6.3 Mixed Methods, Data Triangulation & Research-to-Design Translation

Mixed Methods, Data Triangulation & Research-to-Design Translation

Core Principle: Neither pure quantitative measurements nor isolated qualitative narratives are sufficient to fully understand complex healthcare facilities. EDAC study materials describe the mixed methods approach as one that does not commit to a single philosophical perspective and is free to adopt both quantitative and qualitative approaches, so the methods counterbalance and supplement each other. Triangulation—applying and combining several research methodologies in a study of the same topic—can strengthen the validity and reliability of research inferences; it is not merely a backup plan in case one method fails.

A single research lens often creates dangerous blind spots. For instance, physical sound level dosimeters installed in an Intensive Care Unit might record an average sound level of 58 dBA—violating the World Health Organization's guideline of 35 dBA. However, this quantitative datum alone cannot inform an architect what to fix. Are the decibels generated by HVAC mechanical velocity, medical equipment alarms, linen cart wheel rumble, or staff conversational laughter? Only by pairing continuous physical acoustic logging with qualitative shadow studies and staff interviews can the EBD team pinpoint root causes and design targeted physical interventions.

---

Mixed-Methods Research Typologies in Built Environment Research

Methodologists classify mixed-methods research into three primary typologies based on timing, sequencing, and the relative priority given to quantitative versus qualitative strands.

1. CONVERGENT PARALLEL (QUAN + QUAL Simultaneous)
   Quantitative Data Collection & Analysis  ──┐
                                              ├─> Merged Interpretation & Synthesis
   Qualitative Data Collection & Analysis   ──┘

2. EXPLANATORY SEQUENTIAL (QUAN ──> qual)
   Quantitative Priority ──> Quantitative ──> Qualitative Follow-Up ──> Deep Explanation
   (Large Survey / Data)     Outliers / Gaps  (Interviews / Shadows)    of Root Causes

3. EXPLORATORY SEQUENTIAL (QUAL ──> quan)
   Qualitative Priority  ──> Emerging Themes ──> Quantitative Tool  ──> Broad Scale
   (Interviews / Focus)      & Constructs        (Structured Survey)    Empirical Testing

1. Convergent Parallel Design (QUAN + QUAL)

In a convergent parallel design, the research team collects quantitative data and qualitative data concurrently (simultaneously) during the same research phase. Both strands are analyzed separately using their respective analytical paradigms, and the results are subsequently merged during the final interpretation to evaluate whether the findings converge, complement, or contradict each other.

• EBD Application: During a post-occupancy evaluation (POE) of an orthopedic patient unit, engineers deploy continuous calibrated sound level meters (logging equivalent continuous sound pressure levels, Leq, and maximum sound levels, Lmax, in dBA) for 14 days. Concurrently, researchers administer daily morning sleep-quality surveys to patients and conduct semi-structured interviews with night-shift charge nurses.
• Synthesis: The quantitative data confirms severe acoustic spikes (Lmax > 75 dBA) occurring at 03:00 and 05:00. The qualitative nurse interviews converge perfectly with the physical data, revealing that automated pneumatic tube deliveries crash into metal collection hoppers outside patient rooms at those precise hours.

2. Explanatory Sequential Design (QUAN ──> qual)

In an explanatory sequential design, the research begins with an expansive quantitative data collection phase that carries primary priority. Once quantitative statistical analyses are complete, researchers design a secondary, smaller qualitative phase specifically structured to explain, contextualize, or dig deeper into unexpected statistical patterns, outliers, or non-significant results.

• EBD Application (Hypothetical): A healthcare system monitors nurses wearing RFID location tracking badges across six newly opened medical-surgical hospital units. The quantitative tracking reveals a statistically significant and baffling anomaly: nurses on Unit 3 walk an average of 14.8 kilometers per 12-hour shift, compared to 8.2 kilometers on Units 1, 2, 4, 5, and 6, despite having identical floor-plate layouts and patient nurse ratios.
• Qualitative Follow-Up: To explain this variance, the researchers conduct shadow studies and semi-structured interviews with Unit 3 staff. They discover that the clean supply room door in Unit 3 has a miscalibrated electronic badge-reader that frequently jams, forcing staff to walk across the entire length of the building to use the dirty utility elevator vestibule for supplies.

