Research Problems, Design and Methodology

Introduction

Real estate decisions are often presented as practical questions: Should a developer acquire a site? Which tenant segment should a shopping centre target? Will a rent incentive increase occupancy? Are housing prices being driven by genuine demand or speculative activity? These questions are managerial, but they are not yet research problems. A research problem requires a precise statement of what is unknown, why the uncertainty matters, what evidence is needed, and how that evidence will be generated and evaluated.

In real estate market research, methodological choices are especially consequential because markets are spatially differentiated, heterogeneous, regulated, cyclical and influenced by participants with unequal information. A study of office rents, for example, may be affected by building quality, location, lease structure, transport access, economic growth, planning restrictions and expectations about future demand. Good methodology does not eliminate this complexity; it makes the complexity manageable, visible and analytically defensible.

This chapter explains how to convert managerial concerns into researchable problems, objectives, questions and hypotheses. It then examines research boundaries, paradigms, design choices, sampling, measurement, data collection, validity, reliability, reproducibility and research ethics. The emphasis is on postgraduate-level judgement: selecting a design because it fits the decision and causal logic, rather than because it is familiar or convenient.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish a managerial decision problem from a research problem.
2. Formulate research objectives, questions, hypotheses and study boundaries.
3. Select among exploratory, descriptive, causal, cross-sectional, longitudinal and mixed-method designs.
4. Explain the roles of concepts, constructs, variables, indicators and operational definitions.
5. Design defensible sampling and data-collection procedures for real estate studies.
6. Evaluate validity, reliability, credibility and reproducibility.
7. Identify ethical, governance and regulatory risks in property-market research.
8. Develop a coherent research protocol for a real estate management decision.

1. From Managerial Concern to Research Problem

A managerial problem concerns an action or decision. A research problem concerns a gap in knowledge that prevents an informed decision. The two are related but not identical.

For example:

• Managerial concern: “Occupancy in our suburban retail centre is declining.”
• Weak research problem: “We need to study occupancy.”
• Strong research problem: “It is not known whether declining occupancy is primarily associated with tenant-mix mismatch, reduced customer accessibility, online substitution, or lease-pricing decisions; consequently, management lacks evidence for selecting an appropriate intervention.”

A well-formed research problem normally identifies four elements:

1. Context: the market, asset class, geography and time period.
2. Phenomenon: the behaviour, relationship or process requiring explanation.
3. Knowledge gap: what is unknown, disputed or inadequately measured.
4. Decision relevance: how resolving the uncertainty may improve action.

A problem statement should not predetermine the answer. “Prove that flexible offices increase demand” is advocacy, not neutral research. “Assess whether and under what conditions flexible-office provision is associated with tenant demand” permits evidence to support, qualify or reject the proposition.

1.1 The problem-conversion chain

A useful sequence is:

Managerial situation → research problem → purpose statement → objectives → research questions → hypotheses or propositions → design → evidence → conclusions.

Each stage should remain logically connected. If the objective is to estimate market size, a qualitative design alone is insufficient. If the objective is to understand how households interpret housing security, a purely transactional database may be insufficient.

1.2 Research objectives

Objectives state what the study will accomplish. They should be specific, observable and feasible. Common verbs include identify, estimate, compare, assess, explain, test, evaluate, model and interpret.

A broad objective might be: “To assess the determinants of apartment rental choice among young professionals in City X.” Specific objectives could be to:

• identify the attributes considered in rental choice;
• measure the relative importance of rent, accessibility, amenities and security;
• compare preferences across income and household groups; and
• test whether commuting time is associated with willingness to pay.

Objectives should not be confused with activities. “Conduct 300 surveys” describes a method; “estimate the association between transport access and rental preference” describes an objective.

1.3 Research questions and hypotheses

A research question expresses what the study seeks to answer. A hypothesis is a testable proposition about relationships or differences.

For example:

• Research question: How does proximity to rapid transit influence apartment rents, controlling for unit size and building age?
• Null hypothesis: After controlling for relevant property characteristics, transit proximity has no association with apartment rents.
• Alternative hypothesis: After controlling for relevant property characteristics, transit proximity is associated with apartment rents.

