Summary by James R. Martin
Research Methods Main Page | Decision Theory Main Page
This monograph is part of a larger project developed jointly by the National Association of Accountants and the Society of Industrial Accountants of Canada. It deals with research methodology and research on business decisions. The study is intended for practitioners as well as academics and can serve as a primer on methodology. The purpose of Chapter 1 is to provide a framework for finding and solving research problems. Chapter 2 extends the framework to address the criteria for choosing and evaluating different research methodologies. Chapter 3 is dedicated to business decision problems and Chapter 4 provides a summary of eleven examples of research on decision making.
Chapter 1 - A Framework For Methodology
Need for a Framework
This chapter begins with some questions that are
difficult to answer to the satisfaction of everyone, such as the following.
What is research?
Who is a researcher?
Where do research problems originate?
How should the researcher go about solving his or her problem?
Where should the researcher go for information?
How does the researcher know if and when the problem is solved?
This monograph provides a framework to aid in designing and conducting research in a systemic way.
A Framework for Research Methodology
The authors define research methodology as "the strategy or architectural design by which the researcher maps out an approach to problem-finding or problem-solving." Their framework has six parts as indicated below. Parts I and II are related to problem finding, while parts III through VI relate to problem solving.
I. Problem FindingFormal Approaches:
Analog method - Using knowledge gained in one problem area to formulate
a research question in a related area.
Renovation - Replacing defective components
of a theory. This requires training
in structural analysis, i.e., the ability to diagnose a system to find the
Dialectic method - Developing counter-plans
for challenging, refining or deposing
an existing theory. Price level adjustments to financial statements
is mentioned as an example.
Extrapolation - Extending stable
trends into the future and posing questions related
to predicted outcomes.
Teratological analysis - Similar to
extrapolation, but involves unstable trends, or
unexpected futures. Radical what if questions related to
Morphology - A method of formalizing
combinational possibilities in complex
Decomposition - Breaking problems into
Dichotomy - separating a problem into questions that can be
answered yes or no. Examples include flow charting, network
analysis, algorithms, and computer programming.
Aggregation - Aggregate research findings
and apply them to more complex
problems. Example - Applying input-output analysis, utility theory
and motivational theory to measuring managerial performance.
Conjecture - Based on intuition or hunch.
Phenomenology - Research problems
created by a phenomenon, e.g., computers,
conglomerates, environmentalism, etc.
Consensus - Group agreement that there is a
problem. The FASB approach is
given as an example.
Experiential - Research to find ways to
avoid a reoccurrence of a problem, e.g.,
lawsuits, liquidity problem, customer complaints.
II. Research ProblemWhat is a research problem? A research problem:
in defining the research problem include:
1. Mislabeling the problem. Confusing descriptive with normative. Descriptive
questions relate to empirical research, observing what is. Normative questions
involve opinion research, i.e., what should be. Applying past or future problems
to the present. Using students to study the behavior of business people.
2. Defining a problem that is unsolvable such as how businessmen solve problems.
III. Research Mode
A researcher needs to know which mode they are in to understand the consequences of the methods.Induction: the process of generating theory - from specific facts to generalizations.
Deduction: the process of
testing theory- testing a hypothesis against theory - from
general to specific. Relevant questions include; will questions such as
will it work?, is questions such as is it a good idea?, set-response
questions such as how do people feel?, task-response questions such
as did he do a good job, and if questions such as will this occur if we
Research Strategies, Domains and Techniques
Research methodology includes a set of strategies, domains and techniques used in generating or testing theory or problem solving.
IV. Research Strategy
Research strategy refers to the nature of the data, the process of obtaining the data and the way we go about generating or testing theory. Research strategies include: opinion, empirical, archival and analytic.
Opinion - Includes views, judgments
or appraisals of individuals or groups.
There is no direct observation of the facts. Techniques Include
questionnaires, opinion polls, interviews, delphi (for groups where
opinions are stated, if no consensus, then circulate and include
invitation to revise, continue until a consensus is reached or appears
infeasible) and brainstorming.
