Lecture 1
Scientific reasoning; three main logics
1. Inductive reasoning
• Given a series of observations, we derive an explanation / generalization that is probably
true
2. Deductive reasoning
• Based on premises that are true, we logically come to a conclusion that is true
3. Abductive reasoning
• Based on interactions between observations and theories, we come to a likely explanation
for what we
see.
Research cycle – general logic
Managerial problem
• Typically, a “performance” problem – find out which performance dimension(s) is/are
unsatisfactory
• Often, the managerial problem is (part of) a problem “mess” – multiple interrelated
problems and a complex cause-
and-effect network
• Different problems as perceived by different stakeholders
,• Unravel the spaghetti, and zoom in on one important problem; check that there are
stakeholders who are willing to
sponsor your project
• Define the knowledge question(s) behind the managerial problem
• Use abductive reasoning, collect prior studies around the problem(s),
and do some quick exploratory research
• Bring everything together in a research objective:
What do you aim to achieve with your research project?
Knowledge question
• Start collecting literature around the concepts mentioned in the research objective and the
initial set of questions
• Refine the questions until you have one main research question and a set of sub-questions
(typically 3-5)
• If your research objective is a theory-testing objective, you would typically also formulate
hypotheses
• Four* types of research questions as related to the type of knowledge they generate
(increasing levels of
complexity):
1. Descriptive knowledge (how things are)
2. Explanatory knowledge (why things are that way)
3. Predictive knowledge (how things will be)
4. Prescriptive knowledge (how things should be done)
Review of evidence
• You already started to collect literature in the previous step
• Now that you have your research questions, you start a more systematic search for
relevant literature
• Academic literature (scientific journal articles, scientific books, working papers, conference
papers)
• Professional literature (magazine articles, professional books, blogs, websites, newspaper
articles)
• Your results are captured in a critical synthesis of the literature (a.k.a. literature review)
Research design
• Determine your research type: exploratory (descriptive), theory-building, theory-testing,
decision science
• Check that research objective, research questions and research type are consistent with
each other
• Determine your research strategy: experiment, survey, case study, modeling, etc.
• Your research design will describe:
• Plans for data collection
• Plans for data analysis
• Threats to validity and how to deal with those
• Time plan / project plan
,Research outcomes
• Distinction between the results itself and your discussion of the results
• Results: outcomes of your analyses
• Discussion; five typical elements
• Conclusion: Answers to your research questions
• Contribution to theory: What have we learned on top of what we already knew from the
literature?
• Contribution to practice: (see next slide)
• Limitations: What are the weaker parts of your research project? And how may these have
impacted the results?
• Suggestions for future research: What are the most important follow-up questions?
Recommendation to management
• As this is a BSc program in Business Administration, be explicit about how the research
results support decision-
making by managers*
• Can you capture such recommendations in a CIMO statement? Like: “From this research
we learn that in Context C,
if you do Intervention I, the Mechanism M will help to achieve Outcome O.”
• Be careful not to over-generalize; be mindful of context
• Think about advantages, disadvantages, risks and critical success factors related to a
recommendation or course of
action.
Critical reflection (in all steps)
• What does it mean to be “critical”?
• It does NOT mean that you have to destructively criticize everything you read, see or have
produced yourself
• Some dimensions of working in a “critical” way:
• Force yourself to identify stronger parts and weaker parts
• Search for agreements and disagreements (compare and contrast)
• Check the authority, quality and possible bias of the source
• Reflect on whose voice is amplified and whose voice is muted
• Reflect on the ethics of your own work and the
work of others
,Data is not the same as knowledge. By working with data you can develop knowledge. Data
is information in unprocessed and raw form. Data collection is the process of acquiring
information form different sources. Data is transferred into knowledge through processing
(reflecting, analyzing, linking, etc). Knowledge is always meaningful but data does not have
to be.
Types of data in research:
There are three classifications of data:
- Primary vs. secondary
- Qualitative vs. quantitative
- Empirical vs. simulated.
Primary vs. Secondary
Primary data: collected by the researcher Secondary data is old compared to
or research team for the first time, with a primary. It was collected by any other
particular research problem in mind person or organization in the past for some
other purpose but used by the researcher
to study her own research problem
Always raw and needs processing and Already somehow organized
analysis
Often time-taking and expensive to collect Often cheap and fast access
Typically smaller data samples Typically bigger data samples and sets
Potentially higher quality for research (most Potentially lower quality (often needs
of the time all required variables are adjustment, not al variables or
present) unnecessary variables can be present.
Collected through questionnaires, Collected from newspapers, websites,
interviews, observations, photos, video and books, organization records, government
audio recordings, focus groups, etc. reports, existing databases, etc.
Qualitative vs. Quantitative
Qualitative data Quantitative data
Data comes in words or pictures Data consists of numbers
The data can describe or categorize Data describes counting or continuous
something measurements.
Questions answered are what? How? Questions answered are how many? How
Why? Etc often?
Data is gathered by talking or observing. Gathered by measuring and counting. This
This means the data is subjective and open makes the data universal and objective
to interpretation
Analyzed by reducing data into meaningful Analyzed statistically. Use of
themes or categories. Examples of questionnaires (with numerical scales),
categories are records of conversations analytical tools, sensors and measurement
with people, texts, audio and video devices and secondary data (stock market
recordings and notes. records, sales information, etc.).
, Empirical vs. Simulated
Simulation is the creation of a model that Empirical real data can be historical or
can be manipulated logically to decide how real-time. It cannot give you future records
the ‘real’ physical world works. In research
it is used to check if a method or a
proposed solution works, or to study
alternative solutions.
Simulated data is used for predicting future
events
Sometimes simulated data is used in early
phases of research to better understand
what real data will be needed
Simulated data is often faster and cheaper
to obtain, so also widely used in teaching
Simulated data is always based on
characteristics of real data, so within
meaningful ranges, etc.
Results obtained based on empirical and simulation data are not always the same. There
may be measurement errors, errors in modelling or errors in the formulation of hypotheses.
The type of management problem drives the type of research you use. Examples of how
research depends on the problem you are researching:
- Exploratory research: used if the problem is new or not yet well defined. Academic
literature on these new topics is typically scarce. Some relevant theories from other topics
can be borrowed. The research design in this type of problems is often inductive or
abductive if topic is not completely new. both qualitative and quantitative data can be
involved, the data type determined the data analysis. In terms of recommendations,
exploratory studies cannot give strong advice; however, they can help understand the
problem better and develop next steps in acting on it.
- Theory-building research: problem is new but more clear. We want to understand main
concepts and relationships between them. Academic literature is still limited but there is
typically more professional literature. This also uses inductive or abductive (theory
refinement) logic because it is still building. Both qualitative and quantitative data possible.
Theory building research cannot provide strong advice, but may be used as an attempt to
explain how something works. Recommendations can be very conceptual but no
propositions have been tested yet.
- Theory-testing research: now there is theoretical knowledge but it is not yet
verified/proven. A hypothesis can be formulated and academic literature is present. However
there is no proof or the relationships that form the topic of the research. deductive logic is
used and data is almost always quantitative, analyzed with statistics. Recommendations are
likely to be convincing, if theory is tested and found to be true. Results are perceived as
generalizable. It can be used as evidence for making decisions, but context needs to be
taken into consideration.
- Decision science research: topic can already be studied in general but present a
management problem for the particular company. There are sources present. We do have
information but no particular tool to use for research. the logic is often inductive, you start
