Research methods

Summary of the book Research methods for operations management –

Christer Karlsson

Sharmaine Linker

Contents
Chapter 1 – Introduction of the book ......................................................................................... 3
Chapter 2 – Research in operations management ...................................................................... 3
2.1 Introduction ...................................................................................................................... 3
2.2 Research outputs and targets ............................................................................................ 3
2.3 Roles of the researcher ..................................................................................................... 3
2.4 The research process ........................................................................................................ 4
2.5 Research as contribution to knowledge ............................................................................ 4
2.6 What to research for academia and practice ..................................................................... 5
2.7 Research quality ............................................................................................................... 5
2.8 Assessing research quality and contribution .................................................................... 6
2.9 Getting published ............................................................................................................. 6
2.10 Research ethics and ethics for researchers ..................................................................... 6
3. The research process .............................................................................................................. 6
3.1 Contributing to knowledge ............................................................................................... 6
3.2 Choosing a research topic ................................................................................................ 7
3.3 Using literature to develop the research topic .................................................................. 7
3.4 Developing research question .......................................................................................... 8
3.5 Considerations in choosing a research approach .............................................................. 8
Chapter 4 – Surveys ................................................................................................................... 9
4.1 Introduction ...................................................................................................................... 9
4.2 The survey research process ............................................................................................. 9
4.3 What is needed prior to survey research design? ........................................................... 10
4.4 How a survey should be designed .................................................................................. 10

1
 4.5 Pilot testing the questionnaire ........................................................................................ 11
4.6 Advancements in theory formalization and survey design ............................................ 12
4.7 Survey execution ............................................................................................................ 12
4.8 Data analysis and interpretation of results ..................................................................... 13
4.9 Information that should be included in articles .............................................................. 13
4.10 Ethical issues in survey research .................................................................................. 13
Chapter 5 – Case research ........................................................................................................ 13
5.1 Introduction .................................................................................................................... 13
5.2 When to use case research .............................................................................................. 13
5.3 The research framework, constructs and questions ........................................................ 14
5.4 Choosing cases ............................................................................................................... 14
5.5 Developing research instruments and protocols............................................................. 15
5.6 Conducting the field research ......................................................................................... 15
5.7 Reliability and validity in case research ......................................................................... 15
5.8 Data documentation and coding ..................................................................................... 15
5.9 Analysis .......................................................................................................................... 16
Chapter 9 – Modelling and simulation ..................................................................................... 16
9.1 Introduction .................................................................................................................... 16
9.2 Origins and development of model-based research in operations management ............ 16
9.3 Methodologies in quantitative modelling ....................................................................... 16
9.4 How to conduct quantitative research in operations management ................................. 17

2
Chapter 1 – Introduction of the book
 Nothing special

Chapter 2 – Research in operations management

2.1 Introduction
Research aims at the creation and development of knowledge and research has to be of a high
quality in terms of validity and reliability. Key components of research are new knowledge
and/or new applications. Mode 1 research is traditional pure knowledge creation and
development. Mode 2 research includes the purpose of implementation or using knowledge to
develop applications. Operations are a transformation activity: transforming resources
(machines and people) into outputs of goods and services. The scope of OM is very wide and
includes for example the following strategic perspective: focus on the role of and the
objectives for the function of operations. Operations play an important role in business model
innovation and development.

2.2 Research outputs and targets

A fundamental characteristic of good research is that is significant contribution to knowledge.
Typically it consists of additions to pre-existing knowledge. It also might be significant in
terms of the contribution to theory or practice. Another important factor is the potential value
that can be created when applying the research results. Research is expected to provide
trustworthy knowledge.

The aim of research is the creation and development of knowledge and the output is
contribution to knowledge. Research may have different purposes: confirmation, falsification
and exploration. Confirmation is often done by confirming previous findings in other settings.
Falsification needs to identify the rejection of an earlier hypothesis or theory. Exploration
offers endless opportunities since there are always issues and problems from empirical
situations that have not been studied before.