A directional hypothesis may state that rents are higher nearer to transit, but direction should be specified only when theory and prior evidence justify it. Hypotheses must identify units of analysis and variables. “Location affects value” is too vague because location and value require operational definitions.

2. Scope, Boundaries and the Unit of Analysis

Research boundaries protect a study from becoming unmanageable and prevent conclusions from exceeding the evidence. A protocol should specify:

• Geographic boundary: neighbourhood, municipality, metropolitan region or national market.
• Asset boundary: office, retail, industrial, residential, land or mixed-use property.
• Population boundary: owners, tenants, investors, valuers, households, developers or transactions.
• Time boundary: a particular quarter, market cycle or multi-year period.
• Conceptual boundary: which determinants and outcomes are included or excluded.
• Decision boundary: the decision to which findings will be applied.

The unit of analysis is the entity about which conclusions are made. It may be a property, transaction, household, firm, neighbourhood, lease or time period. Confusing the unit of analysis creates serious errors. Data collected from individual tenants cannot automatically justify claims about all buildings; observations of neighbourhood averages cannot establish individual household behaviour.

The unit of observation is the entity from which data are collected. A study may observe leases but analyse buildings by aggregating leases. The distinction should be stated explicitly.

2.1 Constructs, variables and operationalisation

A construct is an abstract concept, such as perceived safety, investment attractiveness or neighbourhood vitality. A variable is a measurable representation that can take different values. An indicator is an observed measure used to represent a construct.

For example, “neighbourhood vitality” might be operationalised using pedestrian counts, retail vacancy, business diversity and evening activity. No single indicator perfectly represents the construct; operationalisation is therefore a theoretical and methodological decision.

An operational definition should state:

1. what is being measured;
2. how it will be measured;
3. the measurement scale or unit;
4. when and where it will be measured; and
5. how missing, unusual or conflicting observations will be treated.

Measurement scales include nominal categories, ordinal rankings, interval scores and ratio measures. Market rent in currency per square metre is generally a ratio measure, while a five-point satisfaction response is ordinal, even if researchers sometimes analyse it approximately as continuous.

3. Research Paradigms and Logic of Inquiry

A research paradigm consists of assumptions about reality, knowledge and appropriate methods. Real estate research commonly draws on several traditions.

Post-positivist research treats market relationships as real but imperfectly observed. It emphasises measurement, uncertainty, comparison and statistical testing. Hedonic price analysis and quasi-experimental policy evaluation often reflect this orientation.

Interpretivist research examines how actors construct meaning and interpret markets. Interviews with displaced households, studies of investor expectations and research on neighbourhood identity may use this approach.

Pragmatic research selects methods according to the decision and question rather than committing exclusively to one philosophical tradition. It is particularly useful for mixed-method real estate studies in which numerical estimates and stakeholder interpretations are both necessary.

Paradigms should not be treated as labels added after data collection. They influence the formulation of questions, the role of theory, the relationship between researcher and participants, and the standards used to judge evidence.

4. Major Research Designs

4.1 Exploratory design

Exploratory research is appropriate when the problem is poorly understood, concepts are unclear or relevant variables have not been identified. Typical methods include literature review, expert interviews, focus groups, observation, document analysis and preliminary data mining.

In real estate, exploration may reveal that developers define “demand” differently from planners, or that tenants distinguish nominal rent from total occupancy cost. Exploratory findings usually generate concepts, propositions or survey items; they do not normally establish population estimates or causal effects.

4.2 Descriptive design

Descriptive research answers questions such as: What is the vacancy rate? Which tenant segments occupy the market? How are rents distributed by submarket? Descriptive studies may use surveys, transaction records, listing databases, spatial data and administrative sources.

A descriptive design must address measurement quality and coverage. A listing portal may describe asking rents rather than achieved rents. A vacancy estimate based only on advertised units may omit informal or unadvertised stock.

4.3 Causal or explanatory design

Causal research asks whether a change in one factor produces a change in an outcome, and under what conditions. The central challenge is distinguishing causation from correlation. A new transit station may be associated with rising property values because of the station, but also because investment and population growth occurred in the same area.

