Sunday, 15 November 2026

A comparison of research method quality criteria in qualitative and quantitative research: for Housing Studies students

A comparison of research method quality criteria in qualitative and quantitative research: for Housing Studies students

 

In the context of Research Methods study for the Housing Studies program, compare the concepts of (a) "internal validity" of quantitative research and "credibility" of qualitative research and (b) "external validity" of quantitative research and "transferability" of qualitative research, notably on the differences of ways to achieve them.

In Housing Studies research, internal validity and credibility both ask whether a study’s findings are trustworthy within the studied setting; external validity and transferability both ask whether findings can be meaningfully applied beyond it. The key difference is that quantitative research usually seeks statistical generalization from a sample to a defined population, whereas qualitative research enables readers to judge whether its context-rich findings fit another housing setting.

(a) Internal validity vs. credibility

Aspect

Internal validity (quantitative)

Credibility (qualitative)

Core question

Did the intervention or explanatory variable truly cause the observed outcome, rather than another factor?

Do the findings accurately and plausibly represent participants’ experiences and the phenomenon studied?

Main threat

Confounding, selection bias, measurement error, reverse causation, history, maturation, and other extraneous influences.

Researcher misinterpretation, insufficient engagement with participants, bias, and weak correspondence between data and interpretations.

Way to achieve it

Research design controls: random assignment, control/comparison groups, pre–post measurement, matching, statistical controls, validated instruments, and clearly specified procedures.

Interpretive strategies: prolonged engagement, triangulation of sources/methods/theories, member checking, peer debriefing, reflexivity, and transparent analysis.

Housing example

In an experiment on whether a rent-support programme reduces arrears, random assignment and a comparison group help show that the programme—not tenants’ prior income or landlord type—produced the change.

In interviews with tenants in subdivided flats, repeated interviews, observations, multiple data sources, and asking participants to review interpretations help establish that the account reflects their lived housing experiences.

The parallel is useful but not exact: internal validity concerns causal attribution under a controlled design, while credibility concerns the “truth value” of qualitative interpretations. In both cases, the researcher must show that the reported result is not an artefact of the study procedure or the researcher’s assumptions.

(b) External validity vs. transferability

Aspect

External validity (quantitative)

Transferability (qualitative)

Core question

Can the sample-based estimate or causal effect be generalized to a wider population, settings, times, or measures?

Could the findings be applicable to another housing context, group, policy setting, or time?

Main logic

Statistical or population generalization: the sample must represent the target population and the study conditions must resemble the intended real-world conditions.

Contextual reasoning: the reader/user decides, from detailed contextual evidence, whether the findings fit their situation.

Way to achieve it

Probability or representative sampling, adequate sample size, replication across sites/populations, realistic field settings, and clear specification of the population and conditions to which results apply.

Thick or rich description of participants, setting, social relations, housing type, policy context, and study boundaries; the researcher supplies the evidence, but does not claim universal generalization.

Housing example

A survey using a probability sample of private-rental tenants across Hong Kong can support an estimate of the prevalence of overcrowding in that defined population.

A qualitative study of older residents’ experiences in one public-estate renewal project may help policymakers understand similar renewal contexts, provided the study describes the estate’s tenure mix, resident demographics, management arrangements, and local policy conditions.

External validity is strengthened by designing the study to represent the target population and by replication; transferability is strengthened by giving enough contextual detail for an external user to make a reasoned judgement about relevance. This creates an important trade-off: highly controlled experiments may strengthen internal validity but reduce external validity because their artificial conditions may not resemble everyday housing environments.

Practical implication

For a Housing Studies dissertation or research proposal, state the quality standard appropriate to your method rather than applying quantitative labels mechanically to qualitative work. A quantitative study should demonstrate how it minimizes bias and defines its generalization boundary; a qualitative study should demonstrate how it earns confidence in its interpretations and provides the contextual description needed for transferability.

A concise formulation is:

  • Internal validity / credibility: “Can I trust what this study says about the case or sample studied?”
  • External validity / transferability: “Can this study’s conclusion be used beyond the case or sample studied?”
  • Quantitative route: control, representativeness, measurement, and replication.
  • Qualitative route: engagement, triangulation, participant validation, reflexivity, and thick description.

A comparison of research method quality criteria in qualitative and quantitative research: for MBA students

A comparison of research method quality criteria in qualitative and quantitative research: for MBA students

 

In the context of Research Methods study for the MBA program, compare the concepts of (a) "internal validity" of quantitative research and "credibility" of qualitative research and (b) "external validity" of quantitative research and "transferability" of qualitative research, notably on the differences of ways to achieve them.

In an MBA Research Methods framework, internal validity and credibility both ask whether a study’s findings are trustworthy within the setting studied; external validity and transferability both ask whether findings can be applied beyond that setting. The key difference is that quantitative research seeks to establish these qualities through design, measurement, and statistical control, whereas qualitative research establishes them through rich contextual evidence and the reader’s judgement of fit.

