Research Project block 7 2021 Summary FULL slides/notes content compilation

PREPARED BY
Kathleen Gaillot
BT2103 Research Project
RSM Erasmus University




2020 - 2021
IBA Y2 & Pre-Master IBA
Notes primarily based on Lectures &
Extra Quizzes. Coordinator: Yanick Kuper
Office: Mandeville #17, 9th floor
Phone: 010-408-2624
Email:




Summary & Notes
Research Project 1

,Table of Contents
BLOCK 7: EXPERIMENTAL METHODS ............................................................................................................................... 1
Part 1: Recap & Cronbach ............................................................................................................................................. 1
1.1 From Concept to Variable ................................................................................................................................... 1
1.2 Multi-Dimensional Scale ..................................................................................................................................... 1
1.3 Research Model & Hypothesis ............................................................................................................................ 2
1.4 The Right Statistical Test ..................................................................................................................................... 3
Part 2: Experimental Design & Research Models ......................................................................................................... 4
2.1 Causation............................................................................................................................................................. 4
2.2 Experimental Design ........................................................................................................................................... 5
2.3 Experimental Validity .......................................................................................................................................... 7
Part 3: Multivariate Methods: Regression .................................................................................................................... 8
3.1 Background of Regression Analysis ..................................................................................................................... 8
3.2 Application of Regression Analysis ................................................................................................................... 13
3.3 Assumptions & Assessments of Regression Models ......................................................................................... 17
3.4 Advanced Issues: MLG & Prediction ................................................................................................................. 24
3.5 Std.Coef & Prediction ........................................................................................................................................ 27
Part 4: Block 5 & 6 Highlights ...................................................................................................................................... 30
Quizzes & Exam Notes ................................................................................................................................................ 31
Quiz 1 ...................................................................................................................................................................... 31
Quiz 2 ...................................................................................................................................................................... 32
Quiz 3 ...................................................................................................................................................................... 33
Quiz 4 ...................................................................................................................................................................... 34
Quiz 5 ...................................................................................................................................................................... 35
Exam Notes ............................................................................................................................................................. 36




1

, Workload (280h total) Block 5 Trimester 1: 3 EC Block 6 Trimester 2: 3 Block 7 Trimester 3: 4
EC EC
Parts X3 (6 hours) X5 (10 hours) X18 (18 hours)
Workshops X3 (6 hours) X5 (10 hours) X8 (16 hours)
Self-Study X6 weeks 8h (48 hours) X6 weeks 6h (36 hours) X6 weeks 8h (48 hours)
Group Report 24 hours (20%) 25 hours (20%) -
Written Test - 3 hours (20%) 3 hours (20%)
Individual Assignment / Presentation - - 27 hours (20%)
Total 84 hours 84 hours 112 hours

BLOCK 7: EXPERIMENTAL METHODS
Part 1: Recap & Cronbach Lecture 1: Summary of previous blocks (Romain)
1.1 From Concept to Variable
Concepts and constructs vary based on their degree of abstractness: a concept is more abstract than a construct.
 Concept: communication skill (more abstract);
 Construct: vocabulary & syntax skill (less abstract). In this class, we use “concept”/”construct” interchangeably.
 Concepts & constructs are both abstract entities, they are not directly measurable.
 The operationalization of a concept/construct is called a “variable”
o A variable could be a 10 multiple-choice question test on vocabulary and syntax skills with only 1 correct
answer per question (e.g., GMAT, GRE, or even our mid-term exam).

Variables: from measurement to statistical analyses… Two Challenges for Variables:
Challenge 1: for each variables Challenge 2: for several variables within the same
 Find a measure/scale research model
 Determine the variable’s level of measurement  Theoretically: build an empirical model with
 Perform univariate statistics theoretical hypotheses between several variables
 Empirical: test the hypothesis with appropriate
 Special case of multi-dimensional scales
statistical analyses
Levels of Measurement
Type of scale Rules for assigning numbers Examples Univariate statistics
Nominal Object are either identical or Gender, having vs not a feature “do you like this: Frequency, mode
different yes/no” gender…
Ordinal Objects are greater or smaller Relative preferences “how do you rank these Frequency, mode,
beers based on preference?” median
Interval Intervals between adjacent Attitude, agreement, etc “would you like duvel Frequency, mode,
ranks are equal now? Defo/prob/yes…” median, mean
Ratio All properties of interval; 0 Sales, quantity, time “how many beers did you Everything + ‘ratio’
value is meaningful have this week?” statements


