Lecture note on causal research for housing studies dissertation project study

Lecture note on causal research for housing studies dissertation project study

 

In the context of study of research types, please clarify four main ideas of causal research and offer 4 main how-to advices on research methods practices in the context of Housing Studies dissertation project works.

Causal research asks whether a change in one factor produces a change in another, rather than merely being associated with it. In a Housing Studies dissertation, it is especially useful for evaluating the effects of a policy, redevelopment programme, rent intervention, transport improvement, or housing condition on residents and neighbourhoods.

Four ideas in causal research

1.    Specify a cause and an outcome clearly
Define the independent variable (the proposed cause or “treatment”) and dependent variable (the outcome).

o   Example: Does participation in a public-housing redevelopment programme (cause) improve residents’ housing satisfaction or mental well-being (outcomes)?
Operationalise each construct: how exactly will “participation,” “housing satisfaction,” and “well-being” be measured?

2.    Establish temporal order
The cause must occur before the outcome. A one-time survey that finds higher satisfaction among residents in newer estates cannot, on its own, show that the newer estate caused satisfaction: more satisfied or higher-income households may have selected into it. Pre-policy/post-policy data, residential histories, or longitudinal follow-up make causal claims more defensible.

3.    Construct the counterfactual or comparison
The central causal question is: what would have happened to the same residents, households, or neighbourhood if the intervention had not occurred? Since this cannot normally be observed directly, researchers approximate it with a credible comparison group, such as similar non-recipient households, comparable estates, or observations before the policy change. Housing research specifically stresses careful comparison with otherwise comparable households living in alternative housing circumstances.

4.    Address confounding and validity
A third factor may influence both the proposed cause and outcome. For example, income, household size, age, neighbourhood location, transport access, prior housing quality, and self-selection may affect both eligibility for subsidised housing and later well-being. Causal research attempts to reduce such rival explanations through random allocation where feasible, matching, statistical controls, a non-equivalent comparison group, or repeated observations over time. Internal validity is stronger when plausible alternative explanations are minimised; external validity concerns whether findings apply beyond the studied estates or communities.

Four practical recommendations

1.    Align question, design, and claim
Write one precise causal question, then choose a design that can support it. Avoid causal wording if your design is only descriptive.

o   Descriptive: “What are residents’ perceptions of estate renewal?”

o   Causal: “What is the effect of estate renewal on residents’ perceived safety?”
If the project has a cross-sectional survey only, use cautious phrasing such as “is associated with,” rather than “causes.” A stronger feasible MBA design may be a before–after comparison with a carefully selected comparison estate.

2.    Choose a feasible quasi-experimental design
Randomised experiments are usually impractical and ethically unsuitable for housing allocation or public policy. For a dissertation, consider:

o   Difference-in-differences: compare the change over time in an affected estate with the change in a similar unaffected estate.

o   Interrupted time series: analyse several observations before and after a policy or redevelopment event, such as monthly transaction prices, complaint rates, or vacancy rates.

o   Matching/propensity scores: match participating households with non-participants who have similar observable characteristics.

o   Regression discontinuity: use an eligibility threshold, such as an income or score cut-off, when households just above and below it are plausibly similar.
These designs try to approximate the counterfactual when random assignment is unavailable.

3.    Build a defensible sampling and data plan
Define the unit of analysis—individual resident, household, building, estate, district, or transaction—and ensure your sampling follows it. Collect baseline covariates that could confound the relationship: demographics, income band, tenure, household composition, prior residential location, housing condition, neighbourhood amenities, and policy eligibility.
For qualitative interviews, purposively include contrasting experiences—for example, elderly tenants, young families, owners, renters, relocated residents, and frontline housing staff. For quantitative surveys, report response rate, inclusion/exclusion criteria, missing-data treatment, and limitations to representativeness. Probability sampling supports population representation, while purposive sampling supports rich understanding from well-informed participants.

4.    Use mixed methods deliberately and ethically
A highly suitable Housing Studies approach is an explanatory mixed-methods design: first analyse survey, administrative, transaction, or policy data; then interview selected residents and stakeholders to explain the observed patterns. Qualitative evidence should clarify mechanisms—how redevelopment, allocation rules, estate management, stigma, commuting time, or social networks may produce outcomes—not simply provide attractive quotations. Housing research advocates combining quasi-experimental comparison with ethnographically informed accounts because quantitative data may estimate effects while qualitative data reveal lived experience and unintended consequences.
Obtain ethics approval where required; use informed consent, remove identifying details, protect interview recordings and location data, and take extra care when discussing tenancy insecurity, financial hardship, conflict with landlords, or personal housing circumstances. Ethical practice requires continuing reflection and transparency, not merely formal compliance.

Dissertation-ready illustration

Possible question:
“What was the effect of a major estate-renewal programme on residents’ perceived neighbourhood safety in Hong Kong?”

Practical design:

·        Treatment group: residents of the renewed estate.

·        Comparison group: residents of a demographically and locationally similar estate not renewed in the same period.

·        Quantitative evidence: comparable safety surveys or police/management incident records before and after renewal.

·        Analysis: difference-in-differences, controlling for resident composition and broader district-level trends.

·        Qualitative evidence: interviews with residents, estate managers, and community workers about lighting, access control, social cohesion, displacement, and possible reporting changes.

·        Claim: conclude that renewal is associated with an estimated change in safety outcomes only to the extent that comparison trends and confounding have been credibly addressed.

The key principle is modest but rigorous: causal research does not merely show that two housing variables move together; it designs evidence to make the proposed causal explanation more credible than competing explanations.