The strongest design, a randomised experiment, is uncommon for many property-market questions. Researchers therefore use quasi-experiments, natural experiments, longitudinal models, matched comparisons and statistical controls. Causal claims require attention to:

• temporal precedence: the proposed cause occurs before the outcome;
• covariation: the variables change together;
• alternative explanations: competing causes are addressed; and
• a defensible mechanism explaining why the effect should occur.

Statistical adjustment can reduce confounding but cannot guarantee causal identification. A model with many control variables is not automatically causal.

4.4 Cross-sectional design

A cross-sectional study collects data at one point or within a relatively short period. It is efficient for describing market conditions, comparing groups and examining associations. However, it is weak for establishing temporal order and may be vulnerable to period-specific events.

A cross-sectional survey of tenants can compare satisfaction across buildings, but it cannot confidently show that a service intervention caused later retention unless change over time is observed or a credible comparison is available.

4.5 Longitudinal design

Longitudinal research observes units repeatedly over time. Panel studies follow the same properties, households or firms; repeated cross-sectional studies sample different units from the same population at successive times; cohort studies follow a defined group.

Longitudinal designs help analyse market cycles, rent adjustment, displacement, development outcomes and policy effects. They require consistent definitions, stable identifiers and procedures for handling attrition, renovations, demolitions and changes in data sources.

4.6 Mixed-method design

Mixed-method research integrates qualitative and quantitative evidence in one study. Common structures include:

• Sequential exploratory: interviews generate concepts or survey items, followed by a quantitative study.
• Sequential explanatory: quantitative results are followed by interviews to explain unexpected patterns.
• Convergent: qualitative and quantitative data are collected in parallel and compared.
• Embedded: one method provides a supporting role within a primary design.

Mixed methods are not simply the addition of interviews to a survey. Integration must occur through sampling, instrument development, analysis or interpretation. Contradictory findings should be examined rather than concealed.

5. Sampling and Data Sources

The target population is the full set of units to which the study intends to generalise. A sampling frame is the operational list or procedure from which units are selected. These are not always identical. A commercial property database may omit small informal landlords and therefore underrepresent the target population.

Probability sampling gives units a known, non-zero chance of selection. Simple random, systematic, stratified and cluster sampling are common forms. Stratification can improve representation of important submarkets, such as asset class, location or building grade.

Non-probability methods include purposive, quota, convenience and snowball sampling. They may be appropriate for expert interviews or hard-to-reach populations, but the researcher should not present a convenience sample as statistically representative without justification.

For a simple proportion estimate, a commonly used initial sample-size expression is:

Here, n₀ is the preliminary sample size, Z is the standard-normal critical value corresponding to the chosen confidence level, p is the anticipated population proportion, and e is the desired margin of error. The expression assumes a probability sample, independent observations and a sufficiently large population. When no prior estimate of p exists, p = 0.5 produces the largest required sample for a given Z and e. Finite-population correction, design effects, non-response and subgroup analysis may require a larger or adjusted sample.

For example, with 95% confidence, Z ≈ 1.96, p = 0.5 and e = 0.05:

The preliminary requirement is therefore rounded upward to 385 completed observations. This is not a universal answer: a flawed sampling frame or severe non-response can make 10,000 observations less useful than a carefully designed smaller sample.

5.1 Primary and secondary data

Primary data are collected for the current study, such as interviews, surveys, inspections, experiments or field observations. Secondary data already exist, including land registries, planning records, transaction databases, census data, listings, satellite imagery and company reports.

Secondary data require a provenance assessment: who collected them, for what purpose, using which definitions, with what coverage and revision history? Researchers should distinguish asking price from achieved price, nominal from real rent, and stock measures from flow measures.

6. Data Collection and Analysis Planning

A research protocol should specify instruments, recruitment, field procedures, coding rules, quality checks and secure storage before data collection begins. Pilot testing can reveal ambiguous questions, unrealistic interview durations, missing response options and unreliable sensors or software.

Survey questions should be neutral, specific and limited to one idea. “How satisfied are you with the rent and maintenance service?” is double-barrelled because rent and maintenance may receive different evaluations. Recall periods should be realistic, and sensitive questions should be justified.