1.2 Multi-Dimensional Scale
One variable can be measured with several questions, better known as « items »
1. Using multiple items. Example of measuring brand loyalty with the 4 following items:
1 item=1 column. First, we check the
scale’s reliability with Cronbach alpha.
If 𝛼 ≥ 0.70, then compute the
“composite” measure
If 𝛼 < 0.70, then delete 1 or several items
until 𝛼 ≥ 0.70

 If 𝛼 ≥ 0.70 for a 3-item combination (e.g., BR1, BR2,
BR3) then use, if not
 If 𝛼 ≥ 0.70 for a 2-item combination (e.g., BR1, BR2)…
1

, 2. Case where we have a reverse-coded item
First, we need to recode the variable BR5r (r as
in “reverse”) into BR5
Old values = 1, 2, 3, 4, 5, 6, 7
->
New values = 7, 6, 5, 4, 3, 2, 1

R Tutorial on reversing items:
#install.packages(“readxl”) install.packages(“psych”) installe.packages(“plyr”)
library(“readxl”) library(“psych”) library(“plyr”)
Setwd(“C:/users/kathl/documents/R”) #enter location of folder where database is located
database <- read_excel(“CronbachData.xlsx”, sheet = 1) #open database
br1234 <- data.frame(databasep, c(‘BR1’, ‘BR2’, ‘BR3’, BR4’)]) #create data frame with BR1 to BR4
alpha(br1234) #alpha on BR1 to BR4
br123 <- data.frame(databasep, c(‘BR1’, ‘BR2’, ‘BR3’)]) #data frame with BR1 to BR3
alpha(br123)
database$BR5 <- mapvalues(database$BR$r, from = c(1, 2, 3, 4, 55, 6, 7), to = c(7, 6, 5, 4, 3, 2, 1)) #recode
BR5 (reverse) into BR5
br12345 <- data.frane(database[, c(‘BR1’, ‘BR2’, ‘BR3’, ‘BR4’, ‘BR5’)]) #create data frame with BR1 to BR5
alpha(br12345)
3. If items not measured on the same scale
Need to put all items on the same scale, e.g.:
BR2_new=BR2/7*5 Then, compute Cronbach Alpha
(BR1, BR2new)
4. If there is a mix between categorical and interval items
It is NOT possible to merges these two items.



Aggregating several nominal items is only possible if
• The items are binary and the one answer is consistently interpreted as the same in different items (e.g., one
alternative always = “brand loyal”, one alternative always = “true”)
o The aggregated measure would become a ratio variable. Example: Cognitive reflection test
1) A bat & a ball cost €1.10 in total. The bat costs a euro more than the ball. How much does the ball cost?
2) If it takes 5 machines 5 minutes to make 5 widgets, how long would it take 100 machines to make 100 widgets?
3) In a lake, there is a patch of lily pads. Every day, the patch doubles in size. If it takes 48 days for the patch to
cover the entire lake, how long would it take for the patch to cover half of the lake?
a. For each of the items, it’s “right” or “wrong” → Binary
b. Aggregated measure at the individual-level: how many “right”? → Ratio

1.3 Research Model & Hypothesis
A research model determines the relationships between the variables you are exploring in your research.
The simplest research model has two variables: →
 The researcher builds on previous research and
posits a theoretical hypothesis about the
relationship between the variables of interest.
o Example: Researcher X posits that students who wear glasses have better grades than those who do not.
 Note: a theoretical hypothesis without direction is not testable “Wearing glasses influences grades”

A good formulation includes the direction of the effect “Wearing glasses increases grades” “Students who
wear glasses have better grades than student who do not”

Next, the researcher chooses measures for the
two variables of interest: →




2