How to Convert Key Design Issues into Research Questions

in: AIACC / 1 Comment

Share this article:

Architectural designers are prolific generators of research topics. The very essence of the design process is to identify questions and pose answers. This process, emphasized by William Peña and Steve Parshall in Problem Seeking (2012), is a fertile ground for the development of research hypotheses.

A hypothesis is a statement reflecting the opinion of a researcher regarding the impact of a phenomenon or an intervention. The researcher proposes an explanation for an event, the purpose of which is that, when specific circumstances repeat themselves, the results will be identical. Hypotheses are integrated in the scientific method, a series of stages involving identifying an issue, reviewing what is known about the issue, generating a hypothesis, conducting an experiment, analyzing the data, and drawing conclusions. Evidence-based design adheres to the same sequence of operations.

Programming and hypotheses

The distillation and prioritization of design research questions occurs during the programming process, when the primary design goals are generated. While the uninitiated may consider a building program a mere list of spaces, designers are aware that the crux of a good program is the identification of project objectives. These objectives are the equivalent of research hypotheses.

A traditional, competent building program includes:

  • Mission statement
  • Design objectives
  • Functional narrative
  • Adjacency diagrams
  • Space lists
  • Code analysis
  • Preliminary cost analysis, and
  • Support materials (site visit summaries, meeting minutes).

In the interest of substantiating these objectives, an evidence-based design (EBD) program would add:

  • Research articles that address the primary design objectives
  • Studies on EBD return on investment, and
  • Pre-occupancy data embedded in functional narratives, adjacency diagrams and space summaries.

Hypotheses and research topics
A design goal becomes a hypothesis when the designer attempts to gather data and draw conclusions about the effectiveness of the environmental intervention. A hypothesis/design goal may be motivated by one of three circumstances:

  1. A question that arises due to pressing research topics;
  2. A question in response to an untested design innovation; or
  3. An inquiry about the effectiveness of core design goals (Shepley, 2010).

Research topics. Examples of questions that might be considered to be pressing research topics in healthcare are:

  • What are the minimal goals for sustainable design in healthcare buildings?
  • What is the relative effectiveness of decentralized, centralized, or hybrid nursing stations?
  • Do single family NICU rooms have a positive impact on infant, family, and staff outcomes?
  • What are the pros and cons of LDR (Labor, Delivery, and Recovery) versus LDRP (Labor, Delivery, Recovery, and Postpartum) rooms?
  • What is the impact of satellite support spaces in inpatient units?
  • What is the impact of open versus closed nurse stations in psychiatric facilities?
  • Which has a greater impact on staff error reduction, same-handed rooms or mirrored rooms? and
  • Should windows to the outdoors be provided in all NICU (Neonatal Intensive Care Unit) single-family rooms?

Design innovations. Regarding innovation, frequently a designer will introduce a novel feature and need confirmation of its effectiveness. For example, when the first free-standing, single-room maternity care facility was built in California (at Grossmont Women’s Center, San Diego), the building was quickly imitated by hospitals seeking to provide similar care. When a building introduces an innovation that is likely to be replicated elsewhere, the responsibility to conduct research that confirms the appropriateness of the design solution is critical. This is necessary in spite of extensive action research that might have taken place prior to the schematic design process. In the case of the Grossmont Women’s Center, the hospital was advised by a consulting firm that was in the process of developing protocols for LDR/LDRP. Additionally, the design team visited the site where the first LDR/LDRP bed was specified, and built a 6-bed prototype, which was tested for more than a year.

One of the design goals of the Grossmont Women’s Center was to enhance the care of obstetric patients and their families. The research question, therefore, was: Does single-room maternity care enhance the experience of obstetric patients and their families? To address this question, specific hypotheses were developed and a post occupancy evaluation was conducted to confirm whether the design goals were met. Staff completed surveys and participated in interviews, and behavior of staff and patients was observed (Shepley, Bryant & Frohman, 1995).

Healthcare, Evidence-Based Design

Grossmont LDRP room – Photo © The Design Partnership LLP

Core design goals. Regarding research hypotheses that might be generated around design goals, the Bailey Boushay House, a facility for persons with AIDs, included the provision of access to nature as a healing design amenity among its primary design goals. According to the program,

A connection to the earth and its plant life, to fresh air and to sunlight are important attributes of any residential setting. They will be of particular significance in this unique residence. Every opportunity should be taken to promote access to these natural elements. (AIDS Housing of Washington, 1989)

Based on this design goal, a post-occupancy evaluation was conducted to determine whether staff and patients positively received the presence of nature features (solaria, green house, porch, patio). Hypotheses included, “Spaces that provide access to nature will be frequently used by patients and their families,” and “Spaces that provide access to nature will be preferred by staff and patients.” The post-occupancy evaluation included surveys, interviews, and behavior mapping, which corroborated the decision to incorporate these amenities, although in one case a patient lounge/atria was converted to another patient support space (Shepley, Frohman & Wilson, 1999).