2.3 Roles of the researcher

Since there are many different types of research, research approaches, methodologies, used
methods, intended products, contributions and target group(s), there are also many different
types of researchers.
- Observer
- Surveyor ((opzichter) being able to grasp large amounts of data and review it
systematically)
- Analyst (having the ability to break down data and classify into categories)
- Communicator
- Sounding board (calls for an ability to listen, analyse, and then relate to models and /
theories that inform the practitioner)
- Actor and other partner roles
- Consultant (ability to give advice based on the identified problems and developed
knowledge)
- Clinician (curing the research object while at the same time use it as a database)

3
2.4 The research process
1. Identification of a problem or issue to research
2. Literature review
3. Specifying the aim, objective, or purpose of the intended research
4. Determine specific research questions
5. Choice of research approach and methods
6. Development of a conceptual framework
7. Data collection
8. Analysing and interpreting the data
9. Synthesizing and concluding
10. Evaluating the research and suggesting further research
11. Reporting and communicating the research findings

2.5 Research as contribution to knowledge

Explicit knowledge is knowledge that can be captured, documented, transferred, shared and
communicated easily. Tacit knowledge cannot be traced in documents and publications, it is
personal, context-specific and therefore difficult to articulate; it resides only with the
individual. Another distinction is the one between knowledge and skills/craft. Knowledge is
built on a rational formula, while skills can be called know-how. Also a distinction can be
made between skill/craft, virtue and attitude. Virtue are the attributes, attitudes or basic values
of an individual.

At the base of research philosophy is a view of reality that reflects the researcher’s
understanding of what exists and what hence can be said to be, this is called ontology (called
the science of being). Just as there can be different realities, there can be different truths. The
study of reality ontology and the study of knowledge is called epistemology.
There must be good (justified) reasons for believing that something can be verified as
absolutely true regardless of the context, before it can be considered as knowledge.

An aspect of importance is how arguments are build. An argument is said to consist of

premises, inference and conclusions. A premise of an argument is something that is put
forward as a truth, but which is not proven. The conclusion/claim is the statement of the result
of the analysis. The inference is the reasoning process between the premises and conclusions.
There are different types of reasoning:

Logic of argument
Components
Rules Observations Results

Argumentations
Deduction Induction Abduction
Rule Observation Result
↓ ↓ ↓
Observation Result Rule
↓ ↓ ↓
Result Rule Observation

4
Rule: concerns how the world is structured and functions
Observation: the database or the research material

There are different levels of knowledge. The first level is the data, this are those facts that can
be gathered and for which there are measures. When data are put into a system, we form
information. The information can be put in a context and used, and, as a result, add
experience. This enables an interpretation of the information, its applicability, effects and
other strengths and weaknesses. Knowledge is developed.

2.6 What to research for academia and practice

In making a contribution to knowledge, the first step is to think about the intended
contribution. The exact contribution is depended on the outcomes of the research, but an aim
could and should be made. The starting point may also be a research project, a project
possibility, an available case or other empirical material on which to do research. From the
intended contribution it is practical to follow a form of reverse/backward logic, considering
what has to be done to reach the intended contribution.

After deciding on the intended contribution, a more thorough literature review can be
performed. Then the researcher can start to find a way to achieve knowledge development and
plan a study, creating a methodology, analysing the methodology and creating an
empirical/theoretical foundation. The researcher can also plan how to set the contribution in
the research field context.

A researcher should forward a chain of evidence. A report is expected to start with a problem
discussion based on the relevant problem from practice and the theoretical formulation. A
problem definition should be developed. Next a thorough literature review should be given to
develop a picture of the relevant knowledge and gaps in the knowledge. Then the research
questions should be formulated (explorative, hypothesis generating/testing, or whatever). A
conceptual framework of what will be studied should be presented. The next section should be
about the methodology and thereafter the results can be presented. The last section should
contain a conclusion.

