin

ess

Med

ia

New

Yo

rk,

(htt

p://

ww

w-b

cf.u

sc.e

du/~

gare

th/I

SL/

ISLR

%20

Sixt

h%20

Prin

ting.

pdf)

.

Reco

mm

ende

d re

adin

g

¹B

ernh

ard

E. B

oser

and

Isab

elle

M. G

uyon

and

Vla

dim

ir Va

pnik

. (19

92).

Proc

eedi

ngs

of t

he fi

fth

annu

al w

orks

hop

on C

ompu

tatio

nal l

earn

ing

theo

ry (C

OLT

). ¹

S. M

ika,

C. S

chaf

er, P

. Las

kov,

D. T

ax a

nd K

.-R. M

ulle

r, Su

ppor

t Vec

tor

Mac

hine

s, B

ookc

hapt

er :

Han

dboo

k of

Com

puta

tiona

l Sta

tistic

s (C

once

pts

and

Met

hods

). ¹

Mac

hine

Lea

rnin

g w

ith R

, ¹

http:

//w

ww

.ker

nel-m

achi

nes.

org/

¹htt

p://

ww

w.s

uppo

rt-v

ecto

r.net

/ ¹

http:

//w

ww

.lear

ning

-with

-ker

nels

.org

/ ¹

http:

//w

ww

.pas

cal-n

etw

ork.

org/

¹htt

p://

ww

w.r-

proj

ect.o

rg/

(pac

kage

s e1

071

,ker

nlab

)

9(9

h30

- 17h

30)

WO

RKSH

OP

10(9

h30

- 14h

30)

Gro

up p

rese

ntat

ions

, fee

dbac

k an

d cl

osin

g


4. EVALUATION The time of the exams is over. Why? Because I believe that learning for an exam is inefficient. Rather I would like to motivate you to learn for life. So, how does your grad-ing take place? The course consists of three formal assessment oppotunities.

4.1 Contributions to the Multiple Choice Questions (50%)

Each day, during the first exercise session, a set of multiple choice questions (MCQ) will be handed out based on last day s class content. You will have fifteen minutes to solve the MCQs. Overall, the MCQ average counts for 50% of your final grade.

4.2 Group Participation on and around the Workshops (50%)

Will randomly form groups and will randomly distribute topics to them. Each group will be linked to one specific lecture and will receive two method-specific tasks. These tasks are: 1) create a realistic marketing application with your method, collect data to this, apply your method, present the results in the presentation block, 2) create “role-model“ output tables per method, code them in R. We will grade the contribution of your group work which will count to 50% of the overall grade (40% application and R code). By doing this we take into account peer scores (10%), i.e. each group member has to grade the participation of each group member. This helps us avoid and identify free-riders. We reserved time for group work throughout our workshops.

4.3 Bonus for exceptional individual participation (+.25 to final grade)

Overall, less than 5% of the students will most likely perceive this bonus. It is given for exceptional performance and adds 0.25 to the final grade, conditional on passing 4.1 and 4.2.

5. ACADEMIC FRAUDThe Honor Code of the University of Zurich applies to all work in this course, and will be strictly enforced. The intent of the Honor Code in this course is to ensure that each student claims and receives credits for his/her own efforts. Violations to this are considered academic fraud.

Definition

Academic fraud is an act by a student, which may result in a false academic evaluation of that student or of another student.

All documents you will hand-in are going to be checked by software and manually for plagiates. Documents with a score above 10% are going to be intensively validated and in suspicious cases we hand-out penalties for fraud behavior.


6. ADMINISTRATIVE COMMENTS

6.1 Students with disabilities

Any student with a documented disability needing academic adjustment or accommodations is requested to speak with me during the first two days of class. All discussion will remain confidential. Students with disabilities will need to also contact the directors of the school.

6.2 Getting in contact with me

Emails should be short and to the point. I don’t have time to read novels and to search for the point. Before sending an email, make clear that email is the appropriate instrument for your task. Maybe a telephone call is much easier and more personal. Or just ask me in class.

6.3 Registration cards

Right in the beginning of the class you will receive a Word file that we ask you to fill-out. In this file we ask you to add a personal picture and personal address information. Each information is kept confidential and is only accessible to our team. The reasons for doing this are 1) we would like to learn your names by pictures, 2) we use pictures later on if you ask reference letters to better remind ourselves, and 3) we need your contact information for the administration. Delivering these files if of course volun-tary. However, we would highly appreciate your cooperation on this. Many thanks in advance.

6.4 Name cards

Please use name cards regularly in class throughout the term so I can learn your names. I usually have large numbers of students across my class, so this will make it easier for me. If you don’t use name cards, I assume you do not care if I know who you are.

6.5 Class dismissal

You are asked to remain seated and attentive until class is dismissed by me.

6.6 Sound-emitting devices

It is expected that you turn off/mute all devices that emit sounds and noises that may interrupt the class (e.g., mobile phones, pagers, watch alarms). If an aoccasion arises in which you may need to receive a telephone call, please inform mw before the class. If you leave a class to answer a call or pager without previously notifying me, you will not be allowed to return to class.

6.7 Laptops and calculators

Laptops and programmable calculators are allows in class if you are asked for them and as far as their usage supports the individual learning process.


6.8 Group formation

We randomly assign groups and topics. If you are not satisfied with the group forma-tion, you can individually discuss and manage this with your classmates If you found a match for changing group affiliations with another student, you are invited to contact us and ask for change in the FIRST semester week. Later on we cannot consider this.

6.9 Important deadlines and class schedule

All important deadlines and the class schedule are communicated in lecture 1. If you can t participate in this class, it is your duty to inform yourself on the process.

05/09 - 09/09 - First course week

12/09 - 16/09 - Second course week

16/09 - Group presentation Workshop

We are very much looking forward to meeting you in class !

Enjoy!