A note on doing an MBA dissertation project with 1 primary research method: a discussion with generative AI

A note on doing an MBA dissertation project with 1 primary research method: a discussion with generative AI

 

My query posted to generative AI: Some of my part-time MBA students express difficulties to employ more than 1 primary research method for their dissertation projects, notably on the difficulties from obtaining informed consents from organizational gatekeepers. My problem is, to do a part-time 4-month dissertation project, which could be the research weakness of a research methodology design that is pragmatic and evidence-based to reply on a single primary research method, e.g. a questionnaire survey on the researcher friends with 40 respondents? I understand that the students have no problems to carry out 1 to 2 secondary research methods, e.g. Internet-based.

A single primary method is not automatically a weakness in a 4-month part-time MBA dissertation; the real issue is whether the method matches the research question, produces credible evidence, and is transparently justified. The main weakness of a small questionnaire survey of 40 friends is not that it is single-method, but that it may suffer from convenience sampling, limited representativeness, and weak generalizability.

What the real weakness is

A pragmatic research stance does not require mixed methods for its own sake; it asks what method best addresses the problem in the time and access constraints you face. So, if students can only manage one primary method, that is acceptable provided they are honest about scope, sampling, and limitations.

For a survey of 40 respondents drawn from the researcher’s friends, the likely weaknesses are:

·        Sampling bias, because friends are unlikely to represent the target population well.

·        Limited statistical power, because 40 cases is often too small for strong inference unless the study is very exploratory.

·        Response bias and social desirability, since acquaintances may answer in ways that please the researcher.

·        Weak external validity, meaning the findings should be treated as indicative rather than broadly generalizable.

How to frame it well

For a short dissertation, a single-method design can still be evidence-based if the research question is modest and the conclusions are carefully bounded. A survey of 40 can work for exploratory MBA research, especially if the aim is to identify patterns, perceptions, or preliminary relationships rather than to claim population-level truth.

A stronger framing would be:

·        Use the survey as a pilot or exploratory study.

·        Define the population narrowly.

·        State that the sample is non-probability and convenience-based.

·        Emphasize that findings are suggestive, not definitive.

·        Supplement with 1–2 secondary sources, such as industry reports, company websites, annual reports, or publicly available databases.

Managing gatekeeper problems

If organizational gatekeepers are difficult to access, that is a common fieldwork challenge and not a sign of poor research design by itself. In such cases, secondary research and accessible primary data sources are often a realistic compromise, especially for a time-limited dissertation.

A practical solution is to align the research design with what can actually be completed ethically and on time. For example, if access to employees is blocked, a student might study publicly available organizational data, customer perceptions, or professional respondents reached through personal networks, while clearly acknowledging the access constraint.

Best dissertation approach

For a 4-month part-time MBA dissertation, the safest pragmatic design is often:

1.    One primary method with a clearly defined and feasible sample.

2.    One or two secondary methods for triangulation or contextual support.

3.    A narrow research question that does not require deep access to organizations.

In your example, a questionnaire with 40 respondents is not a fatal weakness, but it becomes weak if the dissertation claims broad generalizability or robust theory testing. A better claim is that it provides exploratory evidence from an accessible sample, useful for identifying trends and informing future research.

A thesis-safe wording

You can write something like this in the methodology section:

·        “Given the time constraints of a 4-month part-time dissertation and the limited access to organizational gatekeepers, the study adopts a pragmatic design using one primary survey method supplemented by secondary documentary evidence.”

·        “The use of a convenience sample limits generalizability, but the design is appropriate for an exploratory MBA-level investigation.”

That wording is academically defensible because it shows methodological realism, ethical awareness, and transparent limitation handling.

 

My question posted to generative AI: Kindly offer 3 practice advices in this case, to make the dissertation project with 1 primary research method, to achieve an adequate level of intellectual sophistication up to the expected MBA academic level?