2.7 Research quality

To gain research quality, the first question asked should be whether there is a need for the
product and whether it will be demanded. Other question are whether the research was
performed in a reliable way and if the research can repeated by a reader. Furthermore, the
reader should be able to follow the text, understand the chain of logic, and come to the same
conclusions.

There are four requirements that are relevant to OM research:

- Construct validity the operational measures used to measure the construct should
actually measure the concepts that are intended to measure
- Internal validity that the study actually measures what it is meant to measure and
that demonstrated relationships are explained by the factors
described and not by other factors
- External validity that the results are valid in similar settings outside the studied

5
 objects
- Reliability that the study is objective in the sense that other researchers
should reach the same conclusion in the same setting

2.8 Assessing research quality and contribution

A well-done manuscript review should contain general aspects and careful examination. The
general overview should bring up what has been discussed in quality of research and
especially analyse the chain of evidence.
 In the book (p. 32-34) is an extensive explanation on how to perform a good review.

2.9 Getting published

 Nothing relevant for the exam

2.10 Research ethics and ethics for researchers

In this section the rules and principles about ethics are explained.

3. The research process

3.1 Contributing to knowledge
A theory can be defined as a set of interrelated constructs, definitions and propositions that
present a systematic view of phenomena by specifying relationships among variables, with the
purpose of explaining and predicting the phenomena. Theories explain facts and provide
stories as to why phenomena work as they do. Three elements in a theory:
- Conceptual definitions of the theory’s construct
- Defined relationships between constructs
- A limitation of the theory’s domain; the boundary conditions within which the
proposed relationships hold

Theories are not built, they are invented, they cannot be systematically constructed or deduced
from facts, but require inspiration and creativity. They cannot be proved, but only disproved
by findings that run counter to their predictions/explanations.

- References are not theory

- Data are not theory (data is what was observed, theory why it is observed)
- Lists of variables or constructs are not theory
- Diagrams are not theory
- Hypotheses are not theory

Other research products that could constitute valid contribution, but that should not be
confused with theory
- Proposition there is a relationship between propositions and hypotheses,
hypotheses require measures whereas propositions involve
concepts.
- Model is a way to operationalize theory for empirical testing
- Conceptual model sets out in broad terms the definition of the particular problem to

6
 be solved and it specifies the variables that will be used to define
the nature of the problem.
- Taxonomy used to classify things according to their similarities
- Framework explain either graphically, or in narrative form, the main things
to be studied

Ways of making theoretical contribution

- Adding or subtracting factors from an existing theory
- Find alternative explanations for the occurrence of a phenomenon
- Address the domain limitations of the theory

Other ways of making theoretical contribution are to contribute to existing knowledge, or to

use existing theory to explain phenomena. Another way is to examine and test relationships
between variables.

Maturity of Nascent theory Intermediate Mature theory

existing knowledge theory
Types of studies Exploratory, theory Confirmatory, theory
building testing
Types of Identifying new Testing, confirming,
contributions phenomena and suggesting extending or refining
relationships among them existing theory

The contribution on knowledge is also depending an individual’s regards to ontology &

epistemology’.

3.2 Choosing a research topic

Practical problems are not research problems. Practical problems are solved by action, by
doing something that eliminates the cause of the problem. Research problems are defined by
what you do not know or understand about something.

Research problems can arrive from, but not exclusively:

- Industry problems
- Industry trends
- Literature studies
- Previous research
- Personal experience
- Networking
- Funding bodies

3.3 Using literature to develop the research topic

 8 pages of how to use literature and how to find it.

Use of existing literature

- To motivate and position the research
- Create a framework for the research
- Support more hands-on ‘action” in the research

7
 - Interpreting and explaining research findings
- Discuss the research findings in relation to existing literature

Common misuse of literature

- Only describing the literature and not analyse it
- Review literature without using it
- Employ too many (perhaps even incompatible) grand theories
- Usage of high-level theories is taken as a guide for contributing to theory

3.4 Developing research question

Research questions help you narrow your search to the data that you need to answer the
question. Research question are fundamental in bringing focus to research. Other mechanisms
that help to achieve this focus are the purpose, aim, propositions and hypotheses. The
purposes comprise a statement of what the research is trying to achieve and how this end is to
be achieved. An aim is the purpose of doing something: what somebody is trying to achieve.
The aim and purpose are often used interchangeable.