Qualitative analysis may involve transcription, familiarisation, coding, category development, comparison and interpretation. Codes should be defined sufficiently for consistent application. Quantitative analysis may involve descriptive statistics, cross-tabulation, regression, spatial analysis or time-series methods, but the method must match measurement level and design.

A simple linear model may be used to represent an association between rent and property characteristics:

Here, Rᵢ is observed rent for property i, Aᵢ may represent floor area, Tᵢ transit accessibility, Gᵢ building grade, β₀ is the intercept, β₁, β₂ and β₃ are coefficients, and εᵢ represents unobserved influences and measurement error. The model assumes a correctly specified functional form, appropriately measured variables and an error structure suitable for the chosen estimator. A coefficient is an adjusted association under the model; it is not automatically a causal effect.

7. Validity, Reliability and Reproducibility

Validity concerns whether a study supports the interpretation made from its evidence. Reliability concerns consistency of measurement. A measure can be reliable but invalid: a faulty instrument may produce the same incorrect reading repeatedly.

Important forms of validity include:

• Construct validity: the measure represents the intended concept.
• Content validity: important dimensions of the concept are covered.
• Internal validity: the design supports the stated explanation within the study.
• External validity: findings can reasonably be transferred or generalised to other settings.
• Ecological validity: findings reflect conditions in the real-world context.

In qualitative research, parallel concepts include credibility, dependability, confirmability and transferability. Credibility may be strengthened through prolonged engagement, triangulation, participant feedback and attention to negative cases. These practices do not mean that participants must approve every interpretation; rather, they make the interpretive process more transparent.

Reliability may be assessed through test–retest procedures, inter-rater agreement, internal consistency or repeated measurement. Internal consistency statistics should not substitute for conceptual analysis: items can correlate because they are redundant rather than because they comprehensively measure a construct.

Reproducibility requires that another competent researcher can understand and, where data access permits, repeat the analytical process. Good practice includes:

• version-controlled instruments and code;
• a data dictionary and variable definitions;
• documented inclusion and exclusion rules;
• records of data cleaning and transformations;
• preservation of analysis scripts and model specifications;
• clear reporting of missing data and deviations from the protocol; and
• secure, lawful archiving of de-identified data.

Pre-registration can distinguish planned analyses from exploratory analyses. It is valuable but does not prevent all bias. Reproducibility also depends on data provenance, software versions and access to proprietary databases.

8. Ethics, Governance and Responsible Use of Technology

Real estate research may involve financially sensitive information, household vulnerability, geolocation, tenant records and commercially confidential transactions. Ethical research requires informed consent where appropriate, voluntary participation, minimisation of harm, confidentiality and secure data management.

Researchers should consider whether combining apparently harmless datasets could re-identify individuals or reveal sensitive locations. Small-area maps may expose informal occupants or vulnerable communities. Data should be collected for a legitimate purpose, retained only as long as necessary and accessed according to documented permissions and applicable law.

AI and automated analytics can assist with document classification, image interpretation, forecasting and anomaly detection. However, models may reproduce historical discrimination, confuse listing language, infer sensitive attributes or obscure the basis of recommendations. Human oversight, bias testing, audit trails and disclosure of material limitations are essential. AI-generated text or coding should be checked against source data and the approved protocol; it is not a substitute for methodological accountability.

9. Illustrative Case: Evaluating a Transit-Oriented Residential Development

Background

A developer proposes a 420-unit apartment scheme near a planned rapid-transit station. The investment committee wants to know whether the location supports a premium rent and which household segments should be targeted.

Problem

The managerial question is whether to proceed and how to position the project. The research problem is that the relationship between transit accessibility and achievable rent is uncertain because nearby properties differ in age, size, amenities, parking provision and construction quality.

Analysis and design

The study establishes the following boundaries: apartments within the metropolitan region, transactions and achieved rents during the previous five years, and prospective tenants currently renting in comparable submarkets. It uses a mixed-method sequential explanatory design.

First, a secondary-data analysis assembles achieved rents, floor area, building age, parking, amenities and walking distance to existing stations. A hedonic model estimates the adjusted association between accessibility and rent. Robustness checks use alternative distance bands and exclude unusually large luxury units.