Empirical - Based on observation or
experience. Direct access to the facts.
The researcher either experiences the phenomena or is an eye-witness
to the events.
Case - where there is no attempt to structure the research, i.e., no
experiments, design or control, field and laboratory.
Field - includes experimental design, but no control. Has a structure
and method of collecting and analyzing data. Time and motion
studies given as an example.
Laboratory - Includes experimental design and control. Simulation is
The authors support their definition of empirical research strategy with the following comments. "The term empirical has acquired an ambiguity in the accounting literature. In its broadest sense it has come to mean the use of "real world" data in the application of research, and as such would apply to opinion or recorded data which purports to be of real-world origin. But this fails to distinguish between first and second-hand evidence, or between impressions and reality. Because these distinctions are important, as we will see in the course of this study, we favor a stricter construction of the term. Our definition, we believe, is substantiated widely in the methodology of science literature." (See p. 24).
Archival - Based on recorded facts,
but no direct observation of the facts.
Primary - original documents, e.g., invoices, purchase orders.
Secondary - journals, ledgers, financial reports.
Physical - fingerprints, footprints, dog tracks.
Techniques include scanning, observation, sampling and content analysis, i.e.,
a formal technique for evaluating written or oral communication. Erosion and
accretion measures are used for research in the physical domain.
Analytic - Based on the internal
logic of the researcher.
Inductive argument - from specific to general, generalizing from
incomplete data or drawing conclusions from inadequate evidence.
Rules are: agreement, difference, joint agreement and difference,
residues, and concomitant variations (i.e., correlated).
Deductive argument - from general to specific. Syllogism - a form of
reasoning where a conclusion is drawn from two premises.
Formal techniques include: mathematical logic, mathematical modeling,
formal organization techniques (flowcharting, network analysis,
decision trees, algorithms, and heuristics).
Informal techniques include: philosophical argument, the scenario,
the dialectic (fully developed argument and counter argument), the
dichotomous method (series of yes and no questions) and the
Teratological method (formulating hypotheses beyond the limits of
Summary: Problem - Mode - Strategy - Domain - Techniques
These concepts are summarized in the illustration below adapted from Exhibit 1 in the monograph.
What is Research?
Conceptually, research is defined as the orderly search for
truth. Operationally, research is:
1. an orderly investigation of a defined problem,
2. using an appropriate scientific method or methods,
3. to gather adequate and representative evidence,
4. producing conclusions drawn from logical reasoning without bias,
5. that can be validated,
6. and yield general principles or laws that may be applied in similar conditions.
Artifacts of research form a symbolic and rational system of inquiry and the language of research. These artifacts include facts, experiences, data, concepts, constructs, hypotheses, conjectures, principles and laws. These terms are defined on page 29.
Steps in the Scientific Method
The following steps are designed to achieve reliability and
1. Recognition that knowledge stems from observations.
2. Define the problem. Why and what is to be achieved.
3. Formulate a research plan including appropriate strategies, domains and
techniques. The purpose of the plan should be directed towards hypothesis
construction (induction) or hypothesis testing (deduction).
4. Conduct inquiry in accordance with the plan.
5. State the outcome in explicit terms that may support or refute the hypothesis.
6. State the conclusions with sufficient support and clarity to establish what was
done, what was found, and what significance the findings may have, including the
relationships to the work of others in terms of similarities and differences.
Who is a Researcher?
A researcher is anyone who uses the scientific method.
Research and Creativity
Research does not eliminate the need for creativity. Research supports, but does not replace the intuition and judgment of decision makers.
Chapter 2 - Criteria for Selecting A Research Methodology
A flowchart for selecting a methodology is provided on page 34. It maps out the options and can be used to classify research projects.