Types of research questions

- Exploratory (who, what, where and when)
- Descriptive (who, what, where and when)
- Explanatory (how and why)
- Prescriptive

Nascent theory Intermediate theory Mature theory

Nature of Open-ended inquiry Proposed relationships Focused questions and/or
research question about a phenomenon between new and hypotheses relating
of interest established constructs existing constructs
Types of research Exploratory Explanatory
questions Descriptive Prescriptive

Basic gap spotting modes Specific versions of basic gap spotting modes
Neglect spotting Looking for areas in existing literature that are overlooked or
under-researched or that lack empirical support
Confusion spotting Looking for competing explanations in existing literature
Application spotting Extending and complementing existing literature

3.5 Considerations in choosing a research approach

Overall considerations in choosing an approach: methodological fit, philosophical position
and practical considerations. The core notion of methodological fit is that there should be
internal consistency among all the main elements of a research project (research question,
maturity of existing knowledge, research approach and contribution). The philosophical
position refers to the reflection of the researcher’s position with regard to fundamental issues
of ontology and epistemology. Practical factors that have to considered are for example the
access, the institutional factors and skills & interests.

8
Chapter 4 – Surveys
4.1 Introduction
Exploratory survey research takes place during the early stages of research on a phenomenon
when the objective is to gain preliminary insight into a topic, and provides the basis for more
in-depth survey research.
Confirmatory (or theory testing or explanatory) survey research takes place when knowledge
of a phenomenon has been articulated in a theoretical form using well-defined concepts,
models and propositions. Data collection is carried out with the specific aim of testing the
concepts developed in relation to the phenomenon, the linkages hypothesized among the
concepts and the validity boundaries of the model.

Holistic construal: this approach is neither rigidly deductive nor purely exploratory. Rather it
subsumes a process by which theories and hypotheses are tentatively formulated deductively,
tested on data and, later, reformulated and retested until a meaningful outcome emerges.

Descriptive survey research is aimed at understanding the relevance of a phenomenon and

describing the incidence or distribution of the phenomenon in a population.

A survey is a suitable method when knowledge of the phenomenon under investigation is not
too underdeveloped, when generalization is an important intended contribution, when the
variables and the context can be detailed and when the empirical evidence sought concerns
‘how variables are related’, ‘where the relations hold’ and ‘to what extent a given relation is
present’.

4.2 The survey research process

Link to theoretical level

Design

Pilot test

Collect data

Analyse data

Generate report

9
4.3 What is needed prior to survey research design?
1. Researcher should have a clear idea of what the problem is about and should discuss it
in depth
2. Researcher should map the problem in the existing literature in order to be cognizant
of what is already know and what is still unknown
3. Researcher should clarify the intended contribution of the research.

Exploratory survey research may pose several problems with regard to framing the research,
as little support is provided by the available literature. In contrast, theory-testing is based on
more established knowledge and, therefore, the focus of the researcher is easier to
communicate clearly.

Establish a conceptual model

1. Construct names and nominal definitions

2. Propositions; presentation and discussion of the role of the constructs, the important
linkages between them and an indication of the nature and direction of the
relationships
3. An explanation
4. Boundary conditions; definition of the conditions under which the researcher might
expect these relationships to hold

Before a researcher can talk about how to collect data, it is necessary to

- Define the unit of analysis corresponding to the level of reference of the theory
- Provide and test the operational definitions for the various constructs
- Translate the propositions into hypotheses

The unit of analysis refers to the level of data aggregation during the subsequent analysis.
Defining the unit of analysis at a theoretical and empirical level is important to avoid the
cross-level inference problem and even the possibility that a reader might fail to understand
the level to which the research refers.