Healthcare, Evidence-Based Design

Bailey Boushay House – Photo © Bailey Boushay House

Components of Hypotheses
Most design hypotheses contain four components: the intervention or independent variable, the outcome or dependent variable, subjects, and responses. The physical environment is typically the independent variable, while the dependent variables are the variables that are assessed to measure the impact of the physical environment. The scale of potential independent variables is broad and ranges from the entire site to a piece of furniture. Other physical environmental variables include color, scale, climate, noise levels, and complexity. Dependent variables can be psychological (satisfaction, emotional impact, etc.), behavioral (frequency and duration of use, etc.), physiological (pupil dilation, heart rate, etc.) or mechanical (cost, air quality, etc.).

In the context of healthcare facilities, staff includes: patients, families, visitors, nurses, physicians, technicians, clerical staff, administrators, volunteers, housekeeping, maintenance, food services, and others. Data is frequently gathered about age, duration of employment, duration of visit, gender, etc., as required by some Institutional Review Boards.

Translating design goals into hypotheses
A well-constructed hypothesis should be specific in intent. The following three examples demonstrate this concept.

Case 1: A example of a design goal that has not undergone this refinement would be:

The infusion suite should support the patient’s experience.

This goal includes all four of the components mentioned above. The infusion suite is the independent variable, the outcome is comfort, the dependent variable is support of comfort, and the subjects are patients. However, to translate this into an hypothesis, more specificity is required. If this objective were to be converted for the purpose of a research project, a more appropriate design goal, which might lead to a hypothesis, would be:

The views of nature in the infusion suite should increase patient comfort.

This is more readily translated into a hypothesis, which might be:

Views of nature in an infusion suite increase comfort as measured by patient self-reported perception of comfort.

While some designers might be concerned that this level of specificity limits access to other information, from the perspective of a researcher the question has to be narrow to allow for exploration within a reasonable timeframe.

Case 2: Another example of a broad design goal would be the following:

The psychiatric facility should be designed to reduce undesirable resident behavior.

This would more readily be converted into a hypothesis, if it had a little more detail:

The psychiatric facility should have private rooms to reduce violent events;

which could be translated into the following hypothesis:

Psychiatric facilities with private rooms will have fewer incidents of aggression toward staff and patients than facilities with wards.

The interesting twist to this hypothesis is that it entails a comparison, which is a common technique for evaluating a design decision.

Case 3: A loose design goal might be:

Private NICU rooms are needed to support family-centered care.

A narrower design goal would be:

Private NICU rooms with family sleep space are needed to support family-center care.

To be an appropriate hypothesis, the outcomes need to be yet more specific.

Private NICU rooms with family sleep space result in greater family satisfaction and more time spent in the room with infants than open bay NICUs.

The primary differences among these three hypotheses and their associated design goals is the detail regarding the physical environment and the responses that need to be measured. A secondary difference is the transition from an intention to a statement regarding expected outcomes.

Steps in support of hypothesis generation
For those who are interested in viewing their projects as future research labs, establishing hypotheses is an advantage. To this end, several steps can be taken:

  1. Retain project mission and goal information. Often, once a project is in construction, the preliminary documentation is lost. If firms are interested in conducting research on their projects, the development of hypotheses should take place during the planning and programming phases and be re-visited periodically.
  2. Review EBD journals to examine the structure of hypotheses. Health Environments Design & Research, Environment and Behavior, and the Journal of Environmental Psychology are good sources for hypotheses.
  3. Generate identical hypotheses or design goals across projects, which will enable comparisons and generate more data.
    Create design goals that include independent and dependent variables, subject(s), and response elements.

“A problem well-stated is a problem half-solved.” The generation of hypotheses during the design process is likely to clarify design thinking, support problem solving, and provide the base for potential research.

Peña, W. & Parshall, S. (2012). Problem Seeking. New York, NY: Wiley. 5th edition.
Shepley, M. (2010). Health Facility Evaluation for Design Practitioners. Myersville, MD: Asclepion Publishing.
Shepley, M., Bryant, C., & Frohman, B. (1995). “Using a post-occupancy study to validate a building prototype: An evaluation of a new women’s medical center.” Journal of Interior Design, 21 (2), 19–40.
Shepley, M., Frohman, B, Wilson, P. (1999). “Designing for persons with AIDS: A post-occupancy study at the Bailey-Boushay House.” Journal of Architectural & Planning Research, 16 (1), 17–32.


Mardelle Shepley, FAIA

Mardelle Shepley, FAIA, D. Arch., is Director of the Center for Health Systems & Design at Texas A&M University, where she specializes in architectural design, social architecture, health care facility design, applied research, and environmental psychology.

More Posts

Share this article:

1 Response to How to Convert Key Design Issues into Research Questions


Leave a Reply

Join the discussion!