Strengths and Deficiencies of Different Research Methods
|Suited for research on attitudes, futures research, can use large samples, easiest method to use, and lends itself to many types of data analysis. Herzberg's work on hygiene factors and motivators is mentioned as an example. See the Herzberg summary for an illustration similar to Exhibit 6.||Opinion is not fact, perceptions differ from
reality. Methodology is subject to bias in the survey instrument, and in
biases in the way people respond to questions. Opinions are unstable over
time, and group opinion is difficult to capture and analyze. See the Ethics
topic for more on bias.
|Based on facts, not opinions. Best suited to analyzing actual behavior, fact finding and seeking reality. Gets the researcher involved. Case and field studies provide a rich context for research. Laboratory studies provide the most control. Can use sophisticated techniques such as gaming and simulation.||In general, only a small number of situations can be studied in detail and there is limited ability to generalize from these isolated cases. It is restricted to the present and is the most time consuming method. Case research lacks experimental design and control. Field research has experimental design, but lacks control (See 8 steps below). Laboratory has experimental design and control, but observations may be deficient for several reasons (pp. 38-39).|
|Best suited for analysis of data, in documents, historical analysis, gathering hard evidence, extrapolation of trends, etc. Ability to access a large quantity of hard, often factual information.||Selective depositing - only certain things are recorded, e.g., the accomplishments of white males. Selective survival - much is lost, unpublished, revised etc. Selective retrieval - subject to bias and sampling errors. Filling in the gaps - editorializing. Biases of researchers. Skill deficiencies of the researcher (anachronism, false periodization, interminability, chronicism and staticism (p. 41.)|
|There is no need to search for additional data and analytic research is not limited by existing data. It provides the broadest scope for imagination and creativity. Best suited for the use of logic, philosophy, operation research techniques.||The most abused strategy and the most difficult to criticize. Requires a first rate mental ability that is rare. Can more readily be used to mislead. Often sloppy. It is subject to logical errors (see below), problems of semantics, etc. Temptation to focus on trivial and irrelevant problems. Can only create theory, never proof.|
Field Research - Eight Steps (p. 38)
1. Recognize and identify problems.
2. Explore the problem - gathering information and defining the problems specifics.
3. Formulate the research objectives.
4. Estimate the costs.
5. Define the population.
6. Select the sampling design.
7. Design the research implement.
8. Plan and execute the field operations.
Examples of poor logic in accounting Analytic research (p. 43)
1. Building a strong argument on a
2. Jumping from a few instances to generalizations.
3. Jumping from a generalization to a specific instance.
4. Ignoring other possible causes.
5. Ignoring other possible effects.
6. Circular reasoning.
7. Using an illogical train of thought.
Experimental Design and Control Criteria (p. 46)
Internal validity - Relates to the following questions:
1. Is the research design appropriate?
2. Does the design include all the important factors and relationships?
3. Have the independent variables been controlled to prevent bias?
4. Has randomization been used when possible to reduce systematic bias?
External validity - Relates to the problem of
generalization. Can the findings be applied
to other cases on a broad basis?
Broader Framework for Selecting and Evaluating Methodology
The authors say validity has been stressed to the exclusion of other important criteria and propose a broader framework.
1. Holism - ability to comprehend the
problem in totality.
2. Precision - degree of refinement of a problem for objective measurement.
3. Internal validity - ability to control the variables in the problem.
4. External validity - extent to which the findings are applicable to other similar
5. Quantitative - ability to define or analyze problems in quantitative terms.
6. Qualitative - ability to define and or analyze the problem in qualitative terms.
7. Relevance - value of the findings to significant areas of need and application.
8. Skill transfer - extent to which it is difficult to train and supervise research
The Need For Balance
The authors argue that we should use eclectic methodology. The practice of law is an example where all strategies are used to build a case such as opinions, eye-witness reports, lab evidence, library research, logic, philosophy, and cross-examination. A narrow focus limits the problems that we can study and excludes many problems including research on decision-making.
Research in accounting has been biased in favor of theoretical essays and testing hypotheses. We need more inductive research aimed at generating theory. We need to include greater attention to abstract problems such as decision processes that could yield benefits for the improvement of management practices.