In developing operational definitions, the first problem is to translate the theoretical concepts
into observable and measurable elements. The list of observable elements in each construct
should be developed before writing the items/questions that constitute the measure.
When the operational definition has been developed, the researcher should test it for content
validity. The content validity of a construct measure can be defined as the degree to which a
measure’s items represent a proper sample of the theoretical context domain of a construct.
Evaluating the face validity of a measure can indirectly assess its content validity.

A hypothesis is a logically conjectured relationship between two or more variables

(measures), expressed in the form of testable statements. A hypothesis can also test whether
there are differences between two groups with respect to any variable(s).

4.4 How a survey should be designed

Before embarking on a survey, the research should consider the suitability of the survey
method and the overall feasibility of the research project. In survey research there is a trade-
10
off between time and cost constraint, on the one hand, and minimization of four types of
errors, on the other hand:
- Sampling error
- Measurement error
- Statistical conclusion error
- Internal validity error

The sample
- Population refers to the entire group of people, firms, plants or things that the researcher
wishes to investigate
- An element is a single member of the population
- The population frame is a list of all the elements in the population from which the sample is
to be drawn
- A sample is a subset of the population it comprises some members selected from the
population
- A subject is a single member from the sample
- Sampling is the process of selecting a sufficient number of elements from the population so
that by studying the sample, and understanding the properties or the characteristics of the
sample subjects, the research will be able to generalize the properties or characteristics to the
population elements.
Sampling overcomes the difficulties of collecting data from the entire population, which can
be impossible or prohibitive in terms of time, costs and other human resources.

Data collection methods

- Postal survey
- Face-to-face
- Telephone
- Email/online

 And a lot of other very specific information on how to design a survey

4.5 Pilot testing the questionnaire

After designing the survey, the research has to test it. The pre-test should be done by
submitting the questionnaire to colleagues, industry experts and target respondents. The role
of colleagues is to test whether the questionnaire accomplishes the study objective. The role
of industry experts is to prevent the inclusion of some obvious questions that might reveal the
investigator’s ignorance in some specific areas. The role of the target respondents is to
provide feedback on everything that could affect whether and how the targeted respondents
answer the questions.
Non-respondent can limit the generalizability of the results, in the pilot-testing the researcher
should identify a way to address this problem.

Reliability indicated dependability, stability, predictability, consistency and accuracy and

refers to the extent to which a measuring procedure yields the same results in repeated trials.
- Test-retest method
- Alternative form method
11
 - Split halves method
- Internal consistency method

While content validity refers to the degree to which the meaning of a set of items represents
the domain of the concept under investigation, construct validity refers to the degree to which
scores obtained from using a set of items behave as expected. You can only have construct
validity if there is content validity.

When checking construct validity, first the construct unidimensionality has to be assessed. A
measure must satisfy two conditions in order to be unidimensional. First, an empirical
indicator must be significantly associated with an underlying latent variable. Second, it can be
associated with one and only the latent variable.
The second step in checking construct validity is assessing the convergent and discriminant
validity. Convergent validity refers to the degree to which multiple attempts to measure the
same concept are in agreement. Discriminant validity refers to the degree to which measures
of different concepts are distinct.

Criterion-related validity is when an instrument is intended to perform a prediction function,

validity depends entirely on how well the instrument correlates with what it is intended to
predict. Criterion-related validity is established when the measure differentiates subjects on a
criterion it is expected to predict. Establishing concurrent validity or predictive validity can do
this. Concurrent validity is established when the scale discriminates subjects that are known to
be different. Predictive validity is the ability of the measure to differentiate among subjects
with respect to a future criterion.

4.6 Advancements in theory formalization and survey design

If a construct is conceptualized as a reflective construct, then the latent variable determines
the construct indicators. If the construct is conceptualized as a formative construct, then the
indicators determine the latent variable. Researchers agree that the criteria used to evaluate
reflective constructs, should not apply to formative constructs.

