Top of page
Skip to main content

Research philosophy and theory development: writing a meaningful methodology section

Many quantitative dissertations begin their methodology chapter with “this study adopts positivism” and never return to the statement. A philosophy section matters only when it explains your choice of question, evidence and analytical methods.

Research philosophy concerns beliefs and assumptions about developing knowledge: ontology asks what reality is, epistemology asks what counts as knowledge, and axiology considers the researcher's values. There is no universally best philosophy. Your assumptions should fit your approach to theory development, whether deductive, inductive or abductive, and the methods you actually use.

Start reading

1. Research philosophy and why it belongs in your dissertation

Saunders et al. (2023, p. 131) define research philosophy as a system of beliefs and assumptions about the development of knowledge. Researchers make assumptions about reality, knowledge and their own values at every stage, whether they recognise them or not. Those assumptions shape their understanding of the research question, their methods and their interpretation of results.

The book presents these decisions through the “research onion”. Philosophy and approaches to theory development form its outer layers, followed by methodological choice, strategy and time horizon; data collection and analysis techniques occupy the centre. The layers are related without mechanically determining each other: a straight line cannot be drawn from one philosophy to one technique (Saunders et al., 2023, pp. 128–129).

There is no single best philosophy

Management researchers do not agree on one philosophy. Saunders et al. (2023, p. 133) take a pluralist position: each offers a useful perspective. Recognise your own assumptions, name them and make sure your design is consistent with them.

Journal articles often leave readers to infer philosophy from the methods. In a dissertation, an explicit commitment and its methodological implications help readers understand the basis on which your conclusions should be judged (Saunders et al., 2023, p. 132).

2. Three types of assumption

Table 1. Paraphrased from Saunders et al. (2023, p. 136, Table 4.1).
AssumptionCentral questionObjectivist endSubjectivist end
OntologyWhat is the nature of reality?Real, external, a single ordered reality consisting of thingsConventional, socially constructed, multiple realities consisting of processes
EpistemologyWhat knowledge and evidence are acceptable?Natural-science assumptions; facts, numbers, observable phenomena and general lawsHumanities assumptions; opinions, narratives, texts, images, attributed meanings and particular contexts
AxiologyWhat role do values play?Value-free; researcher detachedValue-bound; researcher involved and reflexive

The book illustrates ontology through resistance to organisational change. If resistance is assumed to be harmful, researchers ask how to remove it. If it is understood as an ongoing process that can expose weaknesses in a change programme, they ask how it might be used constructively. Different assumptions about the same phenomenon generate different questions (Thomas and Hardy, as cited in Saunders et al., 2023, p. 134).

For axiology, the book suggests writing a statement of personal values relevant to the topic. Choosing face-to-face interviews rather than an anonymous questionnaire already expresses a preference for particular kinds of evidence (Saunders et al., 2023, p. 135).

3. Four paradigms for organisational analysis

Burrell and Morgan combine objective–subjective assumptions with regulation–radical change to distinguish four paradigms. Regulation seeks improvements within existing organisational arrangements; radical change questions those arrangements (as cited in Saunders et al., 2023, pp. 140–142).

Table 2. Synthesised from Saunders et al. (2023, Figure 4.3 and Section 4.3).
ParadigmPositionMain concern
FunctionalistObjective; regulationRational explanation and recommendations within existing structures; common in management research and often linked to positivism
InterpretiveSubjective; regulationHow people understand their world and attach meaning to organisational life
Radical structuralistObjective; radical changeStructural power and conflict; often linked to critical realism
Radical humanistSubjective; radical changePower and oppression examined through language and social construction

The terms paradigm, philosophy and methodology are sometimes used interchangeably. In this book, a paradigm is a set of basic assumptions shared by a research community; philosophy concerns knowledge, reality and values; methodology concerns how research should be conducted (Saunders et al., 2023, p. 143).