Chapter 3 - The Scientific Study and Classification of Business Decision Problems
The authors make a distinction between the science of decision-making (Praxeology) and the art of decision-making. The art of decision-making cannot be studied or learned. The science of decision-making can be studied and learned, as any science by applying deductive and inductive analysis.
Exhibit 8 provides a summary (based on the work of Murdick and Ross) showing the development of scientific management from behavioral theory (human relations and social system) to management theory (experimental analysis and functional analysis) to management science (decision analysis, quantitative analysis and technology). The chapter includes a discussion of the benefits from each of the areas of research and provides a classification scheme for business decision problems.
Benefits of Behavioral Research
In this section the authors discuss the work of several authors including Hawthorne, Seashore, Van Zelst, Fiedler, Zander, Maslow, Mayo, Vroom, and Likert, etc.
The work of many of these authors has helped us understand the behavior of groups. Group control often neutralizes company control systems. Intra-group competition can have disastrous effects on group cohesiveness, while inter-group competition has the opposite effect. Formal systems are viewed by employees as propaganda. The grapevine is relied on for information. Organizations have unique cultural, ideological, sociological, attitudinal and technological environments.
Maslow's needs hierarchy has been useful in the area of motivation. Vroom's work (Performance = ability x motivation) revealed that motivation is more important for performance than ability. Likert's work on leadership produced four styles including: exploitive, benevolent, consultative and participative leaders.
Benefits of Management Theory Research
Fayol and others (e.g., Mooney and Reiley, Gulick,Terry, Koontz and O'Donnell) provided and expanded the functional approach to management. For example, Fayol indicated that management functions include: planning, organizing, staffing, directing and control. Harvard developed and promoted the experiential approach using case studies.
Benefits of Management Science Research
Developments in this area include decision analysis, quantitative analysis and management technology.
Barnard promoted the science of decision making.
Simon refuted the concept of the rational economic man, i.e., that a decision
maker can define the problem accurately, search for all possible solutions,
assess the benefits of each and choose the optimal solution. Simon introduced
the concepts of:
Bounded rationality - people have fragmented knowledge of consequences.
Means-ends analysis - defines a sequence of decision points with a limited set
Satisficing - people seek satisfactory rather than optimum solutions due to mental
limitations and information overload. Miller and Haire's work supports.
The concurrency of phases - phases in the decision process run concurrently,
rather than is fixed steps. Goals constantly change.
Quantitative methods developments have been continuous and include time and motion study, the Gantt chart, PERT, CPM, price level models, game theory, decision trees, program planning and budgeting systems (PPBS).
Management Science researchers have considered problems in many areas including inventory, allocation, queuing, sequencing, routing, replacement, competition, and search.
Classification of Business Decision Problems
A classification scheme appears in Exhibit 10 on
page 60. The classifications include:
Cataloging - identification only.
Grouping - Dichotomous - classify into one of two categories or multiple
Dimensioning - Attributes of two or more problems are juxtaposed to produce a
Ordering - Ordinal - establishes rank order.
Interval - determines distance of intervals and levels of complexity.
Ratio - items positioned in relation to scales.
Systemizing - Decomposing - breaking a problem into parts.
Network - connect elements of a system by cause.
Static reciprocity - mapping (static) cause and effect relationships.
Dynamic - mapping (dynamic) cause and effect relationships.
A Typology for Business Problems
Business problems are shown in a matrix of combinations related to the quality of data (objective or subjective) and the nature of the relationships among the decision variables (simple and complex). A flow chart classifying problems into these four categories (subjective-simple, subjective-complex, objective-simple and objective-complex) is presented in exhibit 21.
This chapter includes many examples and
illustrations not mentioned in this brief summary.
Chapter 4 - Concluding Remarks and Application
This chapter includes a summary of eleven examples of research on decision-making (Exhibit 25) and a flowchart for classifying methodology (Exhibit 26). The flowchart below was adapted from Exhibit 26.