Control variables are extraneous, not directly connected with the main theory and hypothesis,
but presumably able to cofound with them. The idea behind control variables is that they
essentially remove the variation explained by a predictor variable that has some readily
identifiable alternative explanation.

4.7 Survey execution

Fundamentally, in this phase the researcher has to repeat the pilot-testing activities with a
large sample:
- Approaching companies/respondents and collecting data
- Controlling and reducing the problems caused by non-respondents
- Performing data input and cleaning
- Treating missing data
- Assessing measurement quality
- Providing feedback to respondents

12
When data is missing three strategies can be used
- Deletion
- Replacement based on estimation
- Mean based
- Regression based
- Model-based
- Hot-deck imputation
- Model-based strategy

4.8 Data analysis and interpretation of results

In order to acquire knowledge about the characteristics and properties of the collected data,
some preliminary data analyses are conducted, usually before performing measurement
quality assessments or testing the hypotheses. Preliminary data analysis is performed by
checking central tendencies, dispersions, frequency distributions and correlations.

Significance tests can be grouped into two classes: parametric and non-parametric. Parametric
tests are considered more powerful because their data are typically derived from interval and
ratio measurements whose likelihood model is known, except from some parameters. Non-
parametric tests are used with nominal and ordinal data.

4.9 Information that should be included in articles

Readers should be able to:
- Understand what has been done
- Critically evaluate what the work has achieved
- Replicate the work or compare the results with similar studies

4.10 Ethical issues in survey research

 Read chapter 2

Chapter 5 – Case research

5.1 Introduction
Pure case research is based on analysis of a limited number of cases to which, at best, only
limited statistical analysis can be applied. There are several challenges in case research: it is
time consuming, it needs skilled interviewers and care is needed in drawing generalizable
conclusions. It can lead to new and creative insights and development of new theory, and have
high validity with practitioners.

5.2 When to use case research

Strengths of case study
- The phenomenon can be studies in its natural setting and meaningful, relevant theory
generated from the understanding gained through observing actual practice
- The case method allows the questions of why, what and how to be answered with a
relatively full understanding of the nature and complexity of the complete
phenomenon

13
 - The case method lends itself to early, exploratory investigations where the variables
are still unknown and the phenomenon not at all understood

Case studies can be used for different types of research

- Exploration: in the early stages of many research programmes exploration is needed
to develop research ideas and questions
- Theory building: theory can be considered as being made up of four components:
definitions of terms/variables, a domain, a set of relationships, and specific
predictions.
- Theory testing: often in conjunction with survey-based research in order to achieve
triangulation
- Theory elaboration/refinement:

In inductive research, propositions or frameworks are the outcome of the analysis. Deductive
research starts with hypotheses that are then empirically tested. In abductive research, an
extant theory is identified and explored through an iterative process eventually leading to
propositions or frameworks and extension to existing theory.

5.3 The research framework, constructs and questions

Case research is especially good in answering how and why questions. Such question can lead
to both theory testing and theory development. First a conceptual framework needs to be
created. The next step is to design the initial research question. Underlying the research
questions, it is likely to be one or more constructs; a priori specification of constructs is
valuable since: it permits researchers to measure constructs more accurately. If these
constructs prove important, then researchers have a firmer empirical grounding for the
emergent theory.

5.4 Choosing cases

For a given set of available resources, the fewer the case studies, the greater the opportunity
for depth of observation. Single in-depth case studies are often used in longitudinal research.
In embedded cases a single firm is studied, but this firm is involved in a number of different
cases or unit of analysis. But single cases have some limitations: first of all the
generalizability, the misjudging of a single event, and of exaggerating easily available data.
Multiple cases may reduce the depth of study when resource is constrained, but can both
augment external validity and help guard against observer bias.