4. Five research philosophies

Table 3. Condensed and paraphrased from Saunders et al. (2023, pp. 145–146, Table 4.3).
PhilosophyOntologyEpistemologyAxiologyTypical methodology
PositivismReal, external, independent and orderedScientific method, measurable facts, general laws, causal explanation and predictionResearcher seeks detachment and neutralityUsually deductive, structured, large samples and quantitative measurement
Critical realismStratified empirical, actual and real domains; objective structuresHistorically situated knowledge and historical causal explanationValue-laden; seeks to minimise biasRetroduction, historical analysis and multiple forms of evidence
InterpretivismComplex, socially constructed through culture and language; multiple meaningsNarratives, perceptions and interpretations; new understandingResearcher involved and reflexiveUsually inductive, qualitative and in-depth with small samples
PostmodernismConventional and constructed through power relationsChallenges dominant accounts of truth and recovers marginalised voicesPower-sensitive and radically reflexiveDeconstruction of texts and realities; qualitative inquiry
PragmatismReality understood through the practical consequences of ideasKnowledge that supports successful action; problem-centred inquiryResearcher's values and doubts guide inquiryQuestion-led; mixed methods, multiple methods and action research

Positivism

Positivists seek observable, measurable facts and causal relationships that support law-like generalisations. Hypotheses are commonly developed from existing theory and tested. Positivism does not necessarily begin with theory: early positivists advocated induction, although contemporary positivist studies are predominantly deductive. Choosing a problem, questionnaire items or experimental conditions can still introduce values (Saunders et al., 2023, pp. 146–147).

Critical realism

Critical realism distinguishes what we observe from the wider reality producing it. In Bhaskar's stratified model, the empirical domain of observations lies within the actual domain of events, while underlying structures and causal mechanisms belong to the real domain. The book contrasts a direct-realist referee who calls events as they are with a critical-realist referee who calls them as they appear to that observer (Saunders et al., 2023, pp. 148–149).

Interpretivism

People create meanings, so interpretivists argue that social research should not simply imitate natural science. Phenomenology attends to lived experience, hermeneutics to texts and symbols, and symbolic interactionism to interaction. Researchers need empathy and should seek to understand participants' worlds from their perspectives (Saunders et al., 2023, pp. 149–150).

Postmodernism and pragmatism

Postmodernism examines language and power, deconstructing accepted realities to reveal excluded perspectives. Pragmatism starts with a practical problem and selects methods that yield relevant, credible evidence for action. The book cautions against choosing pragmatism merely to avoid understanding other philosophies (Saunders et al., 2023, pp. 150–153).

Does using SEM make a dissertation positivist?

A questionnaire study testing theory-derived hypotheses often fits positivism and deduction. A study that develops explanations through interviews and subsequently tests them may instead involve abductive reasoning and a pragmatic or critical-realist position. Describe what the research actually does rather than assigning a philosophy from the software used.

5. Deduction, induction and abduction

Drawing on Ketokivi and Mantere (2010), Saunders et al. (2023, pp. 153–154) distinguish three forms of reasoning. Deduction derives conclusions logically from theoretical premises: if valid premises are true, the conclusion follows. Induction contains a logical gap between observations and conclusions, which can be supported rather than guaranteed. Abduction begins with a surprising fact and seeks premises capable of explaining it.

Table 4. Paraphrased from Saunders et al. (2023, p. 155, Table 4.4). Philosophical associations are common patterns rather than fixed rules.
CriterionDeductionInductionAbduction
LogicTrue premises entail the conclusionKnown premises generate an untested conclusionKnown premises generate a testable conclusion
GeneralisationGeneral to particularParticular to generalMoves between particular and general
Use of dataEvaluates propositions and hypotheses derived from theoryExplores phenomena, identifies themes and develops a conceptual frameworkExplores, builds a framework and tests it with further evidence
TheoryTests or challengesGenerates and developsGenerates or modifies, combining existing theory
Common associationsPositivismInterpretivismCritical realism, postmodernism and pragmatism

Six stages of deduction

Develop a hypothesis

Propose ideas, premises or relationships between concepts that contribute to a theory.

Deduce testable propositions

Use the literature and the conditions under which the theory applies.

