You are on page 1of 14

Lecture 7: Lab-experimental

International Business Research


Anick Bosmans
The seven-step deductive research process
Theory Data
Write-up

Analyze data

Collect data

Choose a research design

Develop a theoretical framework

Formulate the problem statement

Define the business problem


2
Two types of deductive research strategies

Research type Causal Correlational


research research

Research strategies
‒ Lab experiments ‒ Archival research
‒ Field experiments ‒ Survey research

3
Agenda

• Last week: survey research


• Today: lab experimental research
Webclip 7.11. What is experimental research?

Webclip 7.22. Validity (internal and external)

Webclip 7.33. Types of experimental designs


Q&A
starts at
Kick-off 14:00

We welcome your questions:


Webclips, notes, IBR@tilburguniversity.edu (or via
zoom chat)
sending questions • English
• Ask questions following the
order of the webclips
• Be specific! 5
Q&A

6
Webclip Part 1
• I have difficulty with seeing the difference between the ways to
manipulate a variable by frequency vs type. Could you perhaps
give some more examples?

• Type: you want to see the effect of different social media


platforms advertisement on willingness to buy a product
IV in this case can take levels of : facebook vs instagram
• Frequency: you want to see the effect of amount of likes on a
social media post on willingness to buy
IV in this case can take levels of : high (100k likes) medium (50k
likes) low (1k likes)
Webclip Part 3
• Could you explain again the extraneous, confounding, control variables and also make a
comparison between them?

Confound and extraneous variables are used interchangeably in this course (and in the
field) → they contaminate the causal relationship
• Example: Education Level in our study on motivation.
Group 1 (bonus) includes more be high educated employees
Group 2 (no bonus) includes more low educated employees
→ selection bias, and education level is the contaminating extraneous confound variable.

A control variable is a variable that we control for (either statistically, or if possible:


control the environment). For example, if we know that education level is a possible
confound, we can measure it, and include it in our analysis as a covariate.
Webclip Part 3
• Is an extraneous variable the same as a covariate?

Covariate = control variable

Only when we measure the extraneous variable, it becomes a control variable, and
we can include this variable as a covariate in our analysis.
General questions
A lot of experiments in journals seem to be field studies where the
researchers survey a population from various companies, maybe different
positions/ industries. But it still seems like in these studies, [1] various
threats to validity are not addressed. [2] How then should we address
these issues when we want to use them for our own studies. Or should we
not use them at all?

[1] R&V might be addressed in the methodology section, or in the


limitations & future research section. Since you are learning, we put them
the types of R&V very explicit in the research proposals, usually they are
implicitly included in articles.
[2] Only include studies with a sufficient level of R&V, in the top journals
this is in general the case. In the proposal no need to reflect on it too
much (unless it is part of your relevance), in your thesis you will have
more space to reflect on it.
General questions
• Are there certain rules in place for the amount of studies one
has to perform in a row before results can be deemed
generalisable? As randomisation solves a selection bias, yet
it does not ensure that the groups are devided equally across
all possible charactaristics?

• No magic number that makes a study generalizable or not,


depends on multiple factors (environment, sampling etc.)
• Randomization should divide respondent characteristics
• If you are in doubt: Run randomization checks (t-test, chi-
square test)
General questions

• I have a question regarding the content of lecture week 6. There is a


paper mentioned as mandatory material, but to what extent should we
know the content of this?

Papers that are mandatory for the lectures will be asked in the exam. You
have to read, understand and study them to be able to answer questions.
The ‘’study details’’ of the paper does not need to be read.
Individual Assignment 2 – Survey research

• Published 15:00 today


• Business case
• Assignment itself
• Part 1 –PS, theory and methodology (steps 2-4)
• Part 2 –Analysis (step 6)

• Webclips as support for SPSS-analyses

Deadline is March 25th, 15:00 21


Next:
• Practice quiz
• Individual assignment 2

22

You might also like