Sampling cases involves two actions:

- Setting boundaries that define what you can study and connect directly to the
research question.
- Creating a sample frame to help uncover, confirm, or qualify the basic processes or
constructs that underpin the study
Each case should be selected so that it either predicts similar results (a literal replication) or
that it produces contrary results, but not for predictable reasons (theoretical replication).

When selecting cases it is also important to consider what are the parameters or factors that
define the population and are to be held constant across the sample. Controls rely on the

14
selection of the phenomena during the study’s experimental stage for their control. This
allows particular factors to be, in essence, ‘held constant’ while other are left free to vary as
they would naturally.

5.5 Developing research instruments and protocols

The primary source of data in case research is structured interviews, often backed up by
unstructured interviews and interactions. Other sources of data can include personal
observation, informal conversations, attendance at meetings and events, surveys administered
within the organization, collection of objective data and review of archival sources.

Reliability is the degree to which a measure is free from random error components. Validity is
the extent to which a measure only reflects the desired construct without contamination from
other systematically varying constructs.

A common format for the protocol is the funnel model. This start with broad and open-ended
questions first, and as the interview progresses, the questions become more specific and the
detailed questions come last.
Saturation occurs when the addition of respondent does not add any different insights or
theoretical advancement.

5.6 Conducting the field research

An underlying principle of case research in collection of data is that of triangulation, the use
and combination of different methods to study the same phenomenon.

 The chapter contains a very specific explanation of how to conduct field research, steps are
the same as in the assignment and in the slides.

5.7 Reliability and validity in case research

Construct validity is the extent to which we establish correct operational measures for the
concepts being studied. If the construct as measured can be differentiated from other
constructs, it also possesses discriminant validity. Internal validity is the extent to which we
can establish a causal relationship whereby certain conditions are shown to lead to other
conditions, as distinguished from spurious relationships. External validity is knowing whether
a study’s finding can be generalized beyond the immediate case study. Reliability is the extent
to which a study’s operations can be repeated, with the same results.

5.8 Data documentation and coding

The first step in documentation is a detailed write up of each site following the research
protocol structure. Another step is to the coding of the observations and data collected in the
field. It is important to try to reduce data into categories.
In coding, the first step is open coding - data are fragmented or taken apart. The next step is
axial coding – putting together the data in new ways. The final step is selective coding –
selecting a core category and relating it to other categories.

15
5.9 Analysis
There are two steps in analysis: analysis of within-case data, and searching for cross-case
patterns.

To start the analysis of within-case data, it is useful to create an array or display with all the
data. The overall idea is to become intimately familiar with each case as a stand-alone entity,
and to allow the unique patterns of each case to emerge before you seek to generalize across
cases. Now, the researcher can start looking for explanation and causality. Two options are
suggested: using a case dynamics matrix or making predictions and then use the case to test
them. A third method is using the causal network.

When searching for cross-case patterns, also a display should be created first. An effective
analytical approach is then to pick a group or category and to search within groups for
similarities or differences.

Chapter 9 – Modelling and simulation

9.1 Introduction
Model-based quantitative research is research where models of causal relationships between
control variables and performance variables are developed, analysed or tested. Quantitative
models are based on a set of variables that vary over a specific domain, while quantitative and
causal relationships have been defined between these variables.

9.2 Origins and development of model-based research in operations management

 This section explains how research in operations management has developed over the
years.

9.3 Methodologies in quantitative modelling

Quantitative model-based research can be classified as a rational knowledge generation
approach. It is based on the assumption that we can build objective models that explain (part
of) the behaviour of real-life operational processes or that can capture (part of) the decision
making problems that are face by managers in real-life operational processes. It is important
to stress that the relationships between the variables are describes as causal, meaning that it is
explicitly recognized that a change of value α in one variable will lead to a change of f(α) in
another variable.