Examine the reasoning

Compare it with existing theory and ask whether it advances understanding.

Collect and analyse data

Operationalise concepts and evaluate the premises.

If the evidence does not fit

Reject or revise the theory and begin the process again.

If the evidence fits

The theory is corroborated rather than permanently proved.

These stages are attributed to Blaikie and Priest in Saunders et al. (2023, p. 156). Deduction commonly involves causal relationships, operationalisation and carefully sampled evidence for generalisation. The book's example of absence from work in one shop also illustrates the limits: findings about younger, shorter-tenured staff in that shop cannot establish a universal pattern.

Inductive researchers may interview staff and managers before developing theory. Abductive researchers move between theory and evidence, exploring, building and testing with existing and new data. Critical realists often use retroduction to emphasise reasoning backwards towards mechanisms that could have produced the observed phenomenon (Saunders et al., 2023, pp. 157–158).

6. Choosing an appropriate approach

Understanding your approach helps you make informed design decisions, recognise suitable strategies and adapt when access or initial understanding is limited (Easterby-Smith et al., as cited in Saunders et al., 2023, p. 158). Consider:

  • Philosophy: positivism commonly favours deduction and interpretivism induction; other positions may favour abduction.
  • Literature: established literature can support deductive hypotheses; unfamiliar topics may call for induction; evidence from other settings but little from your own may motivate abduction.
  • Time: deduction is often easier to schedule; induction and abduction may take longer and may not produce clear patterns.
  • Your preferences and tolerance for uncertainty: these are also legitimate considerations.

The boundary between deduction and induction is not rigid. Combining them can be beneficial, even when one approach predominates (Saunders et al., 2023, p. 158).

7. A reporting template, common mistakes and checklist

Illustrative paragraph for a quantitative survey dissertation

This study adopts a positivist philosophy. Ontologically, [concepts such as trust and intention to use] are treated as independent of the researcher and measurable with validated instruments. Epistemologically, acceptable knowledge consists of observable relationships evaluated using quantitative evidence from [sample]. Axiologically, the researcher seeks independence through a structured self-administered questionnaire. A deductive approach is used: hypotheses are derived from [theory] and evaluated using [analysis]. Findings support or fail to support these hypotheses within [context]; they do not establish applicability to every setting.

Table 5. Common writing problems, synthesised from the principles in Saunders et al. (2023).
Weak statementProblemBetter approach
“The study is positivist” without explanationNo assumptions or implicationsConnect ontology, epistemology and axiology with the design
Copying five definitions before choosing oneLength without relevanceExplain the selected position and reasons
Claiming deduction when the model emerged from interviewsMismatch with the actual reasoningExplain abductive or combined reasoning where appropriate
Choosing pragmatism only because methods are mixedA label replaces an argumentExplain why the problem requires multiple kinds of knowledge
“The hypothesis has been proved”Testing cannot establish permanent truthUse “supported” or “corroborated” within the study's scope

Checklist

  • State the chosen philosophy and its ontological, epistemological and axiological assumptions.
  • Align assumptions with the question, evidence and analysis.
  • Explain your approach to theory development and the reasons for it.
  • Describe the actual sequence of work rather than a generic template.
  • Use conclusions appropriate to your reasoning: corroboration rather than permanent proof.

Conclusion

The philosophy section is an explanation of why the evidence can address your question and why the findings should be assessed using particular criteria. A short account that connects assumptions with design is more persuasive than pages of copied definitions.

References

  • Ketokivi, M., & Mantere, S. (2010). Two strategies for inductive reasoning in organizational research. Academy of Management Review, 35(2), 315–333. https://doi.org/10.5465/amr.35.2.zok315
  • Saunders, M. N. K., Lewis, P., & Thornhill, A. (2023). Research methods for business students (9th ed.). Pearson. Pearson

This methodological overview draws on the listed scholarly sources. Other authors discussed through Saunders et al. (2023) are identified as secondary citations. The reporting paragraph is an illustrative template written for this article.