Second, a survey measures household composition, commuting patterns, preferred amenities and willingness to trade parking for transit access. Follow-up interviews investigate why some households value transit while others do not. Purposive sampling ensures representation of families, single professionals and older renters.

Findings and decision use

Suppose the quantitative analysis indicates that transit proximity is associated with higher rents, but the effect is concentrated among households commuting to the central business district. Interviews reveal that families place greater value on school access, storage and safety than on commuting distance. The development should therefore avoid a single “transit premium” assumption. It may require differentiated unit layouts, family-oriented amenities and a pricing strategy based on segment-specific value.

Learning points

The case demonstrates that a managerial decision may require both estimation and explanation. It also illustrates why a cross-sectional comparison alone would be insufficient: properties near stations may differ systematically from other properties. Finally, the findings support a bounded recommendation, not a universal claim that every transit-adjacent development commands a premium.

10. Common Methodological Errors and Best Practices

Several errors recur in real estate market studies:

1. Starting with a method rather than a problem. A preferred survey or database should not determine the question.
2. Overclaiming causality. Association, prediction and causal effect are different conclusions.
3. Using vague constructs. Terms such as “demand,” “quality” and “affordability” require operational definitions.
4. Ignoring selection bias. Advertised properties, responding tenants and surviving firms may not represent the population.
5. Changing measures over time. A revised vacancy definition can create artificial market trends.
6. Treating statistical significance as practical importance. A small effect may be statistically detectable but commercially irrelevant.
7. Hiding negative or contradictory evidence. Disconfirming cases improve rather than weaken serious research.
8. Neglecting data governance. Confidentiality and lawful use are part of methodology, not administrative afterthoughts.

Best practice is to maintain an explicit chain from decision problem to evidence, pilot the instruments, document assumptions, test sensitivity to reasonable alternatives and state limitations in the conclusion.

Chapter Summary

A rigorous real estate study begins by converting a managerial concern into a research problem that identifies context, uncertainty and decision relevance. Objectives specify what the study will accomplish; questions guide inquiry; hypotheses express testable relationships where appropriate; and boundaries define the population, geography, time, assets and unit of analysis.

Exploratory, descriptive, causal, cross-sectional, longitudinal and mixed-method designs serve different purposes. Design selection should follow the question, the causal logic and the available evidence. Constructs must be operationalised through defensible variables and indicators, while sampling and data provenance determine the credibility of generalisation.

Validity concerns the defensibility of interpretations, reliability concerns consistency, and reproducibility concerns transparent repetition of the research process. Ethical governance is essential when studies use personal, commercial or geospatial data. The strongest professional studies are transparent about uncertainty, distinguish association from causation, integrate appropriate evidence and align recommendations with the actual scope of the research.

Key Terms

Research problem: A clearly defined gap in knowledge that has decision or theoretical significance.

Managerial problem: A practical decision or action requiring informed judgement.

Research objective: A specific intended accomplishment of a study.

Research question: A focused question that guides evidence collection and analysis.

Hypothesis: A testable proposition about a relationship, difference or effect.

Construct: An abstract concept requiring measurement or interpretation.

Operationalisation: The process of defining how a construct will be observed or measured.

Unit of analysis: The entity about which conclusions are drawn.

Sampling frame: The operational source from which study units are selected.

Exploratory research: Research used to discover concepts, mechanisms or relevant variables.

Descriptive research: Research that estimates or portrays characteristics of a population or market.

Causal research: Research examining whether an intervention or factor produces an outcome.

Cross-sectional design: A design collecting evidence at one time or over a short period.

Longitudinal design: A design observing units or populations across time.

Mixed-method research: Intentional integration of qualitative and quantitative methods.

Validity: The defensibility of the interpretation supported by research evidence.

Reliability: The consistency of a measure or procedure.

Reproducibility: The ability to repeat an analytical process using documented materials and procedures.

Triangulation: The use of multiple sources, methods or perspectives to assess an interpretation.

Data provenance: The documented origin, processing history and characteristics of a dataset.