(a) Internal validity vs credibility

Aspect

Internal validity (quantitative)

Credibility (qualitative)

Core question

Did the study genuinely demonstrate the claimed relationship or causal effect, rather than an artefact of bias, confounding, or measurement error?

Do the findings authentically represent participants’ experiences and the phenomenon as understood in its context?

Main threat

Confounding variables, selection bias, measurement error, researcher effects, history, and maturation.

Misinterpretation of participants’ meanings, researcher subjectivity, superficial engagement, or inaccurate reporting of participants’ views.

Typical standard

Statistical and experimental control; the study must rule out plausible alternative explanations.

Truth value or “believability” from the participants’ perspective; findings must be grounded in the data.

How to achieve internal validity

Quantitative researchers achieve internal validity mainly by designing away alternative explanations:

  • Use random assignment to treatment and control groups, where feasible, so groups are comparable before the intervention.
  • Use control or comparison groups and, where appropriate, pre-tests and post-tests.
  • Use random selection or clearly defined sampling criteria to reduce selection bias.
  • Standardise procedures, instruments, and measurement protocols to reduce error and inconsistency.
  • Measure and statistically control confounders, such as age, prior performance, firm size, or industry, through regression or matching.
  • Use valid and reliable measures of the intended constructs—for example, a validated employee-engagement scale rather than an ad hoc questionnaire.

MBA example: If a firm introduces a training programme and finds higher productivity among trained employees, internal validity requires evidence that the training—not prior ability, motivation, manager support, or simultaneous organisational changes—produced the improvement.

How to achieve credibility

Qualitative researchers achieve credibility mainly by demonstrating that interpretations are firmly grounded in participants’ accounts and the research context:

  • Conduct prolonged engagement and persistent observation in the field, allowing the researcher to understand the setting and detect inconsistencies.
  • Use triangulation: compare interviews, documents, observations, and different participant groups.
  • Conduct member checks by returning interpretations, themes, or summaries to participants for confirmation or correction.
  • Use peer debriefing or peer review, where colleagues challenge the researcher’s interpretations and methods.
  • Analyse negative or disconfirming cases rather than selecting only evidence that supports the emerging conclusion.
  • Practise reflexivity by documenting the researcher’s role, assumptions, and possible influence on data collection and interpretation.

MBA example: In a study of why frontline employees resist a new CRM system, credibility would be strengthened by interviewing employees and managers, observing system use, checking interpretations with participants, and explaining any cases where employees supported the system.

(b) External validity vs transferability

Aspect

External validity (quantitative)

Transferability (qualitative)

Core question

Can the numerical results be generalised to a defined wider population, setting, or time period?

Can the findings be meaningfully applied or adapted to another context judged similar by the reader?

Basis for application

Statistical representativeness and sampling logic.

Contextual similarity and thick description; the reader judges whether the setting is comparable.

Researcher’s role

The researcher designs the study to support generalisation to a specified population.

The researcher provides sufficient contextual information; the user of the research makes the transfer judgement.

Typical claim

“The result is likely to hold for Hong Kong SMEs with fewer than 50 employees.”

“These findings may be useful for managers in comparable service firms facing similar digital-transformation conditions.”

How to achieve external validity

Quantitative researchers achieve external validity mainly through representative sampling and realistic conditions:

  • Define the target population precisely, such as “Hong Kong SMEs in the retail sector.”
  • Use probability sampling—especially random sampling—so the sample represents the population and estimates can be generalised.
  • Use an adequately powered sample size to improve precision and reduce sampling error.
  • Conduct the study in settings, time periods, and participant groups that resemble the intended application context. 
  • Replicate the study across organisations, industries, regions, or time periods to test whether the relationship holds beyond the original sample.
  • Avoid excessive artificiality: highly controlled laboratory conditions may strengthen internal validity but weaken external validity because they may not resemble real organisational settings.

MBA example: A survey of 400 randomly selected Hong Kong SME owners can support stronger external validity for conclusions about Hong Kong SME owners than a convenience sample of 20 business acquaintances.

How to achieve transferability

Qualitative researchers achieve transferability mainly through thick, rich description, rather than claiming statistical generalisation:

  • Describe the research context in detail: organisation, industry, culture, participants, roles, time period, decision-making processes, and relevant events.
  • Provide thick descriptions of participants’ experiences, meanings, interactions, and the setting in which findings emerged.
  • Explain the sampling strategy and the boundaries of the case, such as the type of firm, employee group, and change programme studied.
  • Report the methods, assumptions, and conditions under which the findings were produced, enabling readers to assess similarity with their own context.pmc.ncbi.nlm.nih+1
  • Develop concepts or themes at an appropriate level of abstraction so that they can be considered in other settings without ignoring important contextual differences. 
  • Where relevant, provide “thick description” sufficient for another researcher or manager to judge whether the findings can be recreated or applied elsewhere.