Here are three practice advices that help a single-method MBA dissertation reach an intellectually sophisticated, defensible academic level.

1. Build a tight “chain of reasoning” from philosophy → question → design → safeguards

Do not just list methods; explicitly justify each choice as a logical consequence of your research philosophy and question. For a pragmatic dissertation:

·        Start by stating why pragmatism fits your topic (e.g., you need practical, evidence-based answers to a business problem rather than pure theory testing).

·        Show how pragmatism leads to your chosen approach (often deductive or abductive) and then to a single primary method that is the most feasible and effective way to answer your question.

·        Explain how this shapes your design: sampling strategy, questionnaire structure, variables, and analysis plan.

·        End with how you safeguard quality and ethics: validity, reliability, bias control, informed consent, data protection.

This chain of reasoning shows methodological sophistication: you are not “using a survey because it’s easy,” but “using this survey design because it is the most appropriate pragmatic solution to my question under these constraints”.

2. Narrow the research question and aim for exploratory depth, not broad generalization

A small sample (e.g., 40 respondents) is academically acceptable if the question is narrow and the claims are modest. Instead of trying to “prove” general truths, aim to:

·        Identify patterns, perceptions, or preliminary relationships in a specific context.

·        Explore how a concept operates in a particular setting (e.g., a specific industry, company type, or region in Hong Kong or Singapore).

·        Generate insights that are useful for practice and that can inform future research.

Do this by:

·        Defining a clear, bounded population (e.g., “mid-level managers in Hong Kong retail firms” rather than “all managers”).

·        Stating that the study is exploratory or pilot in nature, and that findings are indicative rather than definitive.

·        Using your single method to go deep on a few well-chosen variables or constructs, with strong theoretical grounding from the literature review.

This approach shows intellectual sophistication: you understand the limits of your design and are not overclaiming, which is a key marker of good research.

3. Strengthen rigor within the single method through careful design, analysis, and triangulation with secondary sources

Even with one primary method, you can demonstrate methodological rigor by:

·        Design rigor:

·        Use validated scales or constructs from existing literature rather than inventing items from scratch.

·        Ensure your questionnaire has clear instructions, logical flow, and avoids bias (e.g., leading questions, double negatives).

·        Explicitly describe how you will collect data (online platform, time window, recruitment channel) and how you ensure informed consent and anonymity.

·        Analysis rigor:

·        Use appropriate statistical techniques for your sample size (e.g., descriptive statistics, correlations, simple regression, or group comparisons if justified).

·        Check basic assumptions (e.g., normality, linearity) and report limitations transparently.

·        Present results clearly with tables and graphs, and interpret them in relation to your research questions and theory.

·        Triangulation with secondary methods:

·        Use 1–2 secondary methods (e.g., analysis of company reports, industry data, publicly available statistics, or policy documents) to contextualize and support your survey findings.

·        Compare and contrast survey results with existing evidence: do they align, contradict, or refine what the literature or industry reports suggest?.

This combination shows that while you rely on one primary method, your overall design is robust, evidence-based, and intellectually mature.

---

If you embed these three practices—(1) a clear chain of reasoning, (2) a narrow, exploratory question with bounded claims, and (3) high internal rigor and secondary triangulation—a single-method dissertation can fully meet MBA-level academic expectations, even with a small sample and limited gatekeeper access.

 

My question posted to Generative AI: Highlight 4 main characteristics of an exploratory research for part-time MBA students with regard to our discussion here.

Based on our discussion about single-method, time-constrained part-time MBA dissertations, here are 4 main characteristics of exploratory research:

1. Purpose is to clarify and map, not to prove or generalize

The goal is to understand "what is happening," "why it might be happening," or "how it works" in a new or under-studied area, rather than to confirm hypotheses or claim population-level truths.

For your students, this means framing their research as identifying patterns, perceptions, or preliminary relationships in a specific context (e.g., Hong Kong retail managers), not as producing definitive generalizable findings.