Review Questions

Short Answer Questions

1. How does a research problem differ from a managerial problem?
2. What is the difference between a unit of analysis and a unit of observation?
3. When is exploratory research particularly appropriate?
4. Distinguish validity from reliability.
5. Why should achieved rent be distinguished from asking rent?

Long Answer Questions

1. Explain how a real estate managerial concern can be converted into objectives, research questions and hypotheses. Illustrate with an example.
2. Compare cross-sectional and longitudinal designs in the study of property-market cycles.
3. Discuss the strengths and limitations of mixed-method research in evaluating a development proposal.
4. Explain the main forms of validity and show how each may be addressed in a tenant survey.
5. Discuss the ethical and governance issues involved in using geospatial and tenant-level data.

Analytical / Application Questions

1. A retail centre has rising vacancy but falling asking rents. Design an exploratory and subsequent descriptive study to identify the likely causes, specifying the population, data sources and boundaries.
2. A city introduces a property-tax incentive for energy-efficient buildings. Propose a defensible causal design, identify possible confounders and explain what evidence would support a causal conclusion.
3. A developer claims that apartments near a transit station command a 12% rent premium. Critically evaluate the claim and propose a research protocol capable of testing it while distinguishing association from causation.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a managerial problem from a research problem?

A. A managerial problem concerns action, whereas a research problem concerns an unresolved knowledge gap relevant to that action.

B. A managerial problem always uses qualitative data, whereas a research problem always uses quantitative data.

C. A managerial problem is theoretical, whereas a research problem is operational.

D. There is no meaningful distinction between the two.

Correct Answer: A. A managerial problem concerns action, whereas a research problem concerns an unresolved knowledge gap relevant to that action.

Question 2

Which design is most appropriate when the relevant concepts and variables are not yet well understood?

A. Exploratory design

B. Descriptive census

C. Randomised experiment

D. Confirmatory causal design

Correct Answer: A. Exploratory design

Question 3

In a study of apartment rents, the unit of analysis is the entity about which the researcher:

A. Draws conclusions

B. Stores the data

C. Conducts the interview

D. Obtains ethical approval

Correct Answer: A. Draws conclusions

Question 4

What is the principal limitation of a cross-sectional design for causal inference?

A. It cannot establish temporal order confidently.

B. It cannot measure categorical variables.

C. It always requires a very large population.

D. It cannot describe market conditions.

Correct Answer: A. It cannot establish temporal order confidently.

Question 5

Using n₀ = (Z²p(1−p))/e², with Z = 1.96, p = 0.5 and e = 0.05, the preliminary sample size is closest to:

A. 96

B. 196

C. 385

D. 1,960

Correct Answer: C. 385

Question 6

Which statement about a regression coefficient is most accurate?

A. It is automatically a causal effect.

B. It is an adjusted association under the specified model and assumptions.

C. It proves that the dependent variable was measured without error.

D. It eliminates all omitted-variable bias.

Correct Answer: B. It is an adjusted association under the specified model and assumptions.

Question 7

A measure that produces highly consistent results but does not represent the intended construct is:

A. Valid but unreliable

B. Reliable but invalid

C. Both valid and reliable

D. Neither measurable nor observable

Correct Answer: B. Reliable but invalid

Question 8

Which mixed-method structure uses quantitative results first and interviews afterward to explain unexpected patterns?

A. Sequential explanatory

B. Sequential exploratory

C. Convergent parallel

D. Embedded experimental

Correct Answer: A. Sequential explanatory

Question 9

Which practice most directly supports reproducibility?

A. Reporting only statistically significant findings

B. Changing variables during analysis without documentation

C. Preserving code, data definitions and cleaning procedures

D. Using a larger sample regardless of its sampling frame

Correct Answer: C. Preserving code, data definitions and cleaning procedures

Question 10

Why might a study of advertised property rents misrepresent the market?

A. Advertised rents may differ from achieved rents and may omit unadvertised properties.

B. Advertised rents are always lower than achieved rents.

C. Advertised rents cannot be recorded numerically.

D. Property rents have no meaningful measurement unit.

Correct Answer: A. Advertised rents may differ from achieved rents and may omit unadvertised properties.