3. Exploratory Sequential Design (QUAL ──> quan)

In an exploratory sequential design, the research begins with an initial qualitative phase to explore an emerging, poorly understood phenomenon. Because existing standardized measurement tools do not exist, the qualitative findings (themes, taxonomy, occupant terminology) are used to develop a new quantitative instrument (such as a structured survey instrument or space-utilization checklist), which is subsequently deployed and psychometrically tested across a large quantitative sample.

• EBD Application (Hypothetical): Hospital designers seek to understand how physical respite spaces relate to staff stress and burnout. The researchers first conducted semi-structured interviews and focus groups with 35 intensive care nurses to identify physical attributes that foster emotional decompression.
• Quantitative Follow-Up: Using the themes identified (e.g., absolute acoustic isolation, no visibility to patient beds, indirect dimmable biophilic lighting, tactile natural materials), the researchers construct a new staff respite environment scale and survey clinicians across many hospitals to test the scale's validity and reliability and examine its relationship with retention.

---

Triangulation: Neutralizing Single-Method Bias

Originally derived from navigation and surveying—where taking bearings from multiple distinct points identifies an exact geographical position—triangulation in EBD refers to combining multiple methodologies, observers, data sources, or theoretical perspectives to study a single phenomenon.

Sociologist Norman Denzin (1978) described four types of triangulation:

                                  ┌───────────────────────────────┐
                                  │    DATA SOURCE TRIANGULATION   │
                                  │  (Day vs. Night Shift Nurses,  │
                                  │   Patients, Environmental Svc)│
                                  └───────────────┬───────────────┘
                                                  │
                                                  ▼
┌──────────────────────────────┐     ┌────────────────────────┐     ┌──────────────────────────────┐
│ METHODOLOGICAL TRIANGULATION │ ──> │    RESEARCH PHENOMENON │ <── │  INVESTIGATOR TRIANGULATION  │
│ (RFID Badges + Sound Meters  │     │ (e.g., Staff Fatigue & │     │ (Architect + Infection Ctrl +│
│  + Semi-Structured Interviews│     │   Medication Errors)   │     │  Nurse Researcher Co-Coding) │
└──────────────────────────────┘     └────────────────────────┘     └──────────────────────────────┘
                                                  ▲
                                                  │
                                  ┌───────────────┴───────────────┐
                                  │     THEORY TRIANGULATION      │
                                  │  (Attention Restoration vs.   │
                                  │    Stress Recovery Theory)    │
                                  └───────────────────────────────┘

1. Methodological Triangulation

Using multiple distinct research methods to gather data on the same operational or physical construct.

• Example: Evaluating nurse physical exertion on a 36-bed progressive care unit by combining three distinct methods: (1) quantitative pedometer/RFID badge logging, (2) qualitative shadow studies logging postural strain, and (3) a validated subjective physical exhaustion survey administered at the end of each shift.

2. Data Source Triangulation (Data Triangulation)

Gathering data from diverse participant groups, multiple timeframes, or differing physical locations to ensure findings are not an artifact of a single cohort.

• Time Triangulation: Sampling during morning medication pass (08:00–10:00), shift handoff (19:00), and the dead of night (02:00–04:00).
• Space Triangulation: Gathering data across three different hospital sites operating the same spatial layout.
• Person Triangulation: Interviewing attending physicians, bedside registered nurses, certified nursing assistants, physical therapists, environmental services workers, and patients regarding room cleanliness and clutter.

3. Investigator Triangulation

Deploying two or more trained researchers with differing professional lenses to observe, collect, and analyze the same dataset.

• Healthcare Design Example: Rather than having an architectural designer analyze post-occupancy focus group transcripts alone (which might bias coding toward aesthetic finishes and sightlines), the research team pairs an architect with a clinical nurse researcher and an industrial systems engineer. Each codes the transcripts independently, followed by inter-rater consensus meetings to eliminate personal professional bias.

4. Theory Triangulation

Interpreting an empirical dataset through multiple competing or complementary theoretical paradigms to test alternative explanations.

• Healthcare Design Example: A team evaluates why access to an outdoor courtyard garden reduces post-surgical patient anxiety. They analyze the empirical data through Roger Ulrich's Stress Recovery Theory (SRT) (which posits psycho-evolutionary neuroendocrine calming triggered by unthreatening natural vegetation) and simultaneously through Stephen and Rachel Kaplan's Attention Restoration Theory (ART) (which posits cognitive fatigue relief through involuntary attention or "soft fascination").

---

The Research-to-Design Translation Pipeline

One of the greatest challenges in the EBD discipline is bridging the chasm between scientific research papers and physical construction documents. Architects do not draw "p-values" or "confidence intervals"; they draw walls, specify materials, select mechanical equipment, and organize spatial adjacencies.

One practical way to convert research into design decisions is a six-step translation pipeline (a working model, not a CHD standard):

Step 1: Clinical Problem & Strategic Visioning
  │
Step 2: Systematic Empirical Literature Review & Appraisal
  │
Step 3: Formulation of Testable EBD Hypotheses
  │
Step 4: Development of Spatial Performance Criteria
  │
Step 5: Drafting Actionable Evidence-Based Design Guidelines
  │
Step 6: High-Fidelity Mock-Up Simulation & Prototyping

1. From Empirical Findings to Spatial Performance Criteria

Scientific research typically reports high-level relational outcomes (e.g., "high acoustic reverberation in corridors increases patient sleep fragmentation"). The translation process converts this empirical finding into a spatial performance criterion—a precise, quantifiable architectural specification:

• Empirical Literature Finding: Corridors with hard gypsum board ceilings reflect mechanical equipment noises and alarm tones, maintaining reverberation times (RT60) above 1.2 seconds, which awakens light-sleep patients.
• Translated Spatial Performance Criterion (example): Patient corridor ceiling finishes will be highly sound-absorbing (for example, NRC ≥ 0.90) with a ceiling attenuation class suited to the adjacent rooms, targeting a short reverberation time set by the project's acoustic consultant.

2. Formulating Actionable Design Guidelines

A design guideline translates criteria into practical spatial layout directives without dictating an overly rigid, aesthetically stifling architectural solution. Guidelines maintain flexibility for site-specific innovation while preserving the empirical evidence baseline.

3. A Four-Part Hypothesis Template

EDAC study materials define a hypothesis as an estimated outcome or research finding for a testable subtopic, developed before data are collected so the team can define variables, testing methods, and expected relationships. A practical template that makes a hypothesis easy to test includes four parts (the template is a teaching tool, not a CHD requirement):

If [Design Intervention X], then [Outcome Y], as measured by [Metric Z], because [Mechanism W].

1. The Design Intervention (X): The physical, spatial, acoustic, or lighting intervention being incorporated into the architecture.
2. The Anticipated Outcome (Y): The specific clinical, operational, or financial improvement expected.
3. The Measurement Metric (Z): The objective metric or standardized measurement instrument that will quantify the outcome.
4. The Theoretical or Biological Mechanism (W): The scientific rationale, physiological pathway, or behavioral theory explaining why the physical intervention produces the outcome.

Example Hypotheses:

• Exemplar 1 (Patient Safety): If single-patient rooms with standardized inboard bathroom orientations and continuous grab rails are installed (X), then unassisted patient fall rates will decrease by 25% (Y), as measured by incident reports per 1,000 patient days over 12 months post-occupancy (Z), because the continuous physical handhold and direct visual sightline eliminate cognitive hesitation and physical postural instability during unassisted transfers (W).
• Exemplar 2 (Infection Control): If all medical-surgical beds are single-patient rooms with private bathrooms and a hand-hygiene sink visible on entry (X), then hospital-onset C. difficile infections will decrease (Y), as measured by laboratory-identified event rates per 10,000 patient days compared with baseline (Z), because patients no longer share toilets and caregivers have convenient hand-hygiene access (W).

4. Full-Scale Mock-Ups and Functional Prototyping

The final translational bridge between design hypothesis and construction is the execution of high-fidelity physical mock-ups.

• Foam-Core vs. Modular Finished Mock-Ups: Teams frequently begin with rapid, low-cost foam-core spatial mock-ups during schematic design to test bulk clearances, equipment swinging arcs, and sightlines. During design development, hospitals construct fully finished, modular functional mock-up rooms equipped with operational medical gases, lighting circuits, headwalls, and bathroom fixtures.
• Simulation Testing: The multidisciplinary team executes high-stress clinical simulations (e.g., adult cardiac resuscitation code, bariatric patient bed-to-toilet transfer, emergency cesarean section handoff) inside the mock-up. Clinicians wear eye-tracking glasses and heart-rate variability monitors, while video cameras record ergonomic collisions and equipment interference.
• Iterative Refinement: Mock-up testing routinely uncovers design errors—such as IV poles striking overhead ceiling lights or medical gas outlets positioned behind patient beds—allowing the design team to revise architectural working drawings before hundreds of millions of dollars are poured into permanent construction.

---

High-Yield Summary Table for Mixed Methods & Translation