The first class of research, axiomatic research, is primarily driven by the model itself. The
primary concern of the researcher is to obtain solutions within the defined model and make
sure that these solutions provide insights into the structure of the problem as defined within
the model. Axiomatic research produces knowledge about the behaviour of certain variables
in the model, based on assumptions about the behaviour of other variables in the model.
Axiomatic research is most of the times prescriptive; prescriptive research is primarily
interested in developing policies, strategies and actions to improve over the results available

16
in the existing literature, to find an optimal solution for a newly defined problem or to
compare various strategies for addressing a specific problem.

The second class of research, empirical research, is primarily driven by empirical findings and
measurements. The primary concern of the researcher is to ensure that there is a model fit
between observations and actions in reality and the model made of that reality. This type of
research can be both prescriptive and descriptive. Descriptive empirical research is interested
in creating a model that adequately describes the causal relationships that may exist in reality,
which leads to understanding of the processes going on.

The methodology for both types is very roughly

- Conceptualization
- Modelling
- Model solving
- Implementation

9.4 How to conduct quantitative research in operations management

Axiomatic descriptive research aims at creating managerial insights into the behaviour of
operational processes and their control. Axiomatic prescriptive research aims at developing
rules and tools for managerial decision making.

Axiomatic descriptive research start with a condensed description of the characteristics of the
operational process or the operational decision problem that is going to be studied. The
conceptual model description should use as much as possible concepts and terms that are
accepted as standards published in scientific literature.
The second phase in the research is the specification of the scientific model of the process or
problem. The scientific model must be presented in formal, mathematical term such that a
mathematical/numerical/simulation analysis or study can be performed. Analytical research
aimed at managerial insights does not try to provide the manager with a direct answer to his
question; instead, it construct an idealized model of the problem, ensuring that an answer for
the idealized problem can be found with the analytical methods and tools available.
At this point validity gets important, three ways to claim validity:
- Researcher may refer to scientifically accepted axiomatic descriptions of the system
studied that contain evident of the occurrence of the characteristics in real life
- Researcher may refer to published empirical research that shows the existence in real
life of the characteristics captured in the model
- Researcher may refer to earlier published empirical research that uses the same
modelling assumptions

 An extensive example can be found in this section

Decision rule oriented research is more ambitious than managerial insight orientated research.
Its goals is to provide the manager with decision rules that, when applied, achieve optimal or
near optimal performance with respect to some criterion function. Results in this field are
more difficult to obtain and modelling assumptions are often less strict. In particular, the

17
results must be obtained for realistic large-scale models, or must be easily scalable to a
realistic size.

A Slightly different approach is taken when the result is obtain with computer simulation
instead of mathematical analysis. This model is used when the model/problem is too complex
for formal mathematical analysis. This type of research generally leads to lower scientific
results than research using mathematical analysis, but the scientific relevance of the
process/problem studies can be much higher.
The first step in simulation study is to build a (very rich) conceptual model. Secondly, a
justification of the research method has to be given since the contribution to the scientific
results is often low. Then a justification of the solution or hypotheses to be tested have to be
given, since no related proof can be generated. After these justifications, the model can be
developed. The next step is the set-up of the experimental design. One of the last steps is to
perform a statistical analysis of the results. The last step is the interpretation of the results.

Model-based empirical research is concerned with either testing the (construct) validity of the
scientific models used in quantitative theoretical research, or with testing the usability and
performance of the problem solutions obtained from quantitative theoretical research, in real
life operational processes.
The first step in model-based empirical research is the identification of the basic assumptions
regarding the operational process underlying the theoretical models or problems. The second
step is to identify the type of operational process and the type of decision problem regarding
this operational process to which the basic assumption are assumed to apply. The third step is
that operational, objective criteria must be developed for deciding whether or not a real-life
operational process belongs to the class of operational processes considered and for
identifying the decision system is the operational process that represents the decision problem
considered. The fourth step is to derive, from the basic assumptions, hypotheses regarding the
behaviour of the operational process. The fifth step is to develop an objective way to do
measurement or to make the observations. The sixth step consists of applying the
measurement and observation systems, collecting and documenting the resulting data. The
seventh step is the processing of the data. The last step is to interpret the results.

18