MBA example: A qualitative study of resistance to CRM adoption in one Hong Kong logistics firm may be transferable to another logistics firm if the researcher clearly describes the firm’s size, management style, employee roles, technology maturity, and implementation process. The reader—not the researcher—decides whether those conditions are sufficiently similar.

Essential distinction

The deepest difference is who establishes applicability:

  • In quantitative research, the researcher aims to demonstrate generalisability through representative sampling, controlled procedures, and replication.
  • In qualitative research, the researcher provides a detailed account of the particular context, while the reader assesses whether the findings are transferable to another comparable situation.

Thus, internal validity is achieved by controlling rival explanations, while credibility is achieved by establishing authentic, participant-grounded interpretations. External validity is achieved by sampling and designing for a defined population, while transferability is achieved by thick description that enables contextual comparison.

A note on ways to improve the findings of a descriptive quantitative housing-studies dissertation

A note on ways to improve the findings of a descriptive quantitative housing-studies dissertation

 

Four main ways to improve the findings of a descriptive quantitative housing-studies dissertation are: strengthen the sample and measurement, improve data quality, present results more clearly and transparently, and interpret findings within their limitations. Descriptive research should systematically describe variables using measures such as frequencies, percentages, means, and ranges, without overstating causal claims.

1. Improve sampling and representativeness

Use a clearly defined target population—for example, public-housing tenants, private-rental tenants, or homeowners in a specified district—and justify the sampling strategy. A larger, more representative sample reduces the risk that the findings reflect only the views of an unrepresentative subgroup.

  • Define the sampling frame, eligibility criteria, and geographic/time boundary.
  • Report the response rate, sample characteristics, and any differences between respondents and the wider population.
  • Where possible, compare your sample with official housing statistics to assess representativeness.
  • Avoid claiming that findings apply to all Hong Kong residents if the sample is limited to one estate, district, or housing type.

Clear study-unit selection and precise variable measurement are central to meaningful descriptive research.

2. Strengthen measurement and operational definitions

Make each housing concept measurable and consistent. For example, “housing affordability” could be operationalised as the ratio of housing cost to household income, while “housing quality” could be measured through validated indicators such as overcrowding, maintenance condition, and access to facilities.

  • Use established survey items or official indicators where available.
  • Pilot-test the questionnaire to identify unclear wording, cultural issues, and response problems.
  • Report how each variable was coded, including categories, units, and treatment of “don’t know” or refused responses.
  • Check reliability where applicable, such as internal consistency for multi-item scales.

Descriptive findings become more credible when the variables are carefully defined and measured rather than treated as vague concepts.

3. Improve data cleaning and analysis

Before reporting findings, document how the dataset was cleaned and analysed. This makes the results more transparent, reproducible, and defensible in a dissertation viva or examination.

  • Check for duplicates, implausible values, missing data, and inconsistent responses.
  • State how missing data were handled and explain why the approach does not seriously distort the findings.
  • Use appropriate descriptive statistics: frequencies and percentages for categorical variables; means, medians, standard deviations, and ranges for continuous variables.
  • Use cross-tabulations, charts, and subgroup comparisons to reveal patterns, such as affordability problems by tenure, household income, or district.
  • Report the software, procedures, and assumptions used; choose the simplest statistical method that answers the research question.

4. Present and interpret findings carefully

Present results in a way that directly answers the research questions, using clear tables, figures, and concise narrative explanations. Descriptive statistics are useful not only for describing the sample but also for transparency, checking distributions, comparing groups, and situating results in the wider housing literature.

  • Label every table and figure fully, including sample size, time period, source, and definitions.
  • Report both visual patterns and exact statistics; do not rely only on a chart.
  • Distinguish description from causation: for example, show that renters report higher affordability stress, but do not claim tenure causes that stress without an appropriate research design.
  • Acknowledge limitations such as sampling bias, self-report bias, missing data, and the cross-sectional nature of the study.
  • Link findings back to housing theory, policy debates, and prior studies, while avoiding unsupported generalisations.

Dissertation example

If your project examines housing affordability in Hong Kong, you could improve the findings by:

1.    Sampling tenants across several districts rather than one estate.

2.    Defining affordability precisely as gross rent or mortgage payments as a percentage of household income.

3.    Cleaning the data, documenting missing responses, and reporting medians as well as means because income distributions are often skewed.

4.    Presenting affordability stress by tenure, household size, and income group, then explaining the patterns without claiming that one factor directly causes another.

Together, these four improvements make the dissertation’s descriptive findings more accurate, transparent, credible, and useful for housing policy discussion. Quantitative research quality is commonly assessed through internal validity, external validity, reliability, and objectivity.