2. Flexible and open-ended design that evolves with insights

Exploratory research is not rigid; researchers adjust their approach based on what they discover as data collection progresses.

In practice, this allows part-time MBA students to refine their questionnaire, focus on key variables, or narrow their scope mid-project if early results show what is most relevant.

3. Foundational: builds stronger questions for future research

Exploratory studies are designed to lay the groundwork for more structured, descriptive, or experimental research later.

This is academically appropriate for a 4-month dissertation: students can honestly state that their findings are indicative and suggest directions for future research, which demonstrates intellectual maturity rather than weakness.

4. Cost-effective and feasible with limited sample sizes

Compared to large-scale studies, exploratory research is typically less expensive and more budget-friendly, making it suitable for students with limited access, time, and resources.

A questionnaire with 40 respondents from accessible networks is acceptable if the study is framed as exploratory, where small samples are standard and the focus is on depth of insight rather than statistical power.

---

These characteristics show that exploratory research is not a "weak" design but a legitimate, academically rigorous approach for MBA students facing real-world constraints like gatekeeper access and tight timelines.

 

*** MBA students need to discuss with their dissertation project supervisors to ensure that they are reasonably satisfied with their research methodology design. This note is an academic discussion paper with generative AI, not an official policy statement from the university.

A collection of blog notes on using chatgpt for research purpose.

Lecture note on case study and case study research: for MBA students

Lecture note on case study and case study research: for MBA students

 

My query posted to generative AI: Highlight 4 main ideas and suggest 4 practice advices, in the context of doing MBA dissertation projects on the following two topics: 1. case study (as a secondary research method) 2. case study research (Robert Yin's version) (as a research methodology with primary research methods to be used).

The reply: Clarifying the two “case” ideas for an MBA dissertation

You are really asking about two different things that often get muddled in MBA projects:

1.    “Case study” as secondary research – using published teaching cases / consultancy cases / company write‑ups as data that you analyse, without collecting new primary data.

2.    Yin-style “case study research” as a full methodology – a systematic empirical inquiry into a contemporary phenomenon in its real‑life context, where you collect primary data (interviews, documents, observations, etc.) following Yin’s design logic.

Below are 4 main ideas and 4 practice advices for each, tailored to MBA dissertations.

---

1. Case study as a secondary research method

Four main ideas

1.    Cases are pre‑existing, “frozen” data sources

You rely on published cases, company reports, analyst notes, news stories, etc., rather than generating new data.

Your methodological contribution is in how you interpret and synthesise them (e.g., theory-driven within-case and cross-case analysis), not in data collection.

2.    Best suited to theory-illustration and conceptual application

This approach works well when your aim is to apply or illustrate theories (e.g., corporate governance, strategy, leadership) to rich real-world examples, or to build tentative propositions.

It is weaker for claims about “lived experience” or causal mechanisms that would require direct access to actors.

3.    Document analysis becomes your core technique

Methodologically, you are doing a form of qualitative document analysis or qualitative content analysis.

The “case” is the unit of analysis; the documents are data sources about that unit.

4.    Validity rests on transparency and triangulation of sources

Because you cannot go back to the field, you strengthen trustworthiness by using multiple independent sources on the same case (reports, media, books, academic articles) and by being explicit about selection, limitations, and potential biases.

Four practice advices (MBA‑practical)

1.    Define the unit and boundaries very explicitly

Decide what exactly the case is (e.g., “DBS Bank’s digital transformation 2013–2020” vs “DBS as a whole bank”).

Write 1–2 clear sentences in your methodology chapter that specify time frame, geography, and organisational scope.

2.    Use a structured coding frame, even for documents

Derive a small set of categories from your theoretical framework (e.g., “board monitoring”, “shareholder engagement”, “risk controls”).

Then systematically code the case material against those categories, instead of narrating the story chronologically.

3.    Plan explicit cross‑case comparisons if you have more than one case

Even with secondary cases, pre‑decide a few comparison dimensions (e.g., ownership structure, regulation, performance trend) and create a simple cross‑case table in your findings chapter.

This stops the dissertation from becoming a series of isolated stories.

4.    Write a tight limitations section focused on secondary data

Acknowledge issues such as survivorship bias, PR spin in corporate material, and incomplete information.

Then link each limitation to how you have tried to mitigate it (triangulation, critical reading, explicit assumptions).

---

2. Yin-style case study research as a full methodology

Four main ideas (Yin’s perspective)

1.    Case study = empirical inquiry in real context

Yin defines case study research as an in‑depth investigation of a contemporary phenomenon within its real‑life context, especially when the boundaries between phenomenon and context are not clearly evident.

It is driven by “how” and “why” questions and is particularly useful for complex managerial phenomena in organisations.

2.    Design logic is central (not just “I did interviews”)

Yin emphasises case study design with components such as: research questions, propositions, unit of analysis (the case), logic linking data to propositions, and criteria for interpreting findings.

This means you justify why a single or multiple case, and how each case is theoretically chosen (replication logic).

3.    Multiple sources of evidence and triangulation

Yin advocates using several primary methods (e.g., interviews, direct and participant observation, documents, archival records, physical artefacts) to build a “chain of evidence”.

Triangulation is not just having many data types but deliberately using them to corroborate or challenge emerging explanations.

4.    Systematic analysis: pattern matching and explanation building

Yin recommends analytic strategies such as pattern matching (comparing empirical patterns to predicted ones), explanation building, time‑series analysis, and cross‑case synthesis.

Analysis is iterative: moving back and forth between data and theory to refine propositions.

Four practice advices (MBA‑practical, Yin‑style)

1.    Start from propositions linked to your conceptual framework

Convert your theoretical framework into a few tentative propositions or “what you expect to see” in the case (e.g., “Firms with higher board independence will show more proactive risk management”).

These propositions then guide your interview protocol, document selection, and pattern matching later.

2.    Be explicit about single vs multiple case logic

For a single case, justify it as a critical, extreme, revelatory, or longitudinal case (Yin’s categories).

For multiple cases, explain how each case serves as a literal or theoretical replication, and show a simple table of cases, selection criteria, and expected theoretical contribution.

3.    Develop and actually use a case study protocol

Yin strongly recommends a protocol covering: field procedures, case study questions, data collection plan, and outline for the case study report.

In an MBA dissertation, include an appendix with your protocol and then, in the methodology chapter, briefly explain how it helped maintain consistency and reliability.

4.    Maintain a case study database and chain of evidence

Even for a student project, organise your raw data (transcripts, notes, documents) and analysis artefacts (coding files, memos) so that an examiner could, in theory, trace how you moved from raw data to conclusions.

In the dissertation, describe this database and give at least one small example of how specific quotes/documents link to a particular finding.

 

4 academic references related to the discussion here, in Harvard reference format

Here are four solid academic references, formatted in a standard Harvard style and aligned with our discussion of (1) case study as secondary research and (2) Yin‑style case study research with primary data.

1.   Core Yin methodology text (must‑have for your dissertation):

Yin, R.K. (2014) Case study research: design and methods. 5th edn. Thousand Oaks, CA: Sage.

2.   Updated Yin applications text (links to practical MBA‑type uses):

Yin, R.K. (2018) Case study research and applications: design and methods. 6th edn. Thousand Oaks, CA: Sage.

3.   General case study methods overview (helpful for positioning Yin vs other approaches):

Tight, M. (2010) ‘The curious case of case study: a viewpoint’, International Journal of Social Research Methodology, 13(4), pp. 329–339.

4.   Case study as (often) secondary-document based research (good to justify secondary case work):

Gerring, J. (2007) Case study research: principles and practices. Cambridge: